Encyclopedia of Mobile Computing and Commerce David Taniar Monash University, Australia
Volume I A-Mobile Hunters
INFORMATION SCIENCE reference Hershey • London • Melbourne • Singapore
Acquisitions Editor: Development Editor: Senior Managing Editor: Managing Editor: Assistant Managing Editor: Copy Editor: Typesetter: Support Staff: Cover Design: Printed at:
Kristin Klinger Kristin Roth Jennifer Neidig Sara Reed Diane Huskinson Maria Boyer and Alana Bubnis Diane Huskinson Sharon Berger, Mike Brehm, Elizabeth Duke, and Jamie Snavely Lisa Tosheff Yurchak Printing Inc.
Published in the United States of America by Information Science Reference (an imprint of Idea Group Inc.) 701 E. Chocolate Avenue, Suite 200 Hershey PA 17033 Tel: 717-533-8845 Fax: 717-533-8661 E-mail:
[email protected] Web site: http://www.idea-group-ref.com and in the United Kingdom by Information Science Reference (an imprint of Idea Group Inc.) 3 Henrietta Street Covent Garden London WC2E 8LU Tel: 44 20 7240 0856 Fax: 44 20 7379 0609 Web site: http://www.eurospanonline.com Copyright © 2007 by Idea Group Inc. All rights reserved. No part of this publication may be reproduced, stored or distributed in any form or by any means, electronic or mechanical, including photocopying, without written permission from the publisher. Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or companies does not indicate a claim of ownership by IGI of the trademark or registered trademark. Library of Congress Cataloging-in-Publication Data Encyclopedia of mobile computing and commerce / David Taniar, editor. p. cm. Summary: “Nowadays, mobile communication, mobile devices, and mobile computing are widely available. The availability of mobile communication networks has made a huge impact to various applications, including commerce. Consequently, there is a strong relationship between mobile computing and commerce. This book brings to readers articles covering a wide range of mobile technologies and their applications”--Provided by publisher. Includes bibliographical references and index. ISBN 978-1-59904-002-8 (hardcover) -- ISBN 978-1-59904-003-5 (ebook) 1. Mobile computing--Encyclopedias. 2. Mobile communication systems--Encyclopedias. 3. Mobile commerce--Encyclopedias. I. Taniar, David. QA76.59.E47 2007 004.16503--dc22 2006039745 British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library. All work contributed to this encyclopedia set is new, previously-unpublished material. The views expressed in this encyclopedia set are those of the authors, but not necessarily of the publisher.
Editorial Advisory Board Bernady O. Apduhan Kyushu Sangyo University, Japan
Kevin H. Liu EMC Corporation, USA
Irfan Awan University of Bradford, UK
Ismail Khalil Ibrahim Johannes Kepler University Linz, Austria
Leonard Barolli Fukuoka Institute of Technology, Japan
Jianhua Ma Hosei University, Japan
Stephane Bressan National University of Singapore, Singapore
Zakaria Maamar Zayed University, UAE
Kuo-Ming Chao Coventry University, UK
Joseph K. Ng Hong Kong Baptist University, Hong Kong
Mieso Kabeto Denko University of Guelph, Canada
Wenny Rahayu La Trobe University, Australia
Mustafa M. Deris Kolej Universiti Teknologi Tun Hussein Onn, Malaysia
Elhadi Shakshuki Acadia University, Canada
Arjan Durresi Louisiana State University, USA John Goh (Assistant Editor-in-Chief) Monash University, Australia D. Frank Hsu Fordham University, USA Hai Jin Huazhong University of Science and Technology, China
Timothy K. Shih Tamkang University, Taiwan Nguyen Manh Tho Vienna University of Technology, Austria Laurence T. Yang St. Francis Xavier University, Canada Muhammad Younas Oxford Brookes University, UK
List of Contributors Abdul-Mehdi, Ziyad Tariq / Multimedia University, Malaysia .......................................................233, 369, 693, 947 Ahmad, Ashraf M. A. / National Chiao Tung University, Taiwan.............................................................................627 Al Kattan, Ibrahim / American University of Sharjah, UAE....................................................................................682 Alam, Muhammad Tanvir / Bond University, Australia...................................................................................724, 778 Aleksy, Markus / University of Mannheim, Germany . .....................................................................................160, 744 Alexiou, Antonios / Patras University, Greece . ..........................................................................................................20 Al-Khalifa, Hend S. / Southampton University, UK .................................................................................................569 Almeida, Hyggo / Federal University of Campina Grande, Brazil . .........71, 249, 260, 341, 621, 877, 974, 978, 1011 Al-Salman, AbdulMalik S. / King Saud University, Saudi Arabia . .........................................................................569 Amara-Hachmi, Nejla / University of Paris 13, France ..........................................................................................717 Anagnostopoulos, Christos / University of Athens, Greece .....................................................................................856 Antonellis, Dimitrios / Research Academic Computer Technology Institute, Greece & University of Patras, Greece.....................................................................................................................................20 Antonellis, Ioannis / University of Patras, Greece ...................................................................................................119 Arbaiy, Nureize / Kolej Universiti Teknologi Tun Hussein Onn, Malaysia ..............................................................763 Avola, Danilo / Instituto di Ricerche Sulla Popolazione e le Politiche Sociali, Italy ..............................................1050 Aziz Basi, Hussein M. / Multimedia University, Malaysia ...............................................................................369, 446 Azzuhri, Saaidal Razalli Bin / Malaysia University of Science and Technology, Malaysia ....................................253 Baba, Takaaki / Waseda University, Japan .......................................................................................................804, 820 Bakhouya, M. / The George Washington University, Washington DC, USA ............................................................954 Bandyopadhyay, Subir K. / Indiana University Northwest, USA ..............................................................................32 Barbosa, Nadia / Federal University of Campina Grande, Brazil . ..........................................................................877 Basole, Rahul C. / Georgia Institute of Technology, USA .........................................................................................481 Beer, Martin / Sheffield Hallam University, UK . ......................................................................................................528 Belsis, Meletis / Telecron, Greece ............................................................................................................................1028 Bin Mamat, Ali / Universiti Putra Malaysia, Malaysia ....................................................................................233, 693 Bodomo, Adams / University of Hong Kong, Hong Kong ........................................................................................562 Bokor, László / Budapest University of Technology and Economics, Hungary ..........................................................51 Bose, Indranil / University of Hong Kong, Hong Kong ..............................................................................................96 Bouras, Christos / Research Academic Computer Technology Institute, Greece & University of Patras, Greece.............................................................................................................................20, 119 Bradley, John F. / University College Dublin, Ireland . ............................................................................................243 Bryant, Barrett R. / University of Alabama at Birmingham, USA ...........................................................................436 Bublitz, Frederico / Federal University of Campina Grande, Brazil .......................................................................877 Byrne, Caroline / Institute of Technology Carlow, Ireland .......................................................................................310 Carroll, John M. / The Pennsylvania State University, USA.....................................................................................291 Caschera, Maria Chiara / Consiglio Nazionale delle Ricerche, Italy . ..........................................................675, 1050 Chalmers, Kevin / Napier University, Scotland ........................................................................................................576 Chan, Alvin T. S. / The Hong Kong Polytechnic University, Hong Kong .................................................................749 Chand, Narottam / Indian Institute of Technology Roorkee, India ..........................................................................172 Chang, Jun-Yang / National Kaohsiung University of Applied Sciences, Taiwan............................................352, 616 Chang, Elizabeth / Curtin University of Technology, Australia . ..............................................................................108
Chen, Jengchung V. / National Cheng Kung University, Taiwan .............................................................................894 Chen, Shuping / Beijing University of Posts & Telecommunications, China ...........................................................940 Chetan, Kumar S / NetDevices India Pvt Ltd., India ...............................................................................................195 Chin, Choong Ming / British Telecommunications (Asian Research Center), Malaysia .........................424, 700, 729 Ching-Bin Tse, Alan / The Chinese University of Hong Kong, Hong Kong .............................................................283 Cho, Vincent / Hong Kong Polytechnic University, Hong Kong . ...............................................................................38 Chochliouros, Stergios P. / Hellenic Telecommunications Organization S.A., Greece ............................................581 Chochliouros, Ioannis P. / Hellenic Telecommunications Organization S.A., Greece .............................................581 Chow, Chi-Yin / University of Minnesota – Twin Cities, USA ..................................................................................749 Chuang, Li-Yeh / I-Shou University, Taiwan.....................................................................................................352, 616 Correia, Eduardo / Christchurch Polytechnic Institute of Technology, New Zealand . ............................................996 Crowther, Paul / Sheffield Hallam University, UK ...................................................................................................528 Curran, Kevin / University of Ulster, Northern Ireland . ................................................................................265, 1022 Cycon, Hans L. / FHTW Berlin, Germany ................................................................................................................589 da Cunha Borelli, Walter / State University of Campinas, Brazil . ..........................................................................272 Dagiuklas, Tasos / Technical Institute of Messolongh, Greece .........................................................................357, 796 Dananjayan, P. / Pondicherry Engineering College, India . .............................................................................149, 810 da Silva Oliveira, Elthon Allex / Federal University of Alagoas – Campus Arapiraca, Brazil . .............................987 de Araújo Lima, Emerson Ferreira / Federal University of Campina Grande, Brazil ..........................................987 de Carvalho Gomes, Yuri / Federal University of Campina Grande, Brazil .........................................................1011 de Figueiredo, Jorge César Abrantes / Federal University of Campina Grande, Brazil . ......................................987 de Leoni, Massimiliano / University of Rome “La Sapienza”, Italy ......................................................................1043 de Oliveira, Juliano Rodrigues Fernandes / Federal University of Campina Grande, Brazil .............................1011 De Rosa, Fabio / University of Rome “La Sapienza”, Italy ....................................................................................1043 Decker, Michael / University of Karlsruhe, Germany . .....................................................................................398, 711 Denko, Mieso Kabeto / University of Guelph, Canada ............................................................................................328 Deris, Mustafa M. / Kolej Universiti Teknologi Tun Hussein Onn, Malaysia ..........................................................763 Di Noia, Tommaso / Politecnico di Bari, Italy ............................................................................................................43 Di Sciascio, Eugenio / Politecnico di Bari, Italy .........................................................................................................43 Diekmann, Thomas / University of Goettingen, Germany .......................................................................................124 Dillon, Tharam S. / University of Technology, Sydney, Australia .............................................................................108 Dirs, Mustafa M. / College University Technology Tun Hussein Onn, Malaysia .....................................233, 693, 947 Djordjevic-Kajan, Slobodanka / University of Nis, Serbia .............................................................................129, 660 Damodaran, Dhilak / Monash University, Australia ..............................................................................................1015 Donini, Francesco Maria / Università della Tuscia, Italy ..........................................................................................43 Dudás, István / Budapest University of Technology and Economics, Hungary ..........................................................51 El Fallah-Seghrouchni, Amal / University of Paris 6, France . ...............................................................................717 El Morr, Christo / York University, Canada .............................................................................................................632 El-Said, Mostafa / Grand Valley State University, USA . ....................................................................................63, 688 Ferreira, Glauber / Federal University of Campina Grande, Brazil . ......................................................................877 Ferri, Fernando / Istituto di Richerche Sulla Popolazione e le Politiche Sociali – CNR, Italy . ....................675, 1050 Fleet, Gregory John / University of New Brunswick at Saint John, Canada .............................................................78 Flores, Andres / University of Comahue, Argentina.....................................................................................................59 Freire de Souza Santos, Danilo / Federal University of Campina Grande, Brazil . ................................................341 Gaber, J. / Université de Technologie de Belfort-Montbéliard, France ....................................................................954 Gandhamaneni, Jayasree / Indiana University Purdue University Indianapolis, USA ...........................................436 Ganoe, Craig H. / The Pennsylvania State University, USA......................................................................................291 Garcia, Juan / Illinois State University, USA ............................................................................................................461 García-Macías, J. Antonio / CICESE Research Center, Mexico ..............................................................................773 Gardikis, G. / University of Aegean, Greece .............................................................................................................889 Garret, Bernie / University of British Columbia, Canada ........................................................................................754 Goldberg, Steve / INET International Inc., Canada ...............................................................................................1004 Grahn, Kaj / Arcada Polytechnic, Finland ...............................................................................................................839 Griffiths, Mark / Nottingham Trent University, UK . ................................................................................................553
Grifoni, Patrizia / Istituto di Ricerche Sulla Popolazione e le Politiche Sociali – CNR, Italy .......................675, 1050 Gritzalis, Stefanos / University of the Aegean, Greece............................................................................................1028 Guan, Jihong / Tongji University, China .....................................................................................................84, 213, 789 Guan, Sheng-Uei / Brunel University, UK ........................................................................................334, 345, 429, 826 Gurău, Călin / Montpellier Business School, France .......................................................................................557, 999 Gyasi-Agyei, Amoakoh / Central Queensland University, Australia .......................................................................165 Hadjiefthymiades, Stathes / University of Athens, Greece ..............................................................................856, 863 Hagenhoff, Svenja / Georg-August-University of Goettingen, Germany .................................................................124 Hartung, Frank / Ericsson GmbH, Germany ...........................................................................................................611 Hegedüs, Péter / Budapest University of Technology and Economics, Hungary . ....................................................393 Herbster, Raul Fernandes / Federal University of Campina Grande, Brazil ..................................................260, 974 Hiew, Pang Leang / British Telecommunications (Asian Research Center), Malaysia ....................................487, 906 Hoh, Simon / British Telecommunications (Asia Research Center), Malaysia .........................................................138 Horn, Uwe / Ericsson GmbH, Germany ....................................................................................................................611 Hosszú, Gábor / Budapest University of Technology and Economics, Hungary ......................................................393 Hsu, Wen-Jing / Nanyang Technological University, Singapore ..............................................................................734 Hu, Wen-Chen / University of North Dakota, USA . .................................................................................................302 Huang, Bo / Waseda University, Japan . ............................................................................................................804, 820 Huang, Hong / New Mexico State University, USA . .................................................................................................202 Hung, Humphry / Hong Kong Polytechnic University, Hong Kong ..........................................................................38 Hussain, Omar Khadeer / Curtin University of Technology, Australia ...................................................................108 Hussain, Farookh Khadeer / University of Technology, Australia ..........................................................................108 Ibrahim, Hamidah / Universiti Putra Malaysia, Malaysia.......................................................................233, 693, 947 Ifinedo, Princely / University of Jyvaskyla, Finland..................................................................................................605 Imre, Sándor / Budapest University of Technology and Economics, Hungary . .........................................................51 Iris, Reychav / Bar-Ilan University, Israel ................................................................................................................413 Jasimuddin, Sajjad M. / University of Wales – Aberystwyth, UK ............................................................................520 Jayaputera, James W. / Monash University, Australia . ...........................................................................................739 Jeong, Eui Jun / Michigan State University, USA . ...........................................................................................185, 928 Jiménez, Leonardo Galicia / CICESE Research Center, Mexico .............................................................................773 Joshi, R. C. / Indian Institute of Technology Roorkee, India .....................................................................................172 Ju, Khoo Wei / Malaysia University of Science and Technology, Malaysia .............................................................912 Kaldanis, Vasileios S. / NTUA, Greece .........................................................................................................................1 Kalliaras, Panagiotis / National Technical University of Athens, Greece ........................................381, 387, 960, 981 Kampmann, Markus / Ericsson GmbH, Germany . .................................................................................................611 Kamthan, Pankaj / Concordia University, Canada . .............................................................................9, 25, 277, 375, Kao, I-Lung / IBM, USA . ..........................................................................................................................................302 Karlsson, Jonny / Arcada Polytechnic, Finland .......................................................................................................839 Karnouskos, Stamatis / SAP AG, Germany . ............................................................................................................706 Kartham, Pankaj / Concordia University, Canada ........................................................................................9, 25, 277 Kaspar, Christian / University of Goettingen, Germany ..........................................................................................124 Katsukura, Akihisa / Dentsu Inc., Japan . ................................................................................................................639 Keegan, Stephen / University College Dublin, Ireland . ...........................................................................................310 Kennedy, David M. / Hong Kong University, Hong Kong . ......................................................................................317 Kerridge, Jon / Napier University, Scotland .............................................................................................................576 Khashchanskiy, Victor / First Hop Ltd., Finland ...............................................................................................15, 785 Kim, Dan J. / University of Houston Clear Lake, USA .....................................................................................185, 928 Kini, Ranjan B. / Indiana University Northwest, USA ...............................................................................................32 Kitisin, Sukumal / Kasetsart University, Thailand ...................................................................................................220 Kleinschmidt, João Henrique / State University of Campinas, Brazil ....................................................................272 Korhonen, Jouni / TeliaSonera Corporation, Finland . ............................................................................................966 Korthaus, Axel / University of Mannheim, Germany . ..............................................................................................160 Kotsopoulos, Stavros / University of Patras, Greece .......................................................................................357, 796 Koukia, Spiridoula / Universtiy of Greece, Greece ..................................................................................................116
Koumaras, H. / N.C.S.R., Demokritos, Greece .................................................................................................758, 889 Kourtis, A. / N.C.S.R., Demokritos, Greece . .....................................................................................................758, 889 Kovács, Ferenc / Budapest University of Technology and Economics, Hungary .....................................................393 Kritzner, Jan / Aachen University, Germany ............................................................................................................611 Kustov, Andrei L. / First Hop Ltd., Finland .......................................................................................................15, 785 Kvasnica, Milan / Tomas Bata University, Zlin, Czech Republic . ....................................................................403, 651 Lalopoulos, George K. / Hellenic Telecommunications Organization S.A., Greece . ...............................................581 Lang, Jia / Nice Business Solutions Finland, Finland . .............................................................................................785 Lau, Chiew-Tong / Nanyang Technological University, Singapore ..........................................................................734 Le, Phu Dung / Monash University, Australia ........................................................................................227, 832, 1015 Lee, Cheon-Pyo / Carson-Newman College, USA ....................................................................................................442 Lee, Dennis / The University of Queensland, Australia & The Australian CRC for Interactive Design, Australia . ........................................................................................933 Lei, Pouwan / University of Bradford, UK ................................................................................................................455 Leong, Hong Va / The Hong Kong Polytechnic University, Hong Kong ..................................................................749 Leu, Huei / Industrial Technology Research Institute, Taiwan...................................................................................178 Liberati, Diego / Italian National Research Council, Italy .........................................................................................68 Lim, Say Ying / Monash University, Australia ..........................................................................................102, 154, 849 Lin, Chad / Edith Cowan University, Australia . .......................................................................................................178 Lin, Koong / Taiwan National University of the Arts, Taiwan ..................................................................................178 Liu, Chao / Waseda University, Japan . .............................................................................................................804, 820 Lívio Vasconcelos Guedes, Ádrian / Federal University of Campina Grande, Brazil ............................................249 Lonthoff, Jörg / Technische Universität Darmstadt, Germany..................................................................................510 Loureiro, Emerson / Federal University of Campina Grande, Brazil.................................................................71, 877 Luís do Nascimento, José / Federal University of Campina Grande, Brazil ...........................................................341 Maamar, Zakaria / Zayed University, UAE ..............................................................................................................190 Mahatanankoon, Pruthikrai / Illinois State University, USA . ................................................................................461 Mahmoud, Qusay H. / University of Guelph, Canada .............................................................................................190 Malik, Haroon / Acadia University, Canada . ...........................................................................................................328 Mamat, Ali Bin / FSKTM – UPM, Malaysia......................................................................................................693, 947 Maricar, Habeebur Rahman / American University of Sharjah, UAE ....................................................................682 Marques, Stefânia / Federal University of Campina Grande, Brazil .......................................................................978 Martakos, D. / University of Athens, Greece . ...........................................................................................................758 Massimi, Michael / University of Toronto, Canada...................................................................................................291 McMeel, Dermott / University of Edinburgh, Scotland ............................................................................................516 Mecella, Massimo / University of Rome, Italy ........................................................................................................1043 Menipaz, Ehud / Ben-Gurion University, Israel .......................................................................................................413 Merten, Patrick S. / University of Fribourg, Switzerland . .......................................................................................466 Misra, Manoj / Indian Institute of Technology Roorkee, India .................................................................................172 Mitchell, Stella / IBM T. J. Watson Research, USA ...................................................................................................644 Morais, Marcos / Federal University of Campina Grande, Brazil ...........................................................................260 Muhlberger, Ralf / The University of Queensland, Australia & The Australian CRC for Interactive Design, Australia . .......................................................................................933 Muldoon, Conor / University College Dublin, Ireland .............................................................................................243 Nanopoulos, Alexandros / Aristotle University, Greece ...........................................................................................660 Nishiyama, Mamoru / Dentsu Communication Institute Inc., Japan .......................................................................639 O’Grady, Michael J. / University College Dublin, Ireland ....................................................................243, 769, 1034 O’Hare, Gregory M. P. / University College Dublin, Ireland ........................................................243, 310, 769, 1034 O’Hare, Peter / University College Dublin, Ireland..................................................................................................310 Okazaki, Shintaro / Autonomous University of Madrid, Spain ........................................................296, 635, 639, 885 Oliveira, Loreno / Federal University of Campina Grande, Brazil ............................................................71, 621, 877 Olla, Phillip / Madonna University, USA ..................................................................................................................504 Olson, Andrew M. / Indiana University Purdue University Indianapolis, USA .......................................................436 Orosz, Mihály / Budapest University of Technology and Economics, Hungary .......................................................393
Pallis, E. / Technological Educational Institute of Crete, Greece ......................................................................758, 889 Papadopoulos, Apostolos N. / Aristotle University, Greece .....................................................................................660 Papageorgiou, P. / National Technical University of Athens, Greece .......................................................................387 Parsons, David / Massey University, New Zealand ...................................................................................................525 Patel, Keyurkumar J. / Box Hill Institute, Australia ................................................................................................365 Patel, Umesh / Box Hill Institute, Australia ..............................................................................................................365 Patrikakis, Charalampos Z. / NTUA, Greece ..............................................................................................................1 Paulo de Assis Barbosa, Luiz / Federal University of Campina Grande, Brazil ...................................................1011 Pavlovski, Christopher J. / IBM Corporation, Australia .................................................................................644, 870 Peng, Mugen / Beijing University of Posts & Telecommunications, China ......................................................921, 940 Perkusich, Angelo / Federal University of Campina Grande, Brazil ......71, 249, 260, 341, 621, 877, 974, 978, 1011 Petrova, Krassie / Auckland University of Technology, New Zealand ..............................................................497, 899 Piscitelli, Giacomo / Politecnico di Bari, Italy . ..........................................................................................................43 Plant, Laurence / IBM Corporation, Australia .........................................................................................................870 Politis, Ilias / University of Patras, Greece .......................................................................................................357, 796 Poulopoulos, Vassilis / Research Academic Computer Technology Institute, Greece & University of Patras, Greece . ................................................................................................................................119 Pousttchi, Key / University of Augsburg, Germany ..................................................................................................547 Predić, Bratislav / University of Nis, Serbia .............................................................................................................129 Protonotarios, Vasileios E. / NTUA, Greece .................................................................................................................1 Pulkkis, Göran / Arcada Polytechnic, Finland .........................................................................................................839 Queiroga, Miguel / Federal University of Campina Grande, Brazil ........................................................................978 Raisinghani, Mahesh S. / TWU School of Management, USA .................................................................................472 Raje, Rajeev R. / Indiana University Purdue University Indianapolis, USA . ..................................................207, 436 Ramamurthy, M. B. / Multimedia University, Malaysia ..........................................................................................446 Ramli, Azizul Azhar / Kolej Universiti Teknologi Tun Hussein Onn, Malaysia . .....................................................763 Reid, Jeffery G. / xwave Saint John, Canada . ............................................................................................................78 Rigou, Maria / University of Patras, Greece & Research Academic Computer Technology Institute, Greece.........116 Romdhami, Imed / Napier University, Scotland . .....................................................................................................576 Rouse, William B / Georgia Institute of Technology, USA . ......................................................................................481 Ruta, Michele / Politecnico di Bari, Italy . ..................................................................................................................43 Ruzzelli, Antonio G. / University College Dublin, Ireland .....................................................................................1034 Saravanan, I. / Pondicherry Engineering College, India . ................................................................................149, 810 Schader, Martin / University of Mannheim, Germany . ....................................................................................160, 744 Schmidt, Thomas C. / HAW Hamburg, Germany .............................................................................................541, 589 Seet, Boon-Chong / Nanyang Technological University, Singapore .........................................................................734 Sekkas, Odysseas / University of Athens, Greece .....................................................................................................863 Serenko, Alexander / Lakehead University, Canada ................................................................................................143 Shakshuki, Elhadi / Acadia University, Canada ......................................................................................................328 Shirali-Shahreza, Mohammad / Sharif University of Technology, Iran ..................................................................666 Silva Rocha, Jerônimo / Federal University of Campina Grande, Brazil . ..............................................................249 Sim, Moh Lim / Multimedia University, Malaysia . ..................................................................................424, 700, 729 Simitsis, Alkis / National Technical University of Athens, Greece...........................................................................1028 Singh, Rohit / Monash University, Australia ...........................................................................................................1015 Sircar, Ranapratap / Wipro Technologies, India ......................................................................................................195 Sirmakessis, Spiros / Technological Institution of Messolongi & Research Academic Computer Technology Institute, Greece ................................................................................116 Sivaradje, G. / Pondicherry Engineering College, India ..................................................................................149, 810 Smyth, Elaine / University of Ulster, Northern Ireland . .........................................................................................1022 So, Simon / Hong Kong Institute of Education, Hong Kong .....................................................................................419 Sotiriou, Athanasios-Dimitrios / National Technical University of Athens, Greece ........................381, 387, 960, 981 Souto, Sabrina / Federal University of Campina Grande, Brazil .............................................................................978 Spiliopoulou, Anastasia S. / Hellenic Telecommunications Organization S.A., Greece . .........................................581 Srikhutkhao, Nopparat / Kasetsart University, Thailand ........................................................................................220
Steinert, Martin / University of Fribourg, Switzerland ............................................................................................466 Stojanović, Dragan / University of Nis, Serbia .................................................................................................129, 660 Suradi, Zurinah / Kolej Universiti Teknologi Tun Hussein Onn, Malaysia . ............................................................763 Tan, Chor Min / British Telecommunications (Asian Research Center), Malaysia ..........................424, 700, 729, 906 Tarkoma, Sasu / Helsinki Institute for Information Technology, Finland..................................................................966 Tay, Yuan Sherng / National University of Singapore, Singapore . ..........................................................................345 Teufel, Stephanie / University of Fribourg, Switzerland . .........................................................................................466 Tjondronegoro, Dian / Queensland University of Technology, Australia ................................................................596 Tong, Carrison K. S. / Pamela Youde Nethersole Eastern Hospital, Hong Kong ....................................................533 Tran, Dai / Arcada Polytechnic, Finland . .................................................................................................................839 Tsagkaropoulos, Michail / University of Patras, Greece...................................................................................357, 796 Tsetsos, Vassileios / University of Athens, Greece .....................................................................................................856 Tuceryan, Mihran / Indiana University Purdue University Indianapolis, USA .......................................................207 Turel, Ofir / McMaster University, Canada ..............................................................................................................143 Tynan, Richard / University College Dublin, Ireland . ...........................................................................................1034 Usaola, Macario Polo / Universidad de Castilla-La Mancha, Spain . ........................................................................57 Venkataram, P. / Indian Institute of Science, India . .................................................................................................195 Vogel, Doug / City University of Hong Kong, Hong Kong ........................................................................................317 Wählisch, Matthias / FHTW Berlin, Germany .................................................................................................541, 589 Wang, Yiling / Monash University, Australia ....................................................................................................227, 832 Wang, JiaJia / University of Bradford, UK ...............................................................................................................455 Wang, Laura / Tongji University, China ...................................................................................................................669 Wang, Yiling / Monash University, Australia ....................................................................................................227, 832 Wang, Wenbo / Beijing University of Posts & Telecommunications, China . ...................................................921, 940 Wang, Yingjie / Beijing University of Posts & Telecommunications, China ............................................................921 Wickramasinghe, Nilmini / IIllinois Institute of Technology, USA ........................................................................1004 Wiedemann, Dietmar Georg / University of Augsburg, Germany ...........................................................................547 Willis, Robert / Lakehead University, Canada . ........................................................................................................143 Wong, Chin Chin / British Telecommunications (Asian Research Center), Malaysia .............................138, 487, 906 Wong, K. Daniel / Malaysia University of Science and Technology, Malaysia ................................................253, 912 Wong, King Yin / The Chinese University of Hong Kong, Hong Kong ....................................................................283 Wong, Eric T. T. / Hong Kong Polytechnic University, Hong Kong .........................................................................533 Wright, David / University of Ottawa, Canada ........................................................................................90, 816, 1038 Xavier, Rodrigo Nóbrega Rocha / Federal University of Campina Grande, Brazil . ............................................1011 Xi, Chen / University of Hong Kong, Hong Kong .......................................................................................................96 Xilouris, G. / N.C.S.R., Demokritos, Greece . ....................................................................................................758, 889 Yan, Lu / Åbo Akademi, Finland . ..............................................................................................................................492 Yan, Hong / City University of Hong Kong, Hong Kong & University of Sydney, Australia.....................................669 Yang, Cheng-Hong / National Kaohsiung University of Applied Sciences, Taiwan .........................................352, 616 Yang, Cheng Huei / National Kaohsiung Marine University, Taiwan...............................................................352, 616 Ye, Yang / Tongji University, China ...........................................................................................................................669 Yeh, Jyh-haw / Boise State University, USA .............................................................................................................302 Yim, Frederick Hong Kit / Drexel University, USA .................................................................................................283 Zervas, Evangelos / Tei-Athens, Greece ....................................................................................................................863 Zhang, Zuopeng (Justin) / Eastern New Mexico University, USA ...........................................................................520 Zhong, Yapin / Shandog Institute of Physical Education and Sport, China .............................................................302 Zhou, Jiaogen / Wuhan University, China . .................................................................................................84, 213, 789 Zhou, Shuigeng / Fudan University, China . ...............................................................................................84, 213, 789 Zhu, Fubao / Wuhan University, China .............................................................................................................213, 789 Zoi, S. / National Technical University of Athens, Greece .........................................................................................387
Contents by Volume volume I Academic Activities Based on Personal Networks Deployment / Vasileios S. Kaldanis, Charalampos Z. Patrikakis, and Vasileios E. Protonotarios .........................................................................................................................................1 Accessibility of Mobile Applications / Pankaj Kamthan ......................................................................................................9 Acoustic Data Communication with Mobile Devices / Victor I. Khashchanskiy and Andrei L. Kustov ............................. 15 Adaptive Transmission of Multimedia Data over UMTS / Antonios Alexiou, Dimitrios Antonellis, and Christos Bouras ............................................................................................................................................................ 20 Addressing the Credibility of Mobile Applications / Pankaj Kamthan................................................................................25 Adoption and Diffusion of M-Commerce / Ranjan B. Kini and Subir K. Bandyopadhyay ................................................32 Adoption of M-Commerce Devices by Consumers / Humphry Hung and Vincent Cho ..................................................... 38 Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce / Michele Ruta, Tommaso Di Noia, Eugenio Di Sciascio, Francesco Maria Donini, and Giacomo Piscitelli . ................................................................... 43 Anycast-Based Mobility / István Dudás, László Bokor, and Sándor Imre ..........................................................................51 Applications Suitability on PvC Environments / Andres Flores and Macario Polo Usaola .............................................. 57 Bio-Inspired Approach for the Next Generation of Cellular Systems, A / Mostafa El-Said ...............................................63 Brain Computer Interfacing / Diego Liberati ...................................................................................................................... 68 Bridging Together Mobile and Service Oriented Computing / Loreno Oliveira, Emerson Loureiro, Hyggo Almeida, and Angelo Perkusich ...................................................................................................................................................71 Browser-Less Surfing and Mobile Internet Access / Gregory John Fleet and Jeffery G. Reid ...........................................78 Building Web Services in P2P Networks / Jihong Guan, Shuigeng Zhou, and Jiaogen Zhou ............................................ 84 Business and Technology Issues in Wireless Networking / David Wright .......................................................................... 90 Business Strategies for Mobile Marketing / Indranil Bose and Chen Xi . ........................................................................... 96 Cache Invalidation in a Mobile Environment / Say Ying Lim . .......................................................................................... 102
Communicating Recommendations in a Service-Oriented Environment / Omar Khadeer Hussain, Elizabeth Chang, Farookh Khadeer Hussain, and Tharam S. Dillon . ...................................................................................................108 Content Personalization for Mobile Interfaces / Spiridoula Koukia, Maria Rigou, and Spiros Sirmakessis .................... 116 Content Transformation Techniques / Ioannis Antonellis, Christos Bouras, and Vassilis Poulopoulos ........................... 119 Context-Adaptive Mobile Systems / Christian Kaspar, Thomas Diekmann, and Svenja Hagenhoff . ..............................124 Context-Aware Mobile Geographic Information Systems / Slobodanka Djordjevic-Kajan, Dragan Stojanović, and Bratislav Predić ...................................................................................................................................................129 Context-Aware Systems / Chin Chin Wong and Simon Hoh ............................................................................................. 138 Contractual Obligations between Mobile Service Providers and Users / Robert Willis, Alexander Serenko, and Ofir Turel .............................................................................................................................................................143 Convergence Technology for Enabling Technologies / G. Sivaradje, I. Saravanan, and P. Dananjayan.......................... 149 Cooperative Caching in a Mobile Environment / Say Ying Lim ........................................................................................154 CORBA on Mobile Devices / Markus Aleksy, Axel Korthaus, and Martin Schader . ....................................................... 160 Cross-Layer RRM in Wireless Data Networks / Amoakoh Gyasi-Agyei . ......................................................................... 165 Data Caching in Mobile Ad-Hoc Networks / Narottam Chand, R. C. Joshi, and Manoj Misra .......................................172 Decision Analysis for Business to Adopt RFID / Koong Lin, Chad Lin, and Huei Leu . ..................................................178 Definitions, Key Characteristics, and Generations of Mobile Games / Eui Jun Jeong and Dan J. Kim . ......................... 185 Design Methodology for Mobile Information Systems / Zakaria Maamar and Qusay H. Mahmoud ..............................190 Distributed Approach for QoS Guarantee to Wireless Multimedia / Kumar S. Chetan, P. Venkataram, and Ranapratap Sircar ......................................................................................................................................................195 Distributed Computing in Wireless Sensor Networks / Hong Huang ............................................................................... 202 Distributed Heterogeneous Tracking for Augmented Reality / Mihran Tuceryan and Rajeev R. Raje ............................. 207 Distributed Web GIS / Jihong Guan, Shuigeng Zhou, Jiaogen Zhou, and Fubao Zhu ..................................................... 213 Dynamic Pricing Based on Net Cost for Mobile Content Services / Nopparat Srikhutkhao, and Sukumal Kitisin . ........220 Efficient and Scalable Group Key Management in Wireless Networks / Yiling Wang and Phu Dung Le ........................227 Efficient Replication Management Techniques for Mobile Databases / Ziyad Tariq Abdul-Mehdi, Ali Bin Mamat, Hamidah Ibrahim, and Mustafa Mat Dirs . ................................................................................................................233 Embedded Agents for Mobile Services / John F. Bradley, Conor Muldoon, Gregory M. P. O’Hare, and Michael J. O’Grady ....................................................................................................................................................243 Enabling Mobile Chat Using Bluetooth / Ádrian Lívio Vasconcelos Guedes, Jerônimo Silva Rocha, Hyggo Almeida, and Angelo Perkusich ......................................................................................................................249
Enabling Mobility in IPv6 Networks / Saaidal Razalli Bin Azzuhri and K. Daniel Wong ................................................ 253 Enabling Multimedia Applications in Memory-Limited Mobile Devices / Raul Fernandes Herbster, Hyggo Almeida, Angelo Perkusich, and Marcos Morais ...........................................................................................260 Enabling Technologies for Mobile Multimedia / Kevin Curran . ......................................................................................265 Enabling Technologies for Pervasive Computing / João Henrique Kleinschmidt and Walter da Cunha Borelli ............. 272 Extreme Programming for Mobile Applications / Pankaj Kartham . ................................................................................277 Factors Affecting Mobile Commerce and Level of Involvement / Frederick Hong Kit Yim, Alan ching Biu Tse, and King Yin Wong . ......................................................................................................................................................283 Game-Based Methodology for Collaborative Mobile Applications, A / Michael Massimi, Craig H. Ganoe, and John M. Carroll .......................................................................................................................................................... 291 Gender Difference in the Motivations of Mobile Internet Usage / Shintaro Okazaki .......................................................296 Handheld Computing and J2ME for Internet-Enabled Mobile Handheld Devices / Wen-Chen Hu, Jyh-haw Yeh, I-Lung Kao, and Yapin Zhong ....................................................................................................................................302 Infrastructural Perspective on U-Commerce, An / Stephen Keegan, Caroline Byrne, Peter O’Hare, and Gregory M. P. O’Hare ..................................................................................................................................................310 Integrating Pedagogy, Infrastructure, and Tools for Mobile Learning / David M. Kennedy and Doug Vogel ..................317 Intelligent Medium Access Control Protocol for WSN / Haroon Malik, Elhadi Shakshuki, and Mieso Kabeto Denko . ...................................................................................................................................................328 Intelligent User Preference Detection for Product Brokering / Sheng-Uei Guan . ............................................................334 Interactive Multimedia File Sharing Using Bluetooth / Danilo Freire de Souza Santos, José Luís do Nascimento, Hyggo Almeida, and Angelo Perkusich ...................................................................................................................... 341 Interactive Product Catalog for M-Commerce / Sheng-Uei Guan and Yuan Sherng Tay . ................................................ 345 Interactive Wireless Morse Code Learning System, An / Cheng-Huei Yang , Li Yeh Chuang, Cheng-Hong Yang, and Jun-Yang Chang, . .................................................................................................................................................352 Interworking Architectures of 3G and WLAN / Ilias Politis, Tasos Dagiuklas, Michail Tsagkaropoulos, and Stavros Kotsopoulos ............................................................................................................................................ 357 iPod as a Visitor’s Personal Guide / Keyurkumar J. Patel and Umesh Patel ....................................................................365 Keyword-Based Language for Mobile Phones Query Services / Ziyad Tariq Abdul-Mehdi and Hussein M. Aziz Basi .................................................................................................................................................. 369 Knowledge Representation in Semantic Mobile Applications / Pankaj Kamthan ............................................................375 Location-Based Multimedia Content Delivery System for Monitoring Purposes / Athanasios-Dimitrios Sotiriou and Panagiotis Kalliaras . .......................................................................................................................................... 381
Location-Based Multimedia Services for Tourists / P. Kalliaras, Athanasios-Dimitrios Sotiriou, P. Papageorgiou, and S. Zoi ......................................................................................................................................................................387 Location-Based Services / Péter Hegedüs, Mihály Orosz, Gábor Hosszú, and Ferenc Kovács .......................................393 M-Advertising / Michael Decker ....................................................................................................................................... 398 Man-Machine Interface with Applications in Mobile Robotic Systems / Milan Kvasnica ............................................... 403 M-Commerce Technology Perceptions on Technology Adoptions / Reychav Iris and Ehud Menipaz ............................. 413 M-Learning with Mobile Phones / Simon So . ...................................................................................................................419 Mobile Ad-Hoc Networks / Moh Lim Sim, Choong Ming Chin, and Chor Min Tan . ....................................................... 424 Mobile Agent Protection for M-Commerce / Sheng-Uei Guan.......................................................................................... 429 Mobile Agent-Based Discovery System / Rajeev R. Raje, Jayasree Gandhamaneni, Andrew M. Olson, and Barrett R. Bryant ................................................................................................................................................. 436 Mobile Business Applications / Cheon-Pyo Lee ............................................................................................................... 442 Mobile Cellular Traffic with the Effect of Outage Channels / Hussein M. Aziz Basi and M. B. Ramamurthy .................446 Mobile Commerce / JiaJia Wang and Pouwan Lei ........................................................................................................... 455 Mobile Commerce Adoption Barriers / Pruthikrai Mahatanankoon and Juan Garcia . ...................................................461 Mobile Computing and Commerce Framework, A / Stephanie Teufel, Patrick S. Merten, and Martin Steinert ............. 466 Mobile E-Commerce as a Strategic Imperative for the New Economy / Mahesh S. Raisinghani ....................................472 Mobile Enterprise Readiness and Transformation / Rahul C. Basole and William B. Rouse ............................................ 481 Mobile Entertainment / Chin Chin Wong and Pang Leang Hiew . ....................................................................................487 Mobile File-Sharing over P2P Networks / Lu Yan . ...........................................................................................................492 Mobile Gaming / Krassie Petrova .....................................................................................................................................497 Mobile Healthcare Communication Infrastructure Networks / Phillip Olla .....................................................................504 Mobile Hunters / Jörg Lonthoff .........................................................................................................................................510
volume II Mobile ICT / Dermott McMeel .......................................................................................................................................... 516 Mobile Knowledge Management / Zuopeng (Justin) Zhang and Sajjad M. Jasimuddin .................................................. 520 Mobile Learning / David Parsons ..................................................................................................................................... 525 Mobile Learning Environments / Paul Crowther and Martin Beer ..................................................................................528
Mobile Medical Image Viewing Using 3G Wireless Network / Carrison K. S. Tong and Eric T. T. Wong ...................... 533 Mobile Multicast / Thomas C. Schmidt and Matthias Wählisch ....................................................................................... 541 Mobile Payment and the Charging of Mobile Services / Key Pousttchi and Dietmar Georg Wiedemann .......................547 Mobile Phone Gambling / Mark Griffiths . ........................................................................................................................553 Mobile Phone Privacy Issues / Călin Gurău ..................................................................................................................... 557 Mobile Phone Texting in Hong Kong / Adams Bodomo . ..................................................................................................562 Mobile Phones for People with Disabilities / Hend S. Al-Khalifa and AbdulMalik S. Al-Salman ....................................569 Mobile Processes and Mobile Channels / Kevin Chalmers, Imed Romdhami, and Jon Kerridge . ................................... 576 Mobile Public Key Infrastructures / Ioannis Chochliouros, George K. Lalopoulos, Stergios P. Chochliouros, and Anastasia S. Spiliopoulou .................................................................................................................................... 581 Mobile Serverless Video Communication / Hans L. Cycon, Thomas C. Schmidt, and Matthias Wählisch ...................... 589 Mobile Sports Video with Total Users Control / Dian Tjondronegoro . ............................................................................ 596 Mobile Telephony in Sub-Saharan Africa / Princely Ifinedo . ...........................................................................................605 Mobile Television / Frank Hartung, Markus Kampmann, Uwe Horn, and Jan Kritzner ................................................. 611 Mobile Text Messaging Interface for Persons with Physical Disabilities / Cheng-Huei Yang, Li-Yeh Chuang, Cheng-Hong Yang, and Jun-Yang Chang .....................................................................................................................616 Mobile Users in Smart Spaces / Loreno Oliveira, Hyggo Almeida, and Angelo Perkusich ..............................................621 Mobile Video Transcoding Approaches and Challenges / Ashraf M. A. Ahmad ...............................................................627 Mobile Virtual Communities / Christo El Morr ................................................................................................................632 Mobile-Based Advertising in Japan / Shintaro Okazaki ....................................................................................................635 Mobile-Based Research Methods / Shintaro Okazaki, Akihisa Katsukura, and Mamoru Nishiyama . ............................. 639 Mobility and Multimodal User Interfaces / Christopher J. Pavlovski and Stella Mitchell ............................................... 644 Modular Sensory System for Robotics and Human-Machine Interaction Based on Optoelectronic Components / Milan Kvasnica .............................................................................................................................................................651 Monitoring and Tracking Moving Objects in Mobile Environments / Dragan Stojanovic, Slobodanka Djordjevic-Kajan, Apostolos N. Papadopoulos, and Alexandros Nanopoulos ......................................660 Multilingual SMS / Mohammad Shirali-Shahreza ............................................................................................................666 Multimedia Contents for Mobile Entertainment / Hong Yan, Laura Wang, and Yang Ye . ................................................669 Multimodality in Mobile Applications and Services / Maria Chiara Caschera, Fernando Ferri, and Patrizia Grifoni . ...........................................................................................................................................................675
Multi-User OFDM in Mobile Multimedia Network / Ibrahim Al Kattan and Habeebur Rahman Maricar . ................... 682 Mutual Biometric Authentication / Mostafa El-Said ......................................................................................................... 688 New Transaction Management Model / Ziyad Tariq Abdul-Mehdi, Ali Bin Mamat, Hamidah Ibrahim, and Mustafa M. Dirs . ........................................................................................................................................................693 Next-Generation Mobile Technologies / Chor Min Tan, Choong Ming Chin, and Moh Lim Sim . ...................................700 NFC-Capable Mobile Devices for Mobile Payment Services / Stamatis Karnouskos ...................................................... 706 Notification Services for Mobile Scenarios / Michael Decker .......................................................................................... 711 Ontology-Based Approach for Mobile Agent’s Context-Awareness, An / Nejla Amara-Hachmi and Amal El Fallah-Seghrouchni ......................................................................................................................................717 Optimal Timer for Push to Talk Controller, An / Muhammad Tanvir Alam ...................................................................... 724 Optimal Utilisation of Future Wireless Resources / Choong Ming Chin, Chor Min Tan, and Moh Lim Sim ................... 729 P2P Models and Complexity in MANETs / Boon-Chong Seet, Chiew-Tong Lau, and Wen-Jing Hsu . ............................734 Partial Global Indexing for Location-Dependent Query Processing / James W. Jayaputera ............................................739 Patterns for Mobile Applications / Markus Aleksy and Martin Schader ........................................................................... 744 Peer-to-Peer Cooperative Caching in Mobile Environments / Chi-Yin Chow, Hong Va Leong, and Alvin T. S. Chan ..... 749 Pen-Based Mobile Computing / Bernie Garret .................................................................................................................754 Perceived Quality Evaluation for Multimedia Services / H. Koumaras, E. Pallis, G. Xilouris, A. Kourtis, and D. Martakos .........................................................................................................................................................758 Pest Activity Prognosis in the Rice Field / Nureize Arbaiy, Azizul Azhar Ramli, Zurinah Suradi, and Mustafa Mat Deris . ....................................................................................................................................................763 Positioning Technologies for Mobile Computing / Michael J. O’Grady and Gregory O’Hare . ...................................... 769 Privacy Concerns for Indoor Location-Based Services / Leonardo Galicia Jiménez, and J. Antonio García-Macías......773 Protocol Analysis for the 3G IP Multimedia Subsystem / Muhammad Tanvir Alam ........................................................ 778 Protocol Replacement Proxy for 2.5 and 3G Mobile Internet / Victor Khashchanskiy, Andrei Kustov, and Jia Lang . ....785 Providing Location-Based Services under Web Services Framework / Jihong Guan, Shuigeng Zhou, Jiaogen Zhou, and Fubao Zhu . ..........................................................................................................................................................789 Provisioning of Multimedia Applications across Heterogeneous All-IP Networks / Michail Tsagkaropoulos, Ilias Politis, Tasos Dagiuklas, and Stavros Kotsopoulos ...........................................................................................796 QoS Routing Framework on Bluetooth Networking, A / Chao Liu, Bo Huang, and Takaaki Baba .................................804 Radio Resource Management in Convergence Technologies / G. Sivaradje, I. Saravanan, and P. Dananjayan ............. 810
RFID and Wireless Personal Area Networks for Supply Chain Management / David Wright . ........................................ 816 Scatternet Structure for Improving Routing and Communication Performance / Bo Huang, Chao Liu, and Takaaki Baba............................................................................................................................................................... 820 Secure Agent Data Protection for E-Commerce Applications / Sheng-Uei Guan .............................................................826 Secure Group Communications in Wireless Networks / Yiling Wang and Phu Dung Le ..................................................832 Security Architectures of Mobile Computing / Kaj Grahn, Göran Pulkkis, Jonny Karlsson, and Dai Tran . .................. 839 Semantic Caching in a Mobile Environment / Say Ying Lim . ...........................................................................................849 Semantic Enrichment of Location-Based Services / Vassileios Tsetsos, Christos Anagnostopoulos, and Stathes Hadjiefthymiades ........................................................................................................................................... 856 Sensor Data Fusion for Location Awareness / Odysseas Sekkas, Stathes Hadjiefthymiades, and Evangelos Zervas .......863 Service Delivery Platforms in Mobile Convergence / Christopher J. Pavlovski and Laurence Plant . ............................ 870 Service Provision for Pervasive Computing Environments / Emerson Loureiro, Frederico Bublitz, Loreno Oliveira, Nadia Barbosa, Angelo Perkusich, Hyggo Almeida, and Glauber Ferreira............................................................... 877 Short Message Service (SMS) as an Advertising Medium / Shintaro Okazaki .................................................................885 Shot Boundary Detection Techniques for Video Sequences / H. Koumaras, G. Xilouris, E. Pallis, G. Gardikis, and A. Kourtis ............................................................................................................................................................. 889 Smartphone Acceptance among Sales Drivers / Jengchung V. Chen.................................................................................. 894 SMS-Based Mobile Learning / Krassie Petrova ...............................................................................................................899 Snapshot Assessment of Asia Pacific BWA Business Scenario / Chin Chin Wong, Chor Min Tan, and Pang Leang Hiew........................................................................................................................................................906 Software Platforms for Mobile Programming / Khoo Wei Ju and K. Daniel Wong ..........................................................912 Standard-Based Wireless Mesh Networks / Mugen Peng, Yingjie Wang, and Wenbo Wang..............................................921 Taxonomies, Applications, and Trends of Mobile Games / Eui Jun Jeong and Dan J. Kim . ........................................... 928 Technology Intervention Perspective of Mobile Marketing, A / Dennis Lee and Ralf Muhlberger . ................................ 933 3G Commercial Deployment / Mugen Peng, Shuping Chen, and Wenbo Wang ............................................................... 940 Transaction Management in Mobile Databases / Ziyad Tariq Abdul-Mehdi, Ali Bin Mamat, Hamidah Ibrahim, and Mustafa M. Dirs...........................................................................................................................................................947 Ubiquitous and Pervasive Application Design / M. Bakhouya and J. Gaber ....................................................................954 “Umbrella” Distributed-Hash Table Protocol for Content Distribution, The / Athanasios-Dimitrios Sotiriou and Panagiotis Kalliaras . ..........................................................................................................................................960
Understanding Multi-Layer Mobility / Sasu Tarkoma and Jouni Korhonen . ................................................................... 966 Using Mobile Devices for Electronic Commerce / Raul Fernandes Herbster, Hyggo Almeida, and Angelo Perkusich ........................................................................................................................................................974 Using Mobile Devices to Manage Traffic Infractions / Stefânia Marques, Sabrina Souto, Miguel Queiroga, Hyggo Almeida, and Angelo Perkusich ........................................................................................................................978 Using Service Proxies for Content Provisioning / Panagiotis Kalliaras and Anthanasios-Dimitrios Sotiriou ................ 981 Verifying Mobile Agent Design Patterns with RPOO / Elthon Allex da Silva Oliveira, Emerson Ferreira de Araújo Lima, and Jorge César Abrantes de Figueiredo ..........................................................987 Virtualization and Mobility in Client and Server Environments / Eduardo Correia . .......................................................996 Voice Recognition Intelligent Agents Technology / Călin Gurău ..................................................................................... 999 Wi-INET Model for Achieving M-Health Success, The / Nilmini Wickramasinghe and Steve Goldberg ...................... 1004 Wireless Access Control System Using Bluetooth / Juliano Rodrigues Fernandes de Oliveira, Rodrigo Nóbrega Rocha Xavier, Yuri de Carvalho Gomes, Hyggo Almeida, and Angelo Perkusich ...................... 1011 Wireless Client Server Application Model Using Limited Key Generation Technique / Rohit Singh, Dhilak Damodaran, and Phu Dung Le......................................................................................................................1015 Wireless Network Security / Kevin Curran and Elaine Smyth ........................................................................................ 1022 Wireless Security / Meletis Belsis, Alkis Simitsis, and Stefanos Gritzalis........................................................................ 1028 Wireless Sensor Networks / Antonio G. Ruzzelli, Richard Tynan, Michael O’Grady, and Gregory O’Hare ................. 1034 Wireless Technologies for Mobile Computing and Commerce / David Wright............................................................... 1038 Workflow Management Systems in MANETs / Fabio De Rosa, Massimiliano de Leoni, and Massimo Mecella .........1043 XML-Based Languages for Multimodality in Mobile Environments / Danilo Avola, Maria Chiara Caschera, Fernando Ferri, and Patrizia Grifoni ......................................................................................................................1050
Contents by Topic 3G Interworking Architectures of 3G and WLAN / Ilias Politis, Tasos Dagiuklas, Michail Tsagkaropoulos, and Stavros Kotsopoulos ..............................................................................................................................................................357 Protocol Analysis for the 3G IP Multimedia Subsystem / Muhammad Alam . ..................................................................778 Protocol Replacement Proxy for 2.5 and 3G Mobile Internet / Victor Khashchanski, Andrei Kustov, and Jia Lang ....... 785 Three 3G Commercial Deployment / Mugen Peng, Shuping Chen, and Wenbo Wang ..................................................... 940
Adhoc Network Data Caching in Mobile Ad-Hoc Networks / Narottam Chand, R.C. Joshi, and Manoj Misra ........................................172 Mobile Ad-Hoc Networks / Moh Lim Sim, Choong Ming Chin, and Chor Min Tan . ....................................................... 424 Workflow Management Systems in MANETs / Fabio De Rosa, Massimiliano de Leoni, and Massimo Mecella .........1043
Converging Technology Acoustic Data Communication with Mobile Devices / Victor I. Khachtchanski and Andrei Kustov . ................................15 Applications Suitability on PvC Environments / Andres Pablo Flores and Macario Polo Usaola .................................... 57 Bio-inspired Approach for Cellular Systems, A / Mostafa El-Said .....................................................................................63 Convergence Technology for Enabling Technologies / G. Sivaradje, I. Saravanan, and P. Dananjayan . ....................... 149 Decision Analysis for Business to Adopt RFID / Koong Lin, Chad Lin, and Huei Leu . ..................................................178 Distributed Web GIS / Jihong Guan, Shuigeng Zhou, Jiaogen Zhou, and Fubao Zhu ..................................................... 213 Enabling Technologies for Pervasive Computing / João H. Kleinschmidt and Walter da Cunha Borelli ........................ 272 Man-Machine Interface with Applications in Mobile Robotic Systems / Milan Kvasnica ............................................... 403 Mobile Users in Smart Spaces / Loreno Oliveira, Hyggo Almeida, and Angelo Perkusich ..............................................621 Modular Sensory System for Robotics and Human-Machine Interaction / Milan Kvasnica ............................................ 651
Mutual Biometric Authentication / Mostafa El-Said.......................................................................................................... 688 Next-Generation Mobile Technologies / Chor Min Tan, Choong Ming Chin, and Moh Lim Sim . ...................................700 Optimal Timer for Push to Talk Controller, An / Muhammad Tanvir Alam ...................................................................... 724 Pen-Based Mobile Computing / Bernie Garrett ................................................................................................................ 754 Pest Activity Prognosis in the Rice Field / Nureize Arbaiy, Azizul Azhar Ramli, Zurinah Suradi, and Mustafa Mat Deris ...................................................................................................................................................763 Using Mobile Devices to Manage Traffic Infractions / Stefânia Daisy Canuto Marques, Sabrina de Figueirêdo Souto, Miguel Queiroga Filho, Hyggo Almeida, and Angelo Perkusich ............................ 978
Human Factor Academic Activities Based on Personal Networks Deployment / Vasileios S. Kaldanis, Charalampos Patrikakis, and Vasileios Protonotarios ................................................................................................................................................1 Adoption and Diffusion of M-Commerce / Ranjan Kini and Subir Bandyopadhyay ......................................................... 32 Adoption of M-commerce Devices by Consumers / Humphry Hung and Vincent Cho ...................................................... 38 Browser-Less Surfing and Mobile Internet Access / Gregory J. Fleet and Jeffery G. Reid ................................................78 Gender Difference in the Motivations of Mobile Internet Usage / Shintaro Okazaki .......................................................296 M-Commerce Technology Perceptions on Technology Adoptions / Reychav Iris and Ehud Menipaz ............................. 413 Mobile Commerce Adoption Barriers / Pruthikrai Mahatanankoon and Juan Garcia . ...................................................461 Mobile Enterprise Readiness and Transformation / Rahul C. Basole and William B Rouse ............................................. 481 Mobile ICT / Dermot McMeel ...........................................................................................................................................516 Mobile Knowledge Management / Zuopeng Zhang and Sajjad M. Jasimuddin ...............................................................520 Mobile Virtual Communities / Christo El Morr ................................................................................................................632
Location and Context Awareness Context-Adaptive Mobile Systems / Christian Kaspar, Thomas Diekman, and Svenja Hagenhoff . ................................124 Context-Aware Mobile Geographic Information Systems / Slobodanka Djordjevic – Kajan, Dragan Stojanovic, and Bratislav Predic . ......................................................................................................................................................129 Context-Aware Systems / Chin Chin Wong and Simon Hoh ............................................................................................. 138 iPod as a Visitor’s Personal Guide / Keyurkumar Patel and Umesh Patel . ......................................................................365 Location-Based Multimedia Content Delivery System for Monitoring Purposes / Athanasios-Dimitrios Sotiriou and Panagiotis Kalliaras ......................................................................................................................................... 381
Location-Based Multimedia Services for Tourists / P. Kalliaras, A. D. Sotiriou, P. Papageorgiou, and S. Zoi ...............387 Location-Based Services / Péter Hegedüs, Mihály Orosz, Gábor Hosszú, and Ferenec Kovács .....................................393 Monitoring and Tracking Moving Objects in Mobile Environments / Dragan Stojanovic, Slobodanka Djordjevic-Kajan, Apostolos N. Papadopoulos, and Alexandros Nanopoulos ....................................660 Notification Services for Mobile Scenarios / Michael Decker .......................................................................................... 711 Ontology-Based Approach for Mobile Agents Context-Awareness, An / Nejla Amara-Hachmi and Amal El Fallah Seghrouchni .................................................................................................................................... 717 Partial Global Indexing for Location-Dependent Query Processing / James Jayaputera .................................................739 Positioning Techologies for Mobile Computing / Michael O’Grady and Gregory O’Hare .............................................769 Privacy Concerns for Indoor Location-based Services / Leonardo Galicia Jimenez and J. Antonio García-Macías ......773 Providing Location-Based Services under Web Services Framework / Jihong Guan, Shuigeng Zhou, Jiaogen Zhou, and Fubao Zhu .........................................................................................................................................................789 Semantic Enrichment of Location-Based Services / Vassileios Tsetsos, Christos Anagnostopoulos, and Stathes Hadjiefthymiades ......................................................................................................................................... 856 Sensor Data Fusion for Location Awareness / Odysseas Sekkas, Stathes Hadjiefthymiades, and Evangelos Zervas .....................................................................................................................................................863
M-Business and M-Commerce Addressing the Credibility of Mobile Applications / Pankaj Kartham ...............................................................................25 Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce / Michele Ruta, Tommaso Di Noia, Eugenio Di Sciascio, Giacomo Piscitelli, Francesco Maria Donini .........................................................................43 Business and Technology Issues in Wireless Networking / David Wright .......................................................................... 90 Business Strategies for Mobile Marketing / Indranil Bose and Chen Xi . ........................................................................... 96 Contractual Obligations between Mobile Service Providers and Users / Robert Willis, Alexander Serenko, and Ofir Turel ..................................................................................................................................................................143 Dynamic Pricing Based on Net Cost for Mobile Content Services / Nopparat Srikhutkhao and Sukumal Kitisin . .........220 Factors Affecting Mobile Commerce and Level of Involvement / Frederick Hong Kit Yim, King-Yin Wong, and Alan Ching-bin Tse ...................................................................................................................................................283 Infrastructural Perspective on U-Commerce, An / Stephen Keegan, Caroline Byrne, Peter O’Hare, and Gregory O’Hare ....................................................................................................................................................... 310 Intelligent User Preference Detection for Product Brokering / Sheng-Uei Guan . ............................................................334 Interactive Product Catalog for M-Commerce / Sheng-Uei Guan and Yuan Sherng Tay . ................................................ 345 M-Advertising / Michael Decker ....................................................................................................................................... 398
Mobile Agent Protection for M-Commerce / Sheng-Uei Guan ......................................................................................... 429 Mobile Business Applications / Cheon-Pyo Lee ............................................................................................................... 442 Mobile Commerce / Jia Jia Wang and Pouwan Lei .......................................................................................................... 455 Mobile Computing and Commerce / Stephanie Teufel, Patrick S. Merten, and Martin Steinert ......................................466 Mobile E-Commerce as a Strategic Imperative for New Economy / Mahesh S. Raisinghani ..........................................472 Mobile Payment and the Charging of Mobile Services / Key Pousttchi and Dietmar Georg Wiedemann .......................547 Mobile-Based Advertising in Japan / Shintaro Okazaki ....................................................................................................635 Mobile-Based Research Methods / Shintaro Okazaki, Akihisa Katsukura, and Mamoru Nishiyama . ............................. 639 NFC-Capable Mobile Devices for Mobile Payment Services / Stamatis Karnouskos ...................................................... 706 RFID and Wireless Personal Area Networks for Supply Chain Management / David Wright . ........................................ 816 Secure Agent Data Protection for E-Commerce Applications / Sheng-Uei Guan .............................................................826 Snapshot Assessment of Asia Pacific BWA Business Scenario / Chin Chin Wong, Chor Min Tan, and Pang Leang Hiew .....................................................................................................................................................906 Technology Intervention Perspective of Mobile Marketing, A / Dennis Lee and Ralf Muhlberger . ................................ 933 Using Mobile Devices for Electronic Commerce / Raul Fernandes Herbster, Hyggo Almeida, and Angelo Perkusich . ....................................................................................................................................................974 Wireless Technologies for Mobile Computing and Commerce / David Wright .............................................................. 1038
M-Entertainment Mobile Hunters / Jörg Lonthoff .........................................................................................................................................510 Short Message Service (SMS) as an Advertising Medium / Shintaro Okazaki .................................................................885 Mobile Phone Gambling / Mark Griffiths . ........................................................................................................................553 Multimedia Contents for Mobile Entertainment / Hong Yan, Laura Wang, and Yang Ye . ................................................669 Mobile Entertainment / Chin Chin Wong and Pang Leang Hiew . ....................................................................................487 Game-Based Methodology for Collaborative Mobile Applications, A / Michael Massimi, Craig Ganoe, and John M. Carroll ........................................................................................................................................................ 291 Mobile Television / Frank Hartung, Markus Kampmann, Uwe Horn, and Jan Kritzner ................................................. 611 Mobile Gaming / Krassie Petrova .....................................................................................................................................497 Definitions, Key Characteristics, and Generations of Mobile Games / Eui Jun Jeong and Dan J. Kim . ......................... 185 Taxonomies, Applications, and Trends of Mobile Games / Eui Jun Jeong and Dan J. Kim . ........................................... 928
M-Health Mobile Medical Image Viewing Using 3G Wireless Network / Carrison KS Tong and Eric TT Wong . .......................... 533 Mobile Healthcare Communication Infrastructure Networks / Phillip Olla .....................................................................504 Wi-INET Model for Achieving M-Health Success, The / Nilmini Wichramasinghe and Steve Goldberg . .................... 1004
M-Learning Integrating Pedagogy, Infrastructure, and Tools for Mobile Learning / David M. Kennedy and Doug Vogel ..................317 Interactive Wireless Morse Code Learning System, An / Cheng-Hong Yang, Li Yeh Chuang, Cheng-Huei Yang, and Jun-Yang Chang .......................................................................................................................................................352 M-Learning with Mobile Phones / Simon So . ...................................................................................................................419 Mobile Learning / David Parsons ..................................................................................................................................... 525 Mobile Learning Environments / Paul Crowther and Martin Beer ..................................................................................528 SMS-Based Mobile Learning / Krassie Petrova ...............................................................................................................899
Mobile Multimedia Adaptive Transmission of Multimedia Data over UMTS / Antonios Alexiou, Dimitrios Antonellis, and Christos J. Bouras ...................................................................................................................................................... 20 Enabling Multimedia Applications in Memory-Limited Mobile Devices / Raul Fernandes Herbster, Hyggo Almeida, Angelo Perkusich, and Marcos Morais .................................................................................................................... 260 Enabling Technologies for Mobile Multimedia / Kevin Curran . ......................................................................................265 Interactive Multimedia File Sharing Using Bluetooth / Danilo Freire de Santos, José Luís do Nascimento, Hyggo Almeida, and Angelo Perkusich . .................................................................................................................. 341 Mobile Serverless Video Communication / Hans L. Cycon, Thomas C. Schmidt, and Matthias Wählisch ...................... 589 Mobile Sports Video with Total Users Control / Dian Tjondronegoro . ............................................................................ 596 Mobile Video Transcoding Approaches and Challenges / Ashraf M. A. Ahmad ...............................................................627 Multi-User OFDM in Mobile Multimedia Network / Ibrahim Al Kattan and Habeebur Rahman Maricar . ................... 682 Perceived Quality Evaluation for Multimedia Services / H. Kourmaras, E. Pallis, G. Xilouris, A. Kourtis, and D. Martakos . ............................................................................................................................................................758 Provisioning of Multimedia Applications across Heterogeneous All-IP Networks / Michail Tsagkaropoulos, Ilias Politis, Tasos Dagiuklas, and Stavros Kotsopoulos .........................................................................................796 Radio Resource Management in Convergence Technologies / G. Sivaradje, I. Saravanan, and P. Dananjayan ............. 810
Shot Boundary Detection Techniques for Video Sequences / H. Kourmaras, G. Xilouris, E. Pallis, G. Gardikis, and A. Kourtis ................................................................................................................................................................. 889
Mobile Phone Enabling Mobile Chat Using Bluetooth / Ádrian Lívio Vasconcelos Guedes, Jerônimo Silva Rocha, Hyggo Almeida, and Angelo Perkusich . .............................................................................................................................................249 Keyword-Based Language for Mobile Phones Query Services / Ziyad Tariq Abdul-Mehdi, and Hussein M. Aziz Basi ................................................................................................................................................ 369 Mobile Phone Privacy Issues / Călin Gurău ..................................................................................................................... 557 Mobile Phone Texting in Hong Kong / Adams Bodomo . ..................................................................................................562 Mobile Phones for People with Disabilities / Hend Al-Khalifa and AbdulMalik S. Al-Salman ........................................569 Mobile Telephony in Sub-Saharan Africa / P. Ifinedo .......................................................................................................605 Mobile Text Messaging Interface for Persons with Physical Disabilities / Cheng-Hong Yan . ......................................... 616 Multilingual SMS / Mohammad Shirali-Shahreza ............................................................................................................666 Smartphone Acceptance among Sales Drivers / Jengchung V. Chen ................................................................................ 894 Voice Recognition Intelligent Agents Technology / Călin Gurău ..................................................................................... 999
Mobile Software Engineering Accessibility of Mobile Applications / Pankaj Kartham . ..................................................................................................... 9 Brain Computer Interfacing / Diego Liberati ...................................................................................................................... 68 Cache Invalidation in a Mobile Environment / Say Ying Lim . .......................................................................................... 102 Content Personalization for Mobile Interfaces / Spiridoula Koukia, Maria Rigou, and Spiros Sirmakessis .................... 116 Content Transformation Techniques / Ioannis Antonellis, Vassilis Poulopoulos, and Christos Bouras ........................... 119 Cooperative Caching in a Mobile Environment / Say Ying Lim ........................................................................................154 CORBA on Mobile Devices / Markus Aleksy, Axel Korthaus, and Martin Schader . ....................................................... 160 Design Methodology for Mobile Information Systems / Zakaria Maamar and Qusay H. Mahmous .............................. 190 Distributed Heterogeneous Tracking for Augmented Reality / Mihran Tuceryan and Rajeev Raje ................................. 207 Efficient Replication Management Techniques for Mobile Databases / Ziyad Tariq Abdul-Mehdi, Ali Bin Mamat, Hamidah Ibrahim, and Mustafa Mat Dirs ............................................................................................................... 233 Extreme Programming for Mobile Applications / Pankaj Kartham . ................................................................................277
Handheld Computing and J2ME for Internet-Enabled Mobile Handheld Devices / Wen-Chen Hu, Jyh-haw Yeh, I-Lung Kao, and Yapin Zhong .................................................................................................................................. 302 Knowledge Representation in Semantic Mobile Applications / Pankaj Kamthan ............................................................375 Mobile Agent-Based Discovery System / Rajeev R. Raje, Jayasree Gandhamaneni, Andrew Olson, and Barrett Bryant . ......................................................................................................................................................... 436 Mobile Processes and Mobile Channels / Kevin Chalmers, Imed Romdhani, and Jon Kerridge ..................................... 576 Mobility and Multimodal User Interfaces / Christopher J. Pavlovski and Stella Mitchell ............................................... 644 Multimodality in Mobile Applications and Services / Maria Chiara Caschera, Fernando Ferri, and Patrizia Grifoni . ....................................................................................................................................................675 New Transaction Management Model / Ziyad Tariq Abdul Mehdi, Ali Bin Mamat, Hamidah Ibrahim, and Mustafa Mat Dirs .....................................................................................................................................................693 Patterns for Mobile Applications / Markus Aleksy and Martin Schader ........................................................................... 744 Semantic Caching in a Mobile Environment / Say Ying Lim . ...........................................................................................849 Software Platforms for Mobile Programming / Khoo Wei Ju and K. Daniel Wong ..........................................................912 Transaction Management in Mobile Databases / Ziyad Tariq Abdul-Mehdi, Hamidah Ibrahim, Mustafa Mat Dirs, and Ali Bin Mamat ...................................................................................................................................................947 Ubiquitous and Pervasive Application Design / Mohamed Bakhouya and J. Gaber ........................................................954 Umbrella Distributed Hash Table Protocol for Content Distribution, The / Athanasios-Dimitrios Sotiriou, and Panagiotis Kalliaras ................................................................................................................................................ 960 Understanding Multi-Layer Mobility / Jouni Korhonen and Sasu A .O. Tarkoma . ..........................................................966 Verifying Mobile Agent Design Patterns with RPOO / Elthon Allex da Silva Oliveira, Emerson Ferreira de Araújo Lima, and Jorge C. A. de Figueiredo . ....................................................................... 987 Virtualization and Mobility in Client and Server Environments / Eduardo Correia . .......................................................996 XML-Based Languages for Multimodality in Mobile Environments / Danilo Avola, Maria Chiara Caschera, Fernando Ferri, and Patrizia Grifoni ....................................................................................................................1050
P2P Building Web Services in P2P Networks / Shuigeng Zhou, Jiaogen Zhou, and Jihong Guan ............................................ 84 Mobile File-sharing Over P2P Networks / Lu Yan ............................................................................................................492 P2P Models and Complexity MANETs / Boon Chong Seet, Chiew-Tong Lau, and Wen-Jing Hsu ..................................734 Peer-to-Peer Cooperative Caching in Mobile Environments / Chi-Yin Chow, Hong Va Leong, and Alvin T.S. Chan ...... 749
Security Mobile Public Key Infrastructures / Ioannis Chochliouros, George K. Lalopoulos, Stergios P. Chochliouros, and Anastasia S. Spiliopoulou . ....................................................................................................................................... 581 Security Architectures of Mobile Computing / Kaj Grahn, Göran Pulkkis, Jonny Karlsson, and Dai Tran . .................. 839 Wireless Network Security / Kevin Curran and Elaine Smyth ........................................................................................ 1022 Wireless Security / Meletis Belsis, Alkis Simitsis, and Stefanos Gritzalis ....................................................................... 1028
Sensor Network Wireless Sensor Networks / Antonio Ruzzelli, Richard Tynan, Michael O’Grady, and Gregory O’Hare ...................... 1034
Service Computing Bridging Together Mobile and Service Oriented Computing / Loreno Oliveira, Emerson Loureiro, Hyggo Almeida, and Angelo Perkusich . ...............................................................................................................................................71 Communicating Recommendations in a Service Oriented Environment / Omar Khadeer Hussain, Elizabeth Chang, Farookh Khadeer Hussain, and Tharam S. Dillon ..................................................................................................108 Embedded Agents for Mobile Services / John F. Bradley, Conor Muldoon, Gregory O’Hare, and Michael O’Grady .....................................................................................................................................................243 Service Delivery Platforms in Mobile Convergence / Christopher Pavlovski and Laurence Plant . ................................ 870 Service Provision for Pervasive Computing Environments / Emerson Loureiro, Frederico Bublitz, Loreno Oliveira, Nadia Barbosa, Hyggo Almeida, Glauber Ferreira, and Angelo Perkusich............................................................. 877 Using Service Proxies for Content Provisioning / P. Kalliaras and A. D. Sotiriou . ......................................................... 981
Wireless Networking Anycast-Based Mobility / István Dudás, László Bokor, and Sándor Imre ..........................................................................51 Cross-Layer RRM in Wireless Data Networks / Amoakoh Gyasi-Agyei . ......................................................................... 165 Distributed Approach for QoS Guarantee to Wireless Multimedia / Kumar S Chetan, P. Venkataram, and Ranapratap Sircar ....................................................................................................................................................195 Distributed Computing in Wireless Sensor Networks / Hong Huang ............................................................................... 202 Efficient and Scalable Group Key Management in Wireless Networks / Yiling Wang and Phu Dung Le ........................227 Enabling Mobility in IPv6 Networks / K. Daniel Wong and Saaidal Razalli Bin Azzuhri ................................................ 253 Intelligent Medium Access Control Protocol for WSN / Haroon Malik, Elhadi Shakshuki, and Mieso Denko ...............328 Mobile Cellular Traffic with the Effect of Outage Channels / Hussein M. Aziz Basi and M. B. Ramamurthy .................446 Mobile Multicast / Thomas C. Schmidt and Matthias Wählisch ....................................................................................... 541
Optimal Utilisation of Future Wireless Resources / Choong Ming Chin, Chor Min Tan, and Moh Lim Sim ................... 729 QoS Routing Framework on Bluetooth Networking, A / Chao Liu, Bo Huang, and Takaaki Baba .................................804 Scatternet Structure for Improving Routing and Communication Performance / Bo Huang, Chao Liu, and Takaaki Baba ............................................................................................................................................................ 820 Secure Group Communications in Wireless Networks / Yiling Wang and Phu Dung Le ..................................................832 Standard-Based Wireless Mesh Networks / Mugen Peng, Yingjie Wang, and Wenbo Wang . ...........................................921 Wireless Access Control System Using Bluetooth / Juliano Rodrigues Fernandes de Oliveira, Rodrigo Nóbrega Rocha Xavier, Luiz Paulo de Assis Barbosa, Yuri de Carvalho Gomes, Hyggo Almeida, and Angelo Perkusich . ........................................................................................................................................... 1011 Wireless Client Server Application Model Using Limited Key Generation Technique / Rohit Singh, Dhilak Domodaran, and Phu Dung Le ..................................................................................................................1015
xxvii
Foreword Let us borrow this quote from the British humorist and cartoonist Ashleigh Brilliant to summarize the role of mobility in the development of the information society: “Unless you move, the place where you are is the place where you will always be.” In more serious terms, it is fundamental to recognize that today’s economic and societal progress is primarily dependent on the technological ability to sustain and facilitate the mobility of persons, physical goods (let us not forget, for instance, that the probably most critical component of global commerce today is deep sea shipping) and digital information (data and programs). Recent years have witnessed a rapid growth of interest in mobile computing and communications. Indicators are the rapidly increasing penetration of the cellular phone market in Europe, and the mobile computing market is growing nearly twice as fast as the desktop market. In addition, technological advancements have significantly enhanced the usability of mobile communication and computer devices. From the first CT1 cordless telephones to today’s Iridium mobile phones and laptops/PDAs with wireless Internet connection, mobile tools and utilities have made the life of many people at work and at home much easier and more comfortable. As a result, mobility and wireless connectivity are expected to play a dominant role in the future in all branches of economy. This is also motivated by the large number of potential users (a U.S. study reports of one in six workers spending at least 20 percent of their time away from their primary workplace, similar trends are observed in Europe). The addition of mobility to data communications systems has not only the potential to put the vision of “being always on” into practice;- but has also enabled new generation of services, for example, location-based services. Mobile commerce leveraging the mobile Web and mobile multimedia is precisely the ability to deploy and utilize modern technologies for the design, development and deployment of a content rich, user and business friendly, integrated network of autonomous, mobile agents (here “agent” is to be taken in the sense of persons, goods and digital information). I am delighted to write the foreword to this encyclopedia, as its scope, content and coverage provides a descriptive, analytical, and comprehensive assessment of factors, trends, and issues in the ever-changing field of mobile computing and commerce. This authoritative research-based publication also offers in-depth explanations of mobile solutions and their specific applications areas, as well as an overview of the future outlook for mobile computing. I am pleased to be able to recommend this timely reference source to readers, be they researchers looking for future directions to pursue when examining issues in the field, or practitioners interested in applying pioneering concepts in practical situations and looking for the perfect tool. Ismail Khalil Ibrahim, Johannes Kepler University Linz, Austria January 2007
xxviii
Preface Nowadays, mobile communication, mobile devices, and mobile computing are widely available. Everywhere people are carrying mobile devices, such as mobile phones. The availability of mobile communication networks has made a huge impact to various applications, including commerce. Consequently, there is a strong relationship between mobile computing and commerce. The Encyclopedia of Mobile Computing and Commerce brings to readers articles covering a wide range of mobile technologies and their applications. Mobile commerce (m-commerce) is expanding, and consequently the impact to the overall economy is considerable. However, there are still many issues and challenges to be addressed, such as mobile marketing, mobile advertising, mobile payment, mobile authorization using voice, and so on. Providing users with more intelligent product catalogues for browsing on mobile devices and product brokering also plays an important role in m-commerce. Furthermore, the impact mobile devices give to the supply chain must be carefully considered. This includes the use of emerging mobile technology, such as RFID, sensor network, and so forth. A wide range of mobile technology is available for m-commerce. Mobile phones are an obvious choice. Additionally, there are many different kinds of mobile phones sold in the market, some of which are labelled as smartphones. There is much research conducted in conjunction with the use of mobile phones. Mobile phone text messaging and SMS are common among mobile users. Subsequently, the use of text messaging and SMS enriches m-commerce, including the ability to support multilingual text messaging. Mobile phone supporting disability has also been a focus lately, which focuses on text messaging to disabled people. More advanced applications now require additional services, such as chatting using Bluetooth, mobile querying, and voice recognition. Mobile privacy issues are also still an important topic. Apart from mobile phones, there is a wide variety of mobile technology, some of which are mobile robots, RFID, penbased mobile computing, and so forth. Many advanced applications have been developed utilizing these technologies. Current research has been focusing on man-machine interfaces and sensory systems, particularly for mobile robots, biometric and voice based authentication, traffic infractions, and so forth. The context of smart spaces also gives a new dimension to mobile technology. The use of mobile technology in entertainment is growing rapidly. Some examples include mobile phone gambling, mobile collaborative games, mobile television, mobile sport videos, and mobile hunting incorporating location-based information. The list is expanding as the technology is advancing. Understanding the success factors for mobile gaming and other entertainment is equally important as the technical aspects of the technology itself. Videos and multimedia undoubted play an important role in mobile entertainment. Video technologies, such as mobile video sequencing, mobile video transcoding, and mobile video communications, have been studied extensively. One of the main limitations of mobile devices is the limited memory capacity, which has to be carefully addressed, especially in the context of mobile multimedia, because these kinds of applications generally require large amount of spaces. Beside videos, radio technology should not be neglected either. There are many other applications of mobile technology. For example, the use of mobile technology in health, called m-health, is expanding. Mobile medical imaging is made possible thru the use of 3G wireless network. Another example is the use of mobile technology in learning, called m-learning, such as the use of SMS and text messaging, although some still argue whether m-learning is the way to go in learning, while others are still looking at how to combine the infrastructures and tools with pedagogy. Developing mobile applications requires a novel software engineering approach. The design for mobile information systems is still maturing. Some researchers are still formulating design patterns for mobile applications, while others are focusing on the user interface aspects. Programming for handheld devices is quite common to use various programming languages and tools, including Java micro edition, J2ME, Corba, and Extreme programming. Since the device generally has a small screen, content transformation and content personalization need to be examined. Other forms of interfaces, includ-
xxix
ing brain computing interfacing, are also interesting. Mobile databases and XML-based mobile technology have received some degree of attention as well. Other issues that have been incorporated into mobile technologies include mobile agents, service-oriented computing, and various forms of caching, such as peer-to-peer, cooperative, and semantic caching. Service delivery and resource discovery are gaining their popularities too. Security—especially in a mobile environment—should not be neglected. Some work on mobile PKI and limited key generations has been carried out by a number of researchers in order to contribute to advancing m-commerce. The impact of mobile technology in commerce needs to be evaluated, including its socio-psychological influence and technological adoption and diffusion, as well as readiness and transformation. We need to understand the adoption, barrier, and influencing factors of m-commerce. Some gender issues have been pointed out by some researchers. All of the abovementioned applications will not be made possible without addressing the advancement of mobile networks. Most of the articles in this encyclopedia may be categorized into the mobile network and communication category. 3G architectures have made their entries lately. Mobile ad-hoc network, IPv6 and P2P are also maturing. Some new work in wireless sensor network is presented. Last but not least, mobile technology and its applications will not be complete without mentioning location-aware and context-aware. New technologies in positioning; either indoor or outdoor, as well as tracking of moving objects, are presented. Some applications of location-aware include ad-hoc mobile querying, use of iPod as a tourist guide, location-based multimedia for monitoring purposes, and location-based multimedia for tourists. Some notable context-aware applications are notification services, context-aware mobile GIS, and semantic mobile agents for context-aware applications. As a final note, the Encyclopedia of Mobile Computing and Commerce covers a broad range of aspects pertaining to mobile computing, mobile communication, mobile devices, and various mobile applications. These technologies and applications will shape mobile computing and commerce into a new era of the 21st century whereby mobile devices are not only pervasive and ubiquitous, but also widely accepted as the main tool in commerce. David Taniar Melbourne, Australia January 2007
xxx
Acknowledgments I would like to acknowledge the help of all involved in the collation and review process of the encyclopedia, without whose support the project could not have been satisfactorily completed. I would like to thank all the staff at IGI, whose contributions throughout the whole process, from inception of the initial idea to final publication, have been invaluable. In particular, our thanks go to Kristin Roth, who kept the project on schedule by continuously monitoring the progress on every stage of the project, and to Mehdi Khosrow-Pour and Jan Travers, whose enthusiasm initially motivated me to accept their invitations to take on this project. I am also grateful to my employer Monash University for supporting this project. A special thank goes to Mr. John Goh of Monash University, who assisted me in almost the entire process of the encyclopedia: from collecting and indexing the proposals, distributing chapters for reviews and re-reviews, constantly reminding reviewers and authors, liaising with the publisher, to many other housekeeping duties, which are endless. I would also like to acknowledge the assistance and advice from the editorial board members. In closing, I wish to thank all of the authors for their insights and excellent contributions to this encyclopedia, in addition to all those who assisted us in the review process. David Taniar Melbourne, Australia January 2007
xxxi
About the Editor David Taniar received a PhD degree in computer science from Victoria University, Australia, in 1997. He is now a senior lecturer at Monash University, Australia. He has published more than 100 research articles and co-authored a number of books in the mobile technology series. He is on the editorial board of a number of international journals in the fields of data warehousing and mining, business intelligence and data mining, mobile information systems, mobile multimedia, Web information systems, and Web and grid services.
xxxii
Category: Human Factor
Academic Activities Based on Personal Networks Deployment Vasileios S. Kaldanis NTUA, Greece Charalampos Z. Patrikakis NTUA, Greece Vasileios E. Protonotarios NTUA, Greece
INTRODUCTION Personal networking has already become an increasingly important aspect of the unbounded connectivity in heterogeneous networking environments. Particularly, personal networks (PNs) based on mobile ad-hoc networking have seen recently a rapid expansion, due to the evolution of wireless devices supporting different radio technologies. Bluetooth can be considered as the launcher of the self-organizing networking in the absence of fixed infrastructure, forming pico nets or even scatternets. Similar other wireless technologies (e.g., WiFi) attract a lot of attention in the context of mobile ad hoc networks, due to the high bandwidth flexibility and QoS selection ranges they feature, leveraging the path to develop advanced services and applications destined to the end user and beyond. Furthermore, personal networks are expected to provide a prosperous business filed for exploitation to thirdparty telecom players such as service and content providers, application developers, integrators, and so forth. In this article, a personal-to-nomadic networking case is presented. Academic PN (AcPN) is a generic case that aims to describe several situations of daily communication activities within a university campus or an extended academic environment through the support of the necessary technological background in terms of communication technologies. The concept is straightforward: a number of mobile users with different characteristics and communication requirements ranging from typical students to instructors and lecturers, researchers and professors, as well as third parties (e.g., visitors, campus staff), are met, work, interact, communicate, educate, and are being educated within such an environment. This implies the presence of a ubiquitous wireless personal networking environment having nomadic characteristics. Several interesting scenarios and use cases are analyzed, along with a number of proposed candidate mobile technology solutions per usage case. The article is organized as follows: first, a general description of the academic case is presented identifying examples of typical communication activities within an academic
environment; the technical requirements necessary for a successful deployment of personal area network (PAN)/PN technologies within the academic environment are also listed. Next, specific deployment scenarios are presented, followed by a business analysis. The article closes with a concluding section.
ACADEMIC CASE DESCRIPTION The AcPN case describes several situations of daily communication activities, taking place within a typical university campus environment. Members of the academic community, such as students, make use of personal networking concepts and related technologies to acquire and maintain constant connectivity among them or with local or remote networks, and utilize offered services—applications discovered at their point of presence. In this fashion, they may exchange files on the move, interact with each other in different ways (e.g., messaging, audio/videoconference), connect to a home desktop PC to download a missing file, or configure remotely a project installation located in a lab. The AcPN case aims to support a number of communication activities known in an academic environment. Typical examples of such activities include: • • • • • •
entering the campus, and making inquiries for local information (maps, buildings, etc.); monitoring information updates (announcements, urgent notices, deadlines, events); meeting with a colleague/friend/other student mates, exchanging data with others (docs, mp3, video clips, etc.), work management, and so on; seeking a friend/colleagues somewhere on campus; communicating with a professor/tutor/technical supervisor; reporting project results to colleagues and real-time discussion;
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
A
Academic Activities Based on Personal Networks Deployment
• • • •
borrowing/returning a book from/to the local library; performing remote home/office network setup (upon returning home); monitoring and controlling a lab experiment/project installation; and responding to emergency situations within the campus area (fire drill, medical assistance, etc).
The objective of developing the AcPN case is to provide the academic users with an easy way to perform their everyday work as efficiently as possible—in the least time and with the least cost. The academic entity concept-model used here is very general and includes all different types of academics existing in a typical university environment. These are undergraduates/postgraduates/PhD students, tutors/lecturers/professors, research associates, and third-party entities such as visitors and permanent/temporary staff. The campus infrastructure is supposed to support as many communication technologies as possible to the academic entities roaming on campus, in order to provide a variety of services, featuring flexibility in constructing different networking configurations. These technologies could range from short-distance wireless protocols (Bluetooth, infrared) to large-scale networking solutions such as WLAN or GSM/ GPRS and 3G/UMTS. In any case, academic users can benefit from PN concepts such as P-PAN, PAN/PN, W-PAN, and so forth in order to acquire access to other networks or services. Each user is equipped with a number of wireless communicating devices such as mobile phones, PDAs, laptops, headsets, and mobile storage devices, featuring GSM/GPRS/UMTS Bluetooth and WiFi technologies. These devices can detect and interact with each other in various ways, providing new communication capabilities and fields for different networking configurations. For example, a student is able to form his own personally attached network or private PAN (P-PAN) by interconnecting his wearable short-range devices (e.g., headset, mp3 player, mobile hard disc, PDA) via Bluetooth or infrared protocol. On a larger scale, the user can also connect to a local network of short-range devices (other users’ devices or local wireless printer) becoming part of the existing personal area network, and interact with users in his or her close vicinity who belong to the same network. The student may use his or her mobile device as a GSM/GPRS or UMTS terminal to extend his or her current P-PAN and PAN configuration in order to connect to his or her home DSL network to download an important file from the remote desktop PC. In this case, the student establishes a personal network that can be further used for numerous other remote actions. In the same way, the ubiquitous campus network provider can interconnect all PANs within the campus area and form a “personal”-like network: the campus PN.
Similarly, any other academic user can form one or more PNs dependent on the following parameters: • • • •
the number of interconnecting devices, the inherent characteristics of used wireless technologies, the connection capabilities per technology in terms of bandwidth and QoS, and the requirements imposed by each service used on a particular PN.
Finally, administration of the campus PN is a very important issue for the successful management of attached users in terms of resources and security and successful service provision. Different security levels can be used, according to the trust policy followed when a foreign user (e.g., visitor) is accepted locally in a PAN or globally in the campus PN.
PN CONCEPT IN ACADEMIC CASE PNs in our case comprise potentially all of a person’s devices capable to detect and connect each other in the real or virtual vicinity. Connection is performed via any known and applicable wireless access technology (Bluetooth, infrared, WiFi, MAGNET low/high data rate, WLAN/GSM/GPRS/UMTS, and so on). PN establishment requires an extension of the present and locally detected PAN by the person’s attached network (set of person’s devices) called private PAN. The physical architecture of the networks and devices (for the AcPN case) has already been mentioned, while all interactions among them is illustrated in the Figure 1. PNs are configured in an ad hoc fashion, establishing any possible peer-to-peer (P2P) connection among users belonging to the same local PAN and other remote PANs or PNs as well, in order to support a person’s private and professional applications. Such applications may be installed and executed on a user’s personal device, but also on foreign devices in the same way. PNs consist of communicating clusters of personal digital devices, possibly shared with others and connected through different communication technologies remaining reachable and accessible via at least a PAN/PN. Obviously, PANs have a limited geographical coverage, while PNs have unrestricted geographical span, incorporating devices into the personal environment, regardless of their physical or geographical location. In order to extend their access range, they need the support of typical infrastructure-based and ad-hoc mobile networks. Strict security policies determine PNs’ performance. Any visiting (foreign to the local PAN) mobile user bearing his or her own P-PAN may acquire trust and become a member of the locally detected PAN, as long as another member of the same PAN can guarantee his or her proper behavior in
Academic Activities Based on Personal Networks Deployment
Figure 1. Academic PN concept topology and interactions
A
Home Cluster Car Cluster `
Infrastructure
Foreign PNs
Campus Area
Cellular Network Student’s P -PAN
Library’s PN Foreign PNs
Other P-PANs
Wireless/Wired LAN
Office Cluster
Table 1. Devices used in AcPN case Laboratory
P-PAN
Home PAN
Office PAN
Campus PN
Desktop PC
-
√
√
√
√
Laptop
Optional
Optional
Optional
Optional
√
PDA
√
√ (Optional)
√ (Optional)
√ (Optional)
√ (Optional)
Mobile Phone
√
Optional
Optional
√
√
MP3 Player
√
√
Optional
Optional
-
Wireless Headset
√
-
-
-
√
Printer
-
√
√
-
√
Scanner
-
√
√
-
√
Mobile Hard Drive
-
√ (Optional)
√ (Optional)
-
√
Camera
-
√
√
-
√
DVD R/Player
-
√
-
-
√
Wall Screen
-
√
√
-
√
Sensor Tx/Rx
-
-
-
-
√
this network. In this way, the new user can become trusted and behave as any other existing user in the PAN. Similar mechanisms exist for AAA functionalities in other clusters and PN domains as well. A list of important devices for the use cases listed formerly is summarized in Table 1.
•
•
PAN / PN
Desktop PC and Laptop: High processing power, unlimited power supply, high storage capacity, graphical UI, support of 802.11/Ethernet/Bluetooth, HDR, Internet connectivity (TCP/IP, UDP, etc.), database software, and so forth. PDA: Low processing power, unlimited power supply, high storage capacity, graphical UI, HDR/LDR,
Academic Activities Based on Personal Networks Deployment
•
• •
• •
support 802.11/Ethernet/Bluetooth/56K, support of TCP/IP, security software integrated, low weight, and so forth. High-Featured Mobile Phone Device: Low processing power, low power consumption, moderate storage space, support of wireless protocols (Bluetooth/GSM/ CDMA), IrDA support, cellular connectivity (GSM/ GPRS/UMTS), WiFi/WLAN connectivity, portability, synchronization with other devices. Printer: Support of various wireless technologies (Bluetooth, IrDA, etc), wired networking, and so forth. MP3 Player: Weak power supply, moderate storage capacity, support of basic wireless access technologies (Bluetooth/WiFi), HDR/LDR, large battery power consumption, low recharging time, handy UI and control, high sound quality, and so forth. Wireless Headset: Support of basic wireless access technologies (Bluetooth, IrDA), LDR, and so forth. Wireless Sensors: Low power consumption, wireless interconnectivity (Bluetooth, IrDA, etc.), LDR, large operation life flexible functionality, light weight device, low volume, remotely controllable, and so forth.
It should be noted that currently, there is ongoing work on specifying devices that support new protocols (especially in the wireless physical layer), and the expansion of the use cases to current networking technologies is also still under development.
SCENARIOS AND USE CASES The scenario generation procedure has been based on the obtained results from an end user workshop held at the NTUA campus. The workshop participants were academic people coming from different knowledge backgrounds and professions (undergraduates/MSc students, PhD candidates/ research associates, tutors, lecturers, professors, and visitors). During the workshop all participants had the chance to exchange thoughts and express their own needs regarding communication solutions and services they wish or expect to have within a typical campus area environment.
Login to the Ubiquitous Campus PN Network This is a fundamental case for the AcPN, since it presents the most important thing an AcPN user must do if he wants to utilize services and applications available in the university domain (single campus or a set of campuses belonging to the same organization). According to this case, the AcPN user must login to the campus network via his mobile device mainly in two cases:
whenever he reaches the real campus area physically (e.g., by car, by bus, or by foot) via a locally detected campus PAN or remotely via a PN which he has previously established dynamically with the campus PN. The AcPN could be a registered user to the campus network (e.g., student, lecturer, researcher, or permanent staff) or a foreign (third-party) user (e.g., visitor) who should follow a registration procedure before attaching to the local network. The login procedure is required for the AcPN case in order to maintain a certain level of security, which is higher for locally connected users in contrast to remotely connected ones. After successful login, the AcPN users can immediately be informed by the campus PN administrator for urgent messages from their colleagues, reminders, scheduled power outages, and other important messages of general importance.
Information Update and Real-Time P2P Interaction In this use case, an AcPN member, after logging into the campus PN network, wishes to have access to any available local services and applications according to his or her educational activity (e.g., student, researcher). At the same time, he or she can be informed about course announcements, important notices (e.g., deadline extensions, change of lecture classrooms, etc.) from the student office or from any other local online source related to his or her studies. Furthermore, using a mobile device he or she may directly connect to a course database to download important files such as handouts, past papers, presentations, or any other material in electronic form. In P2P fashion, the student may have the chance to see on his or her device who is currently roaming into the campus area from among his contact people (friends, colleagues, tutors, etc.) and to interact with them in various ways. He may also publish hello messages everywhere he wants to, arrange a meeting (physical or not) on the fly, be informed by other people who also “see” him on their devices, exchange files with friends (mp3s, pictures, video clips), send an important file to a colleague or to his or her technical supervisor, setup an audio/video conference, and so forth.
Using a Trusted PAN to Connect to Other Networks In this scenario, a mobile user who is not a member in the campus PN currently lies within the campus and wishes to get an Internet connection or to acquire access to the local network for several reasons (e.g., utilize local services, get library access, view local events, etc.). This user is considered a foreign user, since he does not belong to the campus PN or to any other local PAN, as privileged campus PN members do. Obviously, the foreign user is considered by the campus
Academic Activities Based on Personal Networks Deployment
network as a third party-user or a visitor and in some way has to be accepted by the campus PN administrator into the ubiquitous local network. This can be done directly or indirectly. In the direct way, the user can be connected to a locally detected PAN at its point of presence if another registered user of the same PAN can guarantee his or her proper and safe behavior. In other words, the foreign user may be attached to any PAN and consequently to the campus PN if another user of the same PAN can verify him or her as a trusted entity and provide him or her with access rights characterized by the basic required security level. In case the foreign user violates the invitation policy agreement, he or she may be warned or even banned by any other PAN user reporting the event to the campus PN administrator. Then the user who signed his or her trustworthiness may lose credits on his or her membership to the campus PN, or his or her authorization provision to other users in the future may be suspended for some period. Following the indirect way, the foreign user may use the local wide area network (e.g., WLAN) to ask for a temporary registration from the campus PN administrator. For example he or she may use a credit card to register to the ubiquitous campus PN network; buy connection time duration; service access rights, bandwidth, and QoS; and so on. In this way, the registered foreign user may be accepted by any other PAN anywhere in the campus, gaining access to the allowed local services in general. This type of user cannot access individual department resources and services (e.g., engineering department database, ftp software, etc.) but only allowed services for third-party users (e.g., library access, local knowledge base intranets, projects, etc.).
Remote Laboratory Monitoring and Control This is the case where a remote monitoring and controlling of a procedure taking place in a location is required using PAN/PN technologies. Particularly, a group of scientists (students, researchers, professors, etc.) is performing a lab experiment that is long lasting, and the overall progress and results need to be monitored continuously on a 24-hour basis. Furthermore, it is required that according to the collected ongoing results, some experiment parameters may be changed dynamically (locally or remotely). The scientific group must have continuous communication using their mobile devices independent of their point of presence, in order to discuss the change of parameters whenever needed to do so. In this case we consider that there is no physical presence by any member to the lab location and the procedure runs remotely using PAN/PN. The experiment consists of a number of wireless sensors attached on the examined sample under test, forming a P-PAN which sends reports to a report collector. The report collector enriches the raw report signals and forwards them
to the central processing device (high processing power desktop PC) where the experiment software is running. The central processing device sends formatted reports to a local database for data warehousing purposes, while reporting results to the scientific group using the lab PAN as well. Each member of the scientific group has been attached to the lab PAN forming individual PNs and also maintains a direct online connection with the other members for results discussion. Depending on the results, if a parameter change is decided, the user responsible for the experiment sends the required commands to the command executor device, which runs an external application controlling the interaction functionality with the sample under test. The change is verified and archived wirelessly into the database, again using the lab PAN, while a report is sent back to the group about its successful command execution.
Future Library Loaning and Reservation This scenario presents a proposed loaning and reservation system for academic libraries in the future. In this case, the reservation and loaning of a book title may be performed based on the well-known Web service (via the library Web site) and the campus PN infrastructure. The campus PN consists of all PAN/PN clusters in different departments (or offices/labs, etc.) or smaller departmental libraries and the on-campus users equipped with mobile devices. According to this scenario, a requestor for a book is an on-campus entity (normal/MSc/PhD student, research fellow), who is using his or her mobile device and the campus networking infrastructure to get access to the local online library database. The requestor should also be a registered member of the campus PN with a stored profile in the university database already logged in. This profile entry automatically enables a number of useful privileges according to the AcPN user type (user profession) that allows him or her to access specific applications and services. An example use scenario is the following: a requestor gets informed by the library service on his mobile that a requested book is currently loaned and has been delayed to return (i.e., for a day). He is also notified about the priority in the request queue (if any exists) for that title. After that, the system generates an urgent message and forwards it to the loaner of the book using the campus PN. The system, using a tracing mechanism regarding the user status-location, is aware that the loaner is currently active and able to receive notifications via the campus PN, so it prefers to notify the user in this way. The loaner must provide as soon as possible a new book returning date to the library system if he does not want his membership to be blacklisted or in the worst case banned from the campus PN database. Hence, the loaner provides as the new returning date a specific time during the same day. The system forwards the new returning date to the requestor and provides a validity period for
A
Academic Activities Based on Personal Networks Deployment
his request. After that period, his request is no longer valid and a next requestor (on the queue) gets the right to reserve that book. When finally the book returns to the library desk, the system via the campus PN notifies the active requestor about the book availability and his validity period to come and collect it. The requestor may provide himself as the collecting person or another registered PN user.
Remote Course Exam Participation and Distant Learning Finally, using this case, a student currently away from the campus area for several reasons (urgent reasons, recuperation in hospital, etc.) has the option to participate remotely in her course exams using the PN technology. At her current location, she as to scan for a local PAN to attach or to search for another local wireless Internet connection means (e.g., WLAN, UMTS, WiFi). Then she must setup a PN with the campus network, logon to the campus PN using her student account, and connect to the local examinations server who has privately published an exams-related session link for such cases. Then after authorizing and authenticate herself, she must download the support software for this online session or any other auxiliary utility supplied by the exam center administrator, install it properly, and directly connect to the exam server before the actual start time of the exams. It is supposed that she has already applied for a remote exam participation by sending an e-mail to the exam administrator, and on reply she has received all the relevant details—information of that session according to the course requirements (e.g., multiple-choice form), connection bandwidth, QoS, and personal mobile device capabilities (e.g., large viewable display, keyboard, memory, etc). The student using this exam PN session participates remotely in the same way she would if she was present in the real exam center location for the required time period of the exam. It is required that she has an interruptible connection with the campus PN network and particularly with the exam center local server. The student provides her answers to the exam paper questions by ticking the appropriate box in each online XML Web interface, presses the “SEND” button to proceed to the next question, and so on. Each provided answer cannot be changed or undone since it has already been sent to the server and saved to the database system. If any problem occurs (e.g., connection is lost or service application fails), the session state is continuously monitored by the exam administrator and resumed when the problem is solved. At the end, the session is closed and a message informs the student that the application has already completed successfully. The service will later inform the student of her achieved results. In the same way any possible distant learning activity can be supported using similar PN setups and configurations as long as the remote users can create any possible type of
PN with the distant network of interest where a relevant service can run reliably.
BUSINESS PROSPECTS Many players in mobile business may find PN technology to be a prosperous field to extend the market in many dimensions, ranging from high data rate connectivity solutions to advanced services and Web-based applications. The value chain of the mobile market can be dynamically expanded including more than one network and service providers, integrators, service and application developers, or even small-to-medium network operators. The AcPN case exploits PN concepts in a very efficient way, allowing the use of well-known wireless technologies and common networking configurations of the present and the future to be used and easily applied. Target users are people actively involved in educational activities who present high expectations from communication technologies such as increased bandwidth, connection flexibility (among different technologies), use of a wide range of services and applications, more personalized devices, large mobile storage capability, interoperability, friendly user-device interface, and so forth. Based on the collected results from the AcPN end user workshop held in Athens, Greece, a number of important requirements have been identified. These requirements have led to several conclusions regarding the new players in the value chain and the business aspects of PAN/PN concepts within the academic environment. The most important conclusions are: •
•
Regarding Network Infrastructure: The network infrastructure should include the normal mobile networks (GSM/GPRS/UMTS), as well as additional networking infrastructure such as WLAN/WiFi on a single or multioperator environment and the ubiquitous campus PN operator. The campus PN infrastructure must include networking configurations among all campus PANs (different departments, labs, offices) and possibly other PNs (other campuses of the same organization). Regarding Security: The campus PN operator is responsible for network security in the supported connections of the wired/wireless domain, user login/logout functionality, mobility support within the campus (or campuses of the same university), and other required PAN/PN operations. If the particular university operates more than one campus, then a university PN is required to interconnect the different campus PNs and support the previous on a higher administrative level, securing of course the communication between the PNs.
Academic Activities Based on Personal Networks Deployment
•
•
•
Regarding Service Aggregators: In this case, the role of service aggregation and provision to the AcPN users is performed primarily by the campus PN operator and partially by third-party service operators who may have agreements with the campus operator. Any service provided to the campus PN is expected to be controlled and maintained by the unique campus PN operator, which plays the twofold role of the service aggregator and the provider. Other services can be provided on the campus by typical mobile operators through the use of voice, e-mail, SMS, or MMS, but PN services and relevant interconnections must be realized via the campus PN network or service operator. Regarding Terminal Equipment: This requirement takes into account all the different vendors and manufacturers who provide the terminal devices to the end users. The fact that any AcPN user is supposed to be equipped with his or her own P-PAN requires a number of different featured portable devices coming from different vendors to be used. This is feasible as long as the PAN-proposed standards are supported. (It should be noted that for the air interface, the MAGNET LDR/HDR standard has been proposed.) Regarding End Users: These can be divided into two types. The first one includes all the normal students (undergraduates, postgraduates, etc.) who wish to use typical (low QoS) applications and services (Web browsing, chat, e-mail, voice, SMS, MMS, etc.) within the campus PN at a low cost. The second user type includes any other academic person or third party (visitors, temporary staff) who wish to have (and are willing to pay for) a higher bandwidth wireless connection or access to QoS demanding services such as (real-time) audio/videoconference, streaming applications, and so on. Such users could be professors, tutors, researchers, associates, or general university employees who use telecom technology to communicate with their work contacts for several reasons.
CONCLUSION The academic case is very promising for the future deployment of PN technologies for many important reasons. First of all, it attempts to combine and reuse efficiently almost any wireless access technologies of the present with proposed ones for the future in many scalable configurations according to the case. Secondly, it provides the option to choose which type of PN could better serve its purposes in terms of connection bandwidth and cost. The user may choose the most efficient way (in terms of cost) to construct his or her own PN; for example, he or she may prefer a relatively cheap WLAN to connect to his or her office rather than a UMTS. Finally, since the use of PN technology might not
be possible in some cases without the existence of PAN or P-PAN, the definition of clusters eases the PAN or P-PAN formation as a set of preferable devices, but not all.
ACKNOWLEDGMENTS The AcPN case was presented, developed, and analyzed in detail within the IST-MAGNET framework (http://www. ist-magnet.org). Specific documents referenced include: MAGNET WP1 Task 1.1, D.1.1.1a, March 2004; MAGNET, WP1 Task 1.4, D1.4.1a, September 2004; MAGNET WP1 Task 1.1, D.1.1.1b, December 2004; MAGNET WP1 Task 1.1, D.1.1.1b, December 2005; and Academic Case Workshop Results, Internal Report D-1.3.1b. The work acceptance by the academic community is very encouraging and promising for the future. Currently the project group is implementing, based on the previous use cases and scenarios, a number of services.
KEY TERMS Academic PN (AcPN): Use case descriptive name for a PN exploitation into a typical academic environment. Cluster: A network of personal devices and nodes located within a limited geographical area (such as a house or a car) which are connected to each other by one or more network technologies and characterized by a common trust relationship between each other. Context: The information that characterizes a person, place, or object. In that regard, there exist user, environment, and network context. The context information is used to enable context-aware service discovery. Foreign Device: A device that is not personal and cannot be part of the PN. The device can be either trusted, having an ephemeral trust relationship with another device in the PN, or not trusted at all. Private Personal Area Network (P-PAN): A dynamic collection of personal nodes and devices around a person. Personal Area Network (PAN): A network that consists of a set of mobile and wirelessly communicating devices that are geographically close to a person but which may not belong to him. Personal Device: A device related to a given user or person with a pre-established trust attribute. These devices are typically owned by the user. However, any device exhibiting the trust attribute can be considered as a personal device. The same remarks as those for the personal nodes definition hold for devices.
A
Academic Activities Based on Personal Networks Deployment
Personal Network (PN): Network including the PPAN and a dynamic collection of remote personal nodes and devices in clusters that are connected to each other via interconnecting structures.
Category: Mobile Software Engineering
Accessibility of Mobile Applications Pankaj Kamthan Concordia University, Canada
INTRODUCTION The increasing affordability of devices, advantages associated with a device always being handy while not being dependent on its location, and being able to tap into a wealth of information/services has brought a new paradigm to mobile users. Indeed, the mobile Web promises the vision of universality: access (virtually) anywhere, at any time, on any device, and to anybody. However, with these vistas comes the realization that the users of the mobile applications and their context vary in many different ways: personal preferences, cognitive/neurological and physiological ability, age, cultural background, and variations in computing environment (device, platform, user agent) deployed. These pose a challenge to the ubiquity of mobile applications and could present obstacles to their proliferation. This article is organized as follows. We first provide the motivation and background necessary for later discussion. This is followed by introduction of a framework within which accessibility of mobile applications can be systematically addressed and thereby improved. This framework is based on the notions from semiotics and quality engineering, and aims to be practical. Next, challenges and directions for future research are outlined. Finally, concluding remarks are given.
BACKGROUND The issue of accessibility is not new. However, the mobile Web with its potential flexibility on both the client-side and the server-side presents new challenges towards it.
Figure 1 illustrates the dynamics within which the issue of accessibility of a mobile application arises. We define a mobile application as a domain-specific application that provides services and means for interactivity in the mobile Web. For example, education, entertainment, or news syndication are some of the possible domains. The issue of accessibility is intimately related to the user and user context that includes client-side computing environment. To that regard, we define accessibility in context of a mobile application as access to the mobile Web by everyone, regardless of their human or environment properties. A consumer (user) is a person that uses a mobile application. A producer (provider) is a person or an organization that creates a mobile application.
The Consumer Perspective of Mobile Accessibility The accessibility concerns of a consumer are of two types, namely human and environment properties, which we now discuss briefly.
Human Properties Human properties are issues relating to the differences in properties among people. One major class of these properties is related to a person’s ability, and often the degree of absence of such properties is termed as a disability. We will use the term “disability” and “impairment” synonymously. The statistics vary, but according to estimates of the United Nations, about 10% of the world’s population is considered disabled. The number of people with some form of disability that do have access to the Internet is in the millions.
Figure 1. The interrelationships between a consumer, a producer, accessibility, and a mobile application isServedBy
Consumer
isAConcernFor
Accessibility
interactsWith
Producer needsToAddress
isAQualityAttributeOf
isResponsibleFor
Mobile Application
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
A
Accessibility of Mobile Applications
Table 1. A semiotic framework for accessibility of mobile applications Semiotic Level
Quality Attributes Accessibility [T4;E]
Pragmatic
Comprehensibility, Interoperability, Performance, Readability, Reliability, Robustness [T3;E]
Semantic
Completeness and Validity [T2;I]
Syntactic
Correctness (Primary Notation) and Style (Secondary Notation) [T1;I]
There are several types of disabilities that a producer of a mobile application needs to be concerned with. These can include visual (e.g., low visual acuity, blindness, color blindness), neurological (e.g., epilepsy), auditory (e.g., low hearing functionality, deafness), speech (e.g., difficulties in speaking), physical (e.g., problems using an input device), cognitive (e.g., difficulties of comprehending complex texts and complex structures), cultural/regional (e.g., differences in the use of idioms, metaphors leading to linguistic problems).
Environment Properties Environment properties are issues relating to different situations in which people find themselves, either temporarily or permanently. These situations could be related to their connectivity, the location they are in, or the device/platform/user agent they are using. For example, a user using a computer in a vehicle shares many of the issues that some people have permanently due to a disability in hand motorics. Or, for example, a user may be accessing the same information using a personal digital assistant (PDA) or a cellular phone.
The Producer Perspective of Mobile Accessibility The motivation for accessibility for a business is to reach as many users as possible and in doing so reduce concerns over customer alienation. It is the producer of the mobile application that needs to adjust to the user context (and address the issue of accessibility), not the other way around. It is not reasonable for a producer to expect that the consumer environment will be conducive to anything that is delivered to him/her. In certain cases, when a consumer has a certain disability, such adaptation is not even possible. If the success of a mobile application is measured by the access to its services, then improving accessibility is critical for the producers. Still, any steps that are taken by 0
Means for Accessibility Assurance and Evaluation
Decision Support
• Training in Primary and Secondary Notation • “Expert” Knowledge (Principles, Guidelines, Patterns) • Inspections • Testing • Metrics • Tools
Feasibility
a producer related to a mobile application have associated costs and trade-offs, and the same applies to improvements towards accessibility.
Initiatives for Improving Accessibility in Mobile Contexts There are currently only a few efforts in systematically addressing accessibility issues pertaining to mobile applications. There are guidelines available for addressing accessibility (Chisholm, Vanderheiden, & Jacobs, 1999; Ahonen, 2003) in general and language-specific techniques (Chisholm et al., 2000) in particular.
ADDRESSING THE ACCESSIBILITY OF MOBILE APPLICATIONS To systematically address the accessibility of mobile applications, we take the following steps: 1. 2. 3.
View accessibility as a qualitative aspect and address it indirectly via quantitative means. Select a theoretical basis for communication of information (semiotics), and place accessibility in its setting. Address semiotic quality in a practical manner.
Based on this, we propose a framework for accessibility of mobile applications (see Table 1). The external attributes (denoted by E) are extrinsic to the mobile application and are directly the consumer’s concern, while internal attributes (denoted by I) are intrinsic to the mobile application and are directly the producer’s concern. Since not all attributes corresponding to a semiotic level are on the same echelon, the different tiers are denoted by “Tn.” We now describe each component of the framework in detail.
Accessibility of Mobile Applications
Semiotic Levels The first column of Table 1 addresses semiotic levels. Semiotics (Stamper, 1992) is concerned with the use of symbols to convey knowledge. From a semiotics perspective, a representation such as a mobile resource can be viewed on six interrelated levels: physical, empirical, syntactic, semantic, pragmatic, and social, each depending on the previous one in that order. The physical and empirical levels are concerned with the physical representation of signs in hardware and communication properties of signs, and are not of direct concern here. The syntactic level is responsible for the formal or structural relations between signs. The semantic level is responsible for the relationship of signs to what they stand for. The pragmatic level is responsible for the relation of signs to interpreters. The social level is responsible for the manifestation of social interaction with respect to signs, and is not of direct concern here. We note that none of the layers in Table 1 is sufficient in itself for addressing accessibility and intimately depends on other layers. For example, it is readily possible to create a document in XHTML Basic, a markup language for small information appliances such as mobile devices, that is syntactically correct but is semantically non-valid. This, for instance, would be the case when the elements are (mis)used to create certain user-agent-specific presentation effects. Now, even if a mobile resource is syntactically and semantically acceptable, it could be rendered in such a way that it is unreadable (and therefore violates an attribute at the pragmatic level). For example, this could be the case by the use of very small fonts for some text, or the colors chosen for background and text foreground being so close that the characters are hard to discern.
robustness, which are also at the pragmatic level. Since these are perceived as necessary conditions, violations of any of these lead to a deterioration of accessibility.
Means for Accessibility Assurance and Evaluation The third column of Table 1 lists the direct and indirect (and not necessarily mutually exclusive) means for assuring and evaluating accessibility: •
Quality Attributes The second column of Table 1 draws the relationship between semiotic levels and corresponding quality attributes. We contend that the quality attributes we mention are necessary but make no claim of their sufficiency. The internal quality attributes for syntactic and semantic levels are inspired by Lindland, Sindre, and Sølvberg (1994). At the semantic level, we are only concerned with the conformance of the mobile application to the domain it represents (that is, semantic correctness or completeness) and at the syntactic level the interest is in conformance with, with respect to the languages used to produce the mobile application (that is, syntactic correctness). Accessibility belongs to the pragmatic level and depends on the layers beneath it. It in turn depends upon the other external quality attributes, namely comprehensibility, interoperability, performance, readability, reliability, and
•
Training in Primary and Secondary Notation: The knowledge of the primary notation of all technologies (languages) is necessary for guaranteeing conformance to tier T1. The Cognitive Dimensions of Notations (CDs) (Green, 1989) are a generic framework for describing the utility of information artifacts by taking the system environment and the user characteristics into consideration. Our main interest here is in the CD of secondary notation. This CD is about appropriate use (that is, style) of primary notation in order to assist in interpreting semantics. It uses the notions of redundant recoding and escape from formalism along with spatial layout and perceptual cues to clarify information or to give hints to the stakeholder, both of which aid the tiers T2 and T3. Redundant Recoding is the ability to express information in a representation in more than one way, each of which simplifies different cognitive tasks. It can be introduced in a textual mobile resource by making effective use of orthography, typography, and white space. Escape from Formalism is the ability to intersperse natural language text with formalism. Mobile resources could be complemented via natural language annotations (metadata) to make the intent or decision rationale of the author explicit, or to aid understanding of stakeholders that do not have the necessary technical knowledge. Incidentally, many of the language-specific techniques for accessibility (Chisholm et al., 2000) are in agreement with this CD. “Expert” Body of Knowledge: The three types of knowledge that we are interested are principles, guidelines, and patterns. Following the basic principles (Ghezzi, Jazayeri, & Mandrioli, 2003; Bertini, Catarci, Kimani, & Dix, 2005) underlying a mobile application enables a provider to improve quality attributes related to tiers T1-T3 of the framework. However, principles tend to be abstract in nature which can lead to multiple interpretations in their use and not mandate conformance. The guidelines encourage the use of conventions and good practice, and could serve as a checklist with respect to which an application could be heuristically or otherwise evaluated. The guidelines
A
Accessibility of Mobile Applications
•
•
•
available for addressing accessibility (Chisholm et al., 1999; Ahonen, 2003) when tailored to mobile contexts can be used as means for both assurance and evaluation of accessibility of mobile applications. However, guidelines tend to be more useful for those with an expert knowledge than for a novice to whom they may seem rather general to be of much practical use. The problems in using tools to automatically check for accessibility have been outlined in Abascal, Arrue, Fajardo, Garay, and Tomás (2004). Patterns (Alexander, 1979) are reusable entities of knowledge and experience aggregated by experts over years of “best practices” in solving recurring problems in a domain including mobile applications (Roth, 2002; Van Duyne, Landay, & Hong, 2003). They are relatively more structured compared to guidelines and, if represented adequately, provide better opportunities for sharing and reuse. There is, however, an associated cost of learning and adapting patterns to new contexts. Inspections: Inspections (Wiegers, 2002) are a rigorous form of auditing based upon peer review that can address quality concerns for tiers T1, T2, and most of T3, and help improve the accessibility of mobile applications. Inspections could, for example, use the guidelines and decide what information is and is not considered “comprehensible” by consumers at-large, or whether the choice of labels in a navigation system enhances or reduces readability. Still, inspections, being a means of static verification, cannot completely assess interoperability, performance, reliability, or robustness. Furthermore, inspections do involve an initial cost overhead from training each participant in the structured review process, and the logistics of checklists, forms, and reports. Testing: Some form of testing is usually an integral part of most development models of mobile applications (Nguyen, Johnson, & Hackett, 2003). However, due to its very nature, testing addresses quality concerns only at of some of the tiers (T1, subset of T2, subset of T3). Interoperability, performance, reliability, and robustness would intimately depend on testing. Unlike inspections, tool support is imperative for testing. Therefore, testing complements but does not replace inspections. Metrics: In a resource-constrained environment of mobile devices, efficient use of time and space is critical. Metrics (Fenton & Pfleeger, 1997) provide a quantitative means for making qualitative judgments about quality concerns at technical levels. There is currently limited support for metrics for mobile applications in general and for their accessibility (Arrue, Vigo, & Abascal, 2005) in particular. Any dedicated effort of deploying metrics for accessibility measure-
•
ment would inevitably require tool support, which at present is lacking. Tools: Tools that have help improve quality concerns at all tiers. For example, tools can help report violations of accessibility guidelines, or find non-conformance to markup or scripting language syntax. However, at times tools cannot address some of the stylistic issues (such as an “optimal” distance between two text fragments that will improve readability) or semantic issues (like semantic correctness of a resource included in a mobile application). Therefore, the use of tools as means for automatic accessibility evaluation should be kept in perspective.
Decision Support A systematic approach to a mobile application must take a variety of constraints into account: organizational constraints (personnel, infrastructure, schedule, budget, and so on) and forces (market value, competitors, and so on). A producer would need to, for example, take into consideration the cost of an authoring tool vs. the accessibility support it provides; since complete accessibility testing is virtually impossible, determine a stopping criteria that can be attained within the time constraints before the application is delivered; and so on. Indeed, the last column of Table 1 acknowledges that with respect to any assurance and/or evaluation, and includes feasibility as an all-encompassing consideration on the layers to make the framework practical. There are well-known techniques such as analytical hierarchy process (AHP) and quality function deployment (QFD) for carrying out feasibility analysis, and further discussion of this aspect is beyond the scope of this article.
FUTURE TRENDS Much of the development of mobile applications is carried out on the desktop. The tools in the form of software development toolkits (SDK) and simulators such as Nokia Mobile Internet Toolkit, Openwave Phone Simulator, and NetFront Mobile Content Viewer assist in that regard. However, explicit support for accessibility in these tools is currently lacking. The techniques for accessibility for mobile technologies such as XHTML Basic/XHTML Mobile Profile (markup of information) and CSS Mobile Profile (presentation of information) would be of interest. This is especially an imperative considering that the widely used traditional representation languages such as Compact HTML (cHTML), an initiative of the NTT DoCoMo, and the Wireless Markup Language (WML), an initiative of the Open Mobile Alliance (OMA),
Accessibility of Mobile Applications
have evolved towards XHTML Basic or its extensions such as XHTML Mobile Profile. Identification of appropriate CDs, and an evaluation of the aforementioned languages for presentation or representation of information in a mobile context with respect to them, would also be of interest. As mobile applications increase in size and complexity, a systematic approach to developing them arises. Indeed, accessibility needs to be a part of the entire lifecycle of a mobile applicationthat is, in the typical workflows of planning, modeling, requirements, design, implementation, and verification and validation. To that regard, integrating accessibility into “lightweight” process methodologies such as Extreme Programming (XP) (Beck & Andres, 2005) that is adapted for a systematic development of small-to-medium scale mobile applications would be useful. A similar argument can be made for the “heavyweight” case, for example, by instantiating the Unified Process (UP) (Jacobson, Booch, & Rumbaugh, 1999) for medium-to-large scale mobile applications. Finally, a natural extension of the issue of accessibility is to the next generation of mobile applications, namely mobile applications on the semantic Web (Hendler, Lassila, & Berners-Lee, 2001). The mobile applications for the Semantic Web present unique accessibility issues such as inadequacy of current searching techniques (Church, Smyth, & Keane, 2006) and a promising avenue for potential research.
CONCLUSION This article takes the view that accessibility is not only a technical concern, it is also a social right. In that context, the issues of credibility and legality are particularly relevant as both are at a higher echelon (social level) than accessibility within the semiotic framework. Credibility is considered to be synonymous to (and therefore interchangeable with) believability (Hovland, Janis, & Kelley, 1953). Indeed, improvement of accessibility is necessary for a demonstration of expertise, which is one of the dimensions (Fogg, 2003) of establishment of credibility of the producer with the consumer. Accessibility is now a legal requirement for public information systems of governments in Canada, the U.S., Australia, and the European Union. The producers need to be aware of the possibility that, as mobile access becomes pervasive in society, the legal extent could expand to mobile applications. As is well known in engineering contexts, preventative measures such as addressing the problem early are often better than curative measures at late stages when they may just be prohibitively expensive or simply infeasible. If accessibility is to be considered as a first-class concern by the
producer, it needs to be more than just an afterthought; it needs to be integral to mobile Web engineering.
REFERENCES Abascal, J., Arrue, M., Fajardo, I., Garay, N., & Tomás, J. (2004). The use of guidelines to automatically verify Web accessibility. Universal Access in the Information Society, 3(1), 71-79. Ahonen, M. (2003, September 19). Accessibility challenges with mobile lifelong learning tools and related collaboration. Proceedings of the Workshop on Ubiquitous and Mobile Computing for Educational Communities: Enriching and Enlarging Community Spaces (UMOCEC 2003), Amsterdam, The Netherlands. Alexander, C. (1979). The timeless way of building. Oxford, UK: Oxford University Press. Arrue, M., Vigo, M., & Abascal, J. (2005, July 26). Quantitative metrics for Web accessibility evaluation. Proceedings of the 1st Workshop on Web Measurement and Metrics (WMM05), Sydney, Australia. Beck, K., & Andres, C. (2005). Extreme programming explained: Embrace change (2nd ed.). Boston: AddisonWesley. Bertini, E., Catarci, T., Kimani, S., & Dix, A. (2005). A review of standard usability principles in the context of mobile computing. Studies in Communication Sciences, 1(5), 111-126. Chisholm, W., Vanderheiden, G., & Jacobs, I. (1999). Web content accessibility guidelines 1.0. W3C Recommendation, World Wide Web Consortium (W3C). Chisholm, W., Vanderheiden, G., & Jacobs, I. (2000). Techniques for Web content accessibility guidelines 1.0. W3C Note, World Wide Web Consortium (W3C). Church, K., Smyth, B., & Keane, M.T. (2006, May 22). Evaluating interfaces for intelligent mobile search. Proceedings of the International Cross-Disciplinary Workshop on Web Accessibility 2006 (W4A2006), Edinburgh, Scotland. Fogg, B.J. (2003). Persuasive technology: Using computers to change what we think and do. San Francisco: Morgan Kaufmann. Jacobson, I., Booch, G., & Rumbaugh, J. (1999). The unified software development process. Boston: Addison-Wesley. Ghezzi, C., Jazayeri, M., & Mandrioli, D. (2003). Fundamentals of software engineering (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall.
A
Accessibility of Mobile Applications
Fenton, N. E., & Pfleeger, S. L. (1997). Software metrics: A rigorous & practical approach. International Thomson Computer Press. Green, T. R. G. (1989). Cognitive dimensions of notations. In V. A. Sutcliffe & L. Macaulay (Ed.), People and computers (pp. 443-360). Cambridge: Cambridge University Press. Hendler, J., Lassila, O., & Berners-Lee, T. (2001). The semantic Web. Scientific American, 284(5), 34-43. Hovland, C. I., Janis, I. L., & Kelley, J. J. (1953). Communication and persuasion. New Haven, CT: Yale University Press. Lindland, O. I., Sindre, G., & Sølvberg, A. (1994). Understanding quality in conceptual modeling. IEEE Software, 11(2), 42-49. Nguyen, H. Q., Johnson, R., & Hackett, M. (2003). Testing applications on the Web: Test planning for mobile and Internet-based systems (2nd ed.). New York: John Wiley & Sons. Paavilainen, J. (2002). Mobile business strategies: Understanding the technologies and opportunities. Boston: Addison-Wesley. Van Duyne, D. K., Landay, J., & Hong, J. I. (2003). The design of sites: Patterns, principles, and processes for crafting a customer-centered Web experience. Boston: Addison-Wesley.
KEY TERMS Cognitive Dimensions of Notations: A generic framework for describing the utility of information artifacts by taking the system environment and the user characteristics into consideration. Delivery Context: A set of attributes that characterizes the capabilities of the access mechanism, the preferences of the user, and other aspects of the context into which a resource is to be delivered. Mobile Accessibility: Access to the Web by everyone, regardless of their human or environment properties. Mobile Resource: A mobile network data object that can be identified by a URI. Such a resource may be available in multiple representations. Mobile Web Engineering: A discipline concerned with the establishment and use of sound scientific, engineering, and management principles, and disciplined and systematic approaches to the successful development, deployment, and maintenance of high-quality mobile Web applications. Quality: The totality of features and characteristics of a product or a service that bear on its ability to satisfy stated or implied needs. Semantic Web: An extension of the current Web that adds technological infrastructure for better knowledge representation, interpretation, and reasoning. Semiotics: The field of study of signs and their representations.
Category: Converging Technology
Acoustic Data Communication with Mobile Devices Victor I. Khashchanskiy First Hop Ltd., Finland Andrei L. Kustov First Hop Ltd., Finland
INTRODUCTION One of the applications of m-commerce is mobile authorization, that is, rights distribution to mobile users by sending authorization data (a token) to the mobile devices. For example, a supermarket can distribute personalized discount coupon tokens to its customers via SMS. The token can be a symbol string that the customers will present while paying for the goods at the cash desk. The example can be elaborated further—using location information from the mobile operator, the coupons can only be sent to, for example, those customers who are in close vicinity of the mall on Saturday (this will of course require customers to allow disclosing their location). In the example above, the token is used through its manual presentation. However, most interesting is the case when the service is released automatically, without a need for a human operator validating the token and releasing a service to the customer; for example, a vending machine at the automatic gas station must work automatically to be commercially viable. To succeed, this approach requires a convenient and uniform way of delivering authorization information to the point of service—it is obvious that an average user will only have enough patience for very simple operations. And this presents a problem. There are basically only three available local (i.e., shortrange) wireless interfaces (LWI): WLAN, IR, and Bluetooth, which do not cover the whole range of mobile devices. WLAN has not gained popularity yet, while IR is gradually disappearing. Bluetooth is the most frequently used of them, but still it is not available in all phones. For every particular device it is possible to send a token out using some combination of LWI and presentation technology, but there is no common and easy-to-use combination. This is a threshold for the development of services. Taking a deeper look at the mobile devices, we can find one more non-standard simplex LWI, which is present in all devices—acoustical, where the transmitter is a phone ringer. Token presentation through acoustic interface along with general solution of token delivery via SIM Toolkit technology
(see 3GPP TS, 1999) was presented by Khashchanskiy and Kustov (2001). However, mobile operators have not taken SIM Toolkit into any serious use, and the only alternative way of delivering sound tokens into the phone-ringing tone customization technology was not available for a broad range of devices at the time the aforementioned paper was published. Quite unexpectedly, recent development of mobile phone technologies gives a chance for sound tokens to become a better solution for the aforementioned problem, compared with other LWI. Namely, it can be stated that every contemporary mobile device supports either remote customization of ringing tones, or MMS, and in the majority of cases, even both, thus facilitating sound token receiving over the air. Most phone models can playback a received token with only a few button-clicks. Thus, a sound token-based solution meets the set criteria better than any other LWI. Token delivery works the same way for virtually all phones, and token presentation is simple. In this article we study the sound token solution practical implementation in detail. First, we select optimal modulation, encoding, and recognition algorithm, and we estimate data rate. Then we present results of experimental verification.
ACOUSTIC DATA CHANNEL We consider the channel being as follows. The transmitter is a handset ringer; information is encoded as a sequence of sine wave pulses, each with specified frequency and amplitude. Multimedia message sounds and most ringing tones are delivered as sequences of events in MIDI (musical instrument digital interface) format. A basic pair of MIDI events (note on and note off) defines amplitude, frequency, duration of a note, and the instrument that plays this note. MIDI events can be used to produce information-bearing sound pulses with specified frequency and amplitude. Widely used support of polyphonic MIDI sequences allows playback of several notes simultaneously. Nonetheless, this has been proved worthless because in order to get
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
A
Acoustic Data Communication with Mobile Devices
Figure 1. Frequency response measured with test MIDI sequence in hold-max mode 0 dB
-0 dB 0
distinguished, these notes have to belong to different nonoverlapping frequency ranges. Then the bit rate that can be achieved would be the same as if wider frequency range was allocated for a single note. The receiver is a microphone; its analog sound signal is digitized and information is decoded from the digital signal by recognition algorithm, based on fast fourier transform (FFT) technique. FFT is, in our opinion, a reasonable tradeoff between efficiency and simplicity. We investigated acoustics properties of mobile devices. After preliminary comparison of a few mobile phone models, we found that ringer quality is of approximately the same level. All handsets have a high level of harmonic distortions and poor frequency response. The results shown in Figures 1 and 2 are obtained for a mid-class mobile phone SonyEricsson T630 and are close to average. MIDI-based sound synthesis technology applies limitations on pulse magnitude, frequency, and duration. At the same time, ringer frequency response is not linear and the level of harmonic distortions is very high. Figure 1 shows frequency response measured with a sweeping tone or, to be precise, a tone leaping from one musical note to another. To obtain this, the phone played a MIDI sequence of nonoverlapping in-time notes that covered a frequency range from 263 to 4200 Hz (gray area). The frequency response varies over a 40 dB range, reaching its maximum for frequencies from approximately 2.5 to 4 kHz. Moreover, spectral components stretch up to 11 KHz, which is caused by harmonic distortions. This is illustrated also by Figure 2. Horizontal axis is time; overall duration of the test sequence is 15 seconds. Vertical axis is sound frequency, which is in range from 0 to 11025 Hz. Brightness is proportional to sound relative spectral density; its dynamic range is 60 dB, from black to white.
0 Hz
We also found that frequency of the same note may differ in different handsets. Nevertheless, the ratio of note frequencies (musical intervals) remains correct, otherwise melodies would sound wrong. For a simplex channel with such poor parameters, as reliable a data encoding method as possible is to be used. Frequency shift keying (FSK) is known as the most reliable method which finds its application in channels with poor signal-to-noise ratio (SNR) and non-linear frequency response. It is not possible to negotiate transfer rate or clock frequency, as it is usually done in modem protocols because acoustic channel is simplex. To make the channel as adaptive as possible, we have chosen to use differential FSK (DFSK), as it requires no predefined clock frequency. Instead, frequency leaps from one pulse to another provide the channel clocking. The difference between frequencies of consecutive pulses determines the encoded value. Once encoding scheme is selected, let us estimate possible transfer rate before we can find the balance between data
Figure 2. A spectrogram of the test MIDI sequence
Acoustic Data Communication with Mobile Devices
transfer rate and channel reliability. Suppose the transmitter generates a sequence of pulses of duration t, which follow without gaps with repetition frequency f. If each frequency leap between two consequent pulses carries N bits of information, the overall bit rate p is obviously: p=N•f.
(1)
In DFSK, for each frequency leap to carry N bits, we must be able to choose pulse frequencies from a set of 2N +1values. If a pulse frequency can have n values, we will have p = [log2(n-1)] • f ,
(2)
where by [] we denote integer part. It follows from (1, 2), that to increase p, we must increase pulse repetition frequency f and the amount of possible values for pulse carrier frequencies n. However, if the recognition is based on spectral analysis, we cannot increase n and f independently. Let us show it. Assume for simplicity that pulse frequency can have any value within frequency range F. Then the number of available values of coding frequencies will be n = [log2(F/∆f - 1)],
(3)
where ∆f is the minimal shift of pulse frequency between two consecutive pulses. Maximum n is achieved with maximum F and minimum ∆f. Both parameters have their own boundaries. Bandwidth is limited by the ringer capabilities, and frequency shift is dependant on pulse repetition frequency f, due to the fundamental rule of spectral analysis (Marple, 1987), which defines frequency resolution df to be in reverse proportionality to observation time T: df = 1 / T.
(4)
How can (4) be understood in our case of a sequence of pulses? Having converted the signal into frequency domain, we will get the sequence of spectra. As information is encoded in the frequency pulses, we must determine the pulse frequency for every spectrum. This can only be done with certain accuracy df called frequency resolution. The longer time T we observe the signal, the better frequency resolution is. So for given pulse duration t, equation (4) sets the lower limit for frequency difference ∆f between two consecutive pulses: ∆f ≥ 1 / t = f.
Let us now try to estimate the data rate for the system we studied earlier. Figures 1 and 2 show that harmonic distortions are very high, and second and third harmonics often have higher magnitudes than the main tone. Consequently, the coding frequencies must belong to the same octave. Their frequency separation should be no less than defined by (5). An octave contains 12 semitones, so possible frequency values fi are defined by the following formula: fi = f0 • 2i/12, i=0...11.
The minimum spacing between consecutive notes is for i=1; maximum for i=11. In our case, we decided to use the fourth octave—as the closest to the peak area of phone ringer frequency response—in order to maximize SNR and thus make recognition easier. For it, f0= 2093 Hz, and minimum spacing between notes is 125 Hz. Taking the maximum amount of N = 3 (9 coding frequencies), we can estimate transfer rate as: pmax = 3 • 125 = 375 bps.
(7)
Recognition Algorithm (Demodulation) The following algorithm was developed to decode information transferred through audio channel. Analog audio signal from the microphone is digitized with sampling frequency Fs satisfying Nyquist theorem (Marple, 1987). A signal of duration Ts is then represented as a sequence of Ts/Fs samples. FFT is performed on a sliding vector of M signal samples, where M is a power of 2. •
• •
(5)
This means, that if we increase pulse repetition rate f, then we have to correspondingly increase frequency separation ∆f for the consecutive pulses; otherwise the spectral analysis-based recognizing device will not principally be able to detect signal.
(6)
•
First, sequence of instant power spectra is obtained from the signal using discrete Fourier transform with sliding window (vector) of M samples. To get consecutive spectra overlapped by 50%, the time shift between them was taken M/2Fs. Overlapping is needed to eliminate the probability of missing the proper position of a sliding window corresponding to the pulse existence interval, when the pulse duration is not much longer than analysis time significantly (at least twice). Second, the synchronization sound is found as sine wave with a constant, but not known in advance frequency, and a certain minimum duration. Third, the spectrum composed of maximum values over the spectra sequence (so-called hold-max spectrum) is used to find the pulse carrier frequencies. This step relies on the assumption that used frequency range does not exceed one octave. In other words, the highest frequency is less than twice the value of the lowest one. Forth, time cross-sections of spectra sequence at found carrier frequencies are used to recognize moments of sound pulse appearances.
A
Acoustic Data Communication with Mobile Devices
•
The last step is reconstruction of encoded bit sequence having the time-ordered set of frequency leaps.
Figure 3. Encoded “hello world”; note the leading synchronization header. Overall duration is approximately 2 seconds.
Such an algorithm does not need feedback and can work with unknown carrier frequencies in unknown but limited frequency range. Recognizing the beginning of the transmission is critical for the correct work, so we added “synchronization header” in the beginning of the signal. The length of this header is constant, so the throughput of the system will rise with the message length.
Recognizer Parameters Here we explain how the parameters of analyzer (Fs, M) are defined from that of signal (f, ∆f). After FFT, we have M/2 of complex samples in frequency domain, corresponding to frequency range from 0 to Fs/2. So for this particular case, frequency resolution obviously equals the difference between the consecutive samples in the frequency domain; namely, df = Fs / M.
(7)
According to (4), minimum required time of analysis is T = M / Fs.
(8)
It is obvious, that T must not exceed burst duration t. Combining (8) and (5), we get: M / Fs ≤ 1 / f
(9)
On the other hand, frequency resolution df must not exceed spacing ∆f between carrier frequencies: Fs / M ≤ ∆f
(10)
Combining (9) and (10), we will finally get: f ≤ Fs / M ≤ ∆f
(11)
which shows that values of analyzer parameters may be restricted when (5) is close to the equation. This imposes requirements on the sound recognition algorithm to work reliably nearby the “critical points,” where the recognition becomes principally impossible.
EXPERIMENTAL RESULTS We implemented a prototype of acoustic data channel with the mobile phone SonyEricsson T630, whose characteristics are seen in Figures 1 and 2. For encoding, we developed software that encoded symbol strings in ASCII to melody played by an electric
organ. The instrument was chosen from 127 instruments available in MIDI format, because its sound is the closest to the sine wave pulses model we used in calculations. It is maintained at approximately the same level over the whole note duration. The recognizer consisted of a Sony ECM-MS907 studio microphone for signal recording, and a conventional PC with a sound card was used for signal analysis. FFT processing was done by our own software. In the beginning of our experiments, we used the parameters described in the theoretical section. Later we found that at the highest possible transfer rate, data recognition is not reliable. So we gradually increased pulse duration until recognition became reliable. Eventually we selected the following modulation parameters: n=5 (each frequency leap carries two data bits), notes were evenly distributed over the octave (C, D#, F, G, A in musical notation, and they correspond to frequencies 2093, 2489, 2794, 3136, and 3520 Hz), and pulse duration was 46 ms. Figure 3 shows a spectrogram of recognizable signal from the microphone. Horizontal axis is time, and overall signal duration is 2 seconds. Vertical axis is frequency, and one can see the leaps between consecutive pulses. Brightness is proportional to the signal intensity. This example signal carries 88 bits of information (a string “hello world,” coded as 11 ASCII characters), which makes the data transfer rate approximately 40 bps. Overhead from the synchronization header is ca. 25%; for longer messages the average transfer rate would be higher.
DISCUSSION We have managed to implement a reliable data channel from the phone; the advantage of the proposed recognition algo-
Acoustic Data Communication with Mobile Devices
rithm is that it can work in the same way for every mobile device, independent on acoustic properties of different brands and models, although encoding frequencies are different. The channel is principally one way: the handset cannot receive any feedback that can be used, for example, for error correction. Nevertheless, developed recognition algorithm provided good reliability. For a handset placed 30 cm from the microphone, in a room environment, recognition was 100% reliable. This condition corresponds to the output of the average phone in a “normal” room environment. Ensuring reliability does not seem to be a very difficult task. First of all, SNR can be improved by increasing the number of receiving microphones. On the other hand, in practical systems simple shielding is very easy to implement. And finally, even one error in recognition is not fatal: the user can always have another try. A recognizing device can easily identify cases of unsuccessful recognition and indicate the former case for the user to retry. The recognition system can be implemented on any PC equipped with a sound card. The algorithm is so simple that the system can also be implemented as an embedded solution based on digital signal processors. Microphone requirements are not critical either: both the frequency response and SNR of entry level microphones are much better than those of mobile device ringers. This means that cheap stand-alone recognizers can be implemented and deployed at the points of service. It is interesting to note that other devices capable of playing MIDI sequences (e.g., PDAs) can be used as well as mobile phones. Measured transfer rate (40 bps) was considerably less than the estimation, obtained in our simple model—375 bps. We think that the reason for this was slow pulse decay rate in combination with non-linear frequency response. Amplitude of the note with frequency close to a local frequency response maximum might remain higher than amplitude of the consecutive note through the whole duration of the latter. Thus, the weaker sound of the second note might be not recognized. However, we consider even such relatively slow transmission still suitable for the purposes of mobile authorization applications, because authorization data is usually small and its transmission time is not critical. Our example (Figure 3) seems to be a quite practical situation—transmitting 11-symbol password during 2s is definitely not too long for a user. Typing the same token on the vending machine keyboard would easily take twice as long. The acoustic presentation method might be an attractive feature for teenagers (e.g., mobile cinema tickets being one conceivable application).
ACKNOWLEDGMENTS The authors would like to thank Petteri Koponen for the original idea.
REFERENCES Khashchanskiy, V., & Kustov, A. (2001). Universal SIM Toolkit-based client for mobile authorization system. Proceedings of the 3rd International Conference on Information Integration and Web-Based Applications & Services (IIWAS 2001) (pp. 337-344). Marple, S. Lawrence Jr. (1987). Digital spectral analysis with applications. Englewood Cliffs, NJ: Prentice-Hall. 3GPP TS 11.14. (1999). Specification of the SIM application toolkit for the Subscriber Identity Module-Mobile Equipment (SIM-ME) interface. Retrieved from http://www.3gpp. org/ftp/Specs/html-info/1114.htm
KEY TERMS Fast Fourier Transform (FFT): An optimized form of the algorithm that calculates a complex spectrum of digitized signals. It is most widely used to obtain a so-called power spectrum as a square of a complex spectrum module. Power spectrum represents energy distribution along frequency axis. Frequency Resolution: The minimum difference in frequencies which can be distinguished in a signal spectrum. Frequency Response: For a device, circuit, or system, the ratio between output and input signal spectra. Frequency Shift Keying (FSK): The digital modulation scheme that assigns fixed frequencies to certain bit sequences. Differential FSK (DFSK) uses frequency differences to encode bit sequences. Harmonic Distortions: Alteration of the original signal shape caused by the appearance of higher harmonics of input signal at the output. IR: Short-range infrared communication channel. Musical Instrument Digital Interface (MIDI): A standard communications protocol that transfers musical notes between electronic musical instruments as sequences of events, like ‘Note On’, ‘Note Off’, and many others. Sampling Frequency: The rate at which analogue signal is digitized by an analogue-to-digital converter (ADC) in order to convert the signal into numeric format that can be stored and processed by a computer.
A
0 Category: Mobile Multimedia
Adaptive Transmission of Multimedia Data over UMTS Antonios Alexiou Patras University, Greece Dimitrios Antonellis Patras University, Greece Christos Bouras Patras University, Greece
INTRODUCTION As communications technology is being developed, users’ demand for multimedia services raises. Meanwhile, the Internet has enjoyed tremendous growth in recent years. Consequently, there is a great interest in using the IP-based networks to provide multimedia services. One of the most important areas in which the issues are being debated is the development of standards for the universal mobile telecommunications system (UMTS). UMTS constitutes the third generation of cellular wireless networks which aims to provide high-speed data access along with real-time voice calls. Wireless data is one of the major boosters of wireless communications and one of the main motivations of the next-generation standards. Bandwidth is a valuable and limited resource for UMTS and every wireless network in general. Therefore, it is of extreme importance to exploit this resource in the most efficient way. Consequently, when a user experiences a streaming video, there should be enough bandwidth available at any time for any other application that the mobile user might need. In addition, when two different applications run together, the network should guarantee that there is no possibility for any of the above-mentioned applications to prevail against the other by taking all the available channel bandwidth. Since Internet applications adopt mainly TCP as the transport protocol, while streaming applications mainly use RTP, the network should guarantee that RTP does not prevail against the TCP traffic. This means that there should be enough bandwidth available in the wireless channel for the Internet applications to run properly.
BACKGROUND Chen and Zachor (2004) propose a widely accepted rate control method in wired networks which is the equationbased rate control, also known as TFRC (TCP-friendly rate control). In this approach the authors use multiple TFRC
connections as an end-to-end rate control solution for wireless streaming video. Another approach is presented by Fu and Liew (2003). As they mention, TCP Reno treats the occurrence of packet losses as a manifestation of network congestion. This assumption may not apply to networks with wireless channels, in which packet losses are often induced by noise, link error, or reasons other than network congestion. Equivalently, TCP Vegas uses queuing delay as a measure of congestion (Choe & Low, 2003). Thus, Fu and Liew (2003) propose an enhancement of TCP Reno and TCP Vegas for the wireless networks, namely TCP Veno. Chen, Low, and Doyle (2005) present two algorithms that formulate resource allocation in wireless networks. These procedures constitute a preliminary step towards a systematic approach to jointly design TCP congestion control algorithms, not only to improve performance, but more importantly, to make interaction more transparent. Additionally, Xu, Tian, and Ansari (2005) study the performance characteristics of TCP New Reno, TCP SACK, TCP Veno, and TCP Westwood under the wireless network conditions and they propose a new TCP scheme, called TCP New Jersey, which is capable of distinguishing wireless packet losses from congestion. Recent work provides an overview of MPEG-4 video transmission over wireless networks (Zhao, Kok, & Ahmad, 2004). A critical issue is how we can ensure the QoS of video-based applications to be maintained at an acceptable level. Another point to consider is the unreliability of the network, especially of the wireless channels, because we observe packet losses resulting in a reduction of the video quality. The results demonstrate that the video quality can be substantially improved by preserving the high-priority video data during the transmission.
THE TCP-FRIENDLY RATE CONTROL PROTOCOL TFRC is not actually a fully specified end to-end transmission protocol, but a congestion control mechanism that is designed
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Adaptive Transmission of Multimedia Data over UMTS
Figure 1. Typical scenario for streaming video over UMTS
to operate fairly along with TCP traffic. Generally TFRC should be deployed with some existing transport protocol such as UDP or RTP in order to present its useful properties (Floyd, Handley, Padhye, & Widmer, 2000). The main idea behind TFRC is to provide a smooth transmission rate for streaming applications. The other properties of TFRC include slow response to congestion and the opportunity of not aggressively trying to make up with all available bandwidth. Consequently, in case of a single packet loss, TFRC does not halve its transmission rate like TCP, while on the other hand it does not respond rapidly to the changes in available network bandwidth. TFRC has also been designed to behave fairly when competing for the available bandwidth with concurrent TCP flows that comprise the majority of flows in today’s networks. A widely popular model for TFRC is described by the following equation (Floyd & Fall, 1999): T=
kS RTT p
(1)
T represents the sending rate, S is the packet size, RTT is the end-to-end round trip time, p is the end-to-end packet loss rate, and k is a constant factor between 0.7 and 1.3 (Mahdavi & Floyd, 1997) depending on the particular derivation of equation (1). The equation describes TFRC’s sending rate as a function of the measured packet loss rate, round-trip time, and used packet size. More specifically, a potential congestion in the nodes of the path will cause an increment in the packet loss rate and in the round trip time according to the current packet size. Given this fluctuation, it is easy to determine the new transmission rate so as to avoid congestion and packet losses. Generally, TFRC’s congestion control consists of the following mechanisms: 1. 2.
The receiver measures the packet loss event rate and feeds this information back to the sender. The sender uses these feedback messages to calculate the round-trip-time (RTT) of the packets.
A
3.
The loss event rate and the RTT are then fed into the TRFC rate calculation equation (described later in more detail) in order to find out the correct data sending rate.
ANALYSIS OF THE TFRC MECHANISM FOR UMTS The typical scenario for streaming video over UMTS is shown in Figure 1, where the server is denoted by Node1 and the receiver by UE1. The addressed scenario comprises a UMTS radio cell covered by a Node B connected to an RNC. The model consists of a UE connected to DCH, as shown in Figure 1. In this case, the DCH is used for the transmission of the data over the air. DCH is a bi-directional channel and is reserved only for a single user. The common channels are the forward access channel (FACH) in the downlink and the random access channel (RACH) in the uplink. The wireless link is assumed to have available bandwidth BW, and packet loss rate pw, caused by wireless channel error. This implies that the maximum throughput that could be achieved in the wireless link is BW (1 – pW). There could also be packet loss caused by congestion at wired nodes denoted by pnode name (node name: GGSN, SGSN, RNC, Node B). The end-to-end packet loss rate observed by the receiver is denoted as p. The streaming rate is denoted by T. This means that the streaming throughput is T (1 - p). Under the above assumptions we characterize the wireless channel as underutilized if T (1 - p) < BW (1 – pW). Given the above described scenario, the following are assumed: 1. 2.
The wireless link is the long-term bottleneck. This means that there is no congestion due to streaming traffic to the nodes GGSN, SGSN, and RNC. There is no congestion at Node B due to the streaming application, if and only if the wireless bandwidth is underutilizedthat is, T (1 - p) < BW (1 – pW). This also implies that no queuing delay is caused at Νode
Adaptive Transmission of Multimedia Data over UMTS
3. 4.
B and hence, the round trip time for a given route has the minimum value (i.e., RTTmin). Thus, this assumption can be restated as follows: for a given route, RTT = RTTmin if and only if T (1 - p) ≤ BW (1 – pW). This in turn implies that if T (1 - p) > BW (1 – pW) then RTT ≥ RTTmin. The packet loss rate caused by wireless channel error (pW) is random and varies from 0 to 0.16. The backward route is error-free and congestionfree.
The communication between the sender and the receiver is based on RTP/RTCP sessions; and the sender, denoted by Node 1 (Figure 1), uses the RTP protocol to transmit the video stream. The client, denoted by UE1 (Figure 1), uses the RTCP protocol in order to exchange control messages. The mobile user in recurrent time space sends RTCP reports to the media server. These reports contain information about the current conditions of the wireless link during the transmission of the multimedia data between the server and the mobile user. The feedback information contains the following parameters: • •
Packet Loss Rate: The receiver calculates the packet loss rate during the reception of sender data, based on RTP packets sequence numbers. Timestamp of Every Packet Arrived at the Mobile User: This parameter is used by the server for the RTT calculation of every packet.
The media server extracts the feedback information from the RTCP report and passes it through an appropriate filter. The use of filter is essential for the operation of the mechanism in order to avoid wrong estimations of the network conditions. On the sender side, the media server using the feedback information estimates the appropriate rate of the streaming video so as to avoid network congestion. The appropriate transmission rate of the video sequence is calculated from equation (1), and the media server is responsible for adjusting the sending rate with the calculated value. Obviously, the media server does not have the opportunity to transmit the video in all the calculated sending rates. However, it provides a small variety of them and has to approximate the calculated value choosing the sending rate from the provided transmission rates. This extends the functionality of the whole congestion control mechanism. More specifically, the sender does not have to change the transmission rate every time it calculates a new one with a slight difference from the previous value. Consequently, it changes the transmission rate of the multimedia data to one of the available sending rates of the media server as has already been mentioned. In this approach, the number of the changes in the sending rate is small and the mobile user does not deal with a continually different transmission rate.
As mentioned above, it is essential to keep a history of the previous calculated values for the transmission rate. Having this information, the media server can estimate the smoothed transmission rate, using the m most recent values of the calculated sending rate from the following equation:
T
Smoothed
∑ =
m i =1
wi ⋅ TmSmoothed +1−i
∑
m i =1
wi
(2)
The value m, used in calculating the transmission rate, determines TFRC’s speed in responding to changes in the level of congestion (Handley, Floyd, Padhye, & Widmer, 2003). The weights wi are appropriately chosen so that the most recent calculated sending rates receive the same high weights, while the weights gradually decrease to 0 for older calculated values. Equivalently to the calculation of the transmission rate, the mobile user (client) measures the packet loss rate p1 based on the RTP packets sequence numbers. This information is sent to the media server via the RTCP reports. In order to prevent a single spurious packet loss having an excessive effect on the packet loss estimation, the server smoothes the values of packet loss rate using the filter of the following equation, which computes the weighted average of the m most recent loss rate values (Vicisiano, Rizzo, & Crowcroft, 1998). p1Smoothed
∑ =
m i =1
wi ⋅ p1,Smoothed m +1−i
∑
m i =1
wi
(3)
The value of p1Smoothed is then used by equation (1) for the estimation of the transmission rate of the multimedia data. The weights wi are chosen as in the transmission rate estimation.
FUTURE TRENDS The most prominent enhancement of the adaptive real-time applications is the use of multicast transmission of the multimedia data. The multicast transmission of multimedia data has to accommodate clients with heterogeneous data reception capabilities. To accommodate heterogeneity, the server may transmit one multicast stream and determine the transmission rate that satisfies most of the clients (Byers et al., 2000). Additionally, Vickers, Albuquerque, and Suda (1998) present different approaches where the server transmits multiple multicast streams with different transmission rates allocating the client at these streams, as well as using layered encoding and transmitting each layer to a different
Adaptive Transmission of Multimedia Data over UMTS
multicast stream. An interesting survey of techniques for multicast multimedia data over the Internet is presented in Li, Ammar, and Paul (1999). Single multicast stream approaches have the disadvantage that clients with a low-bandwidth link will always get a high-bandwidth stream if most of the other members are connected via a high-bandwidth link, and the same is true the other way around. This problem can be overcome with the use of a multi-stream multicast approach. Single multicast stream approaches have the advantages of easy encoder and decoder implementation and simple protocol operation, due to the fact that during the single multicast stream approach, there is no need for synchronization of clients’ actions (as the multiple multicast streams and layered encoding approaches require). The subject of adaptive multicast of multimedia data over networks with the use of one multicast stream has engaged researchers all over the world. During the adaptive multicast transmission of multimedia data in a single multicast stream, the server must select the transmission rate that satisfies most the clients with the current network conditions. Totally, three approaches can be found in the literature for the implementation of the adaptation protocol in a single stream multicast mechanism: equation based (Rizzo, 2000; Widmer & Handley, 2001), network feedback based (Byers et al., 2000), or a combination of the above two approaches (Sisalem & Wolisz, 2000).
tional Workshop on Networked Group Communication (pp. 71-81), Palo Alto, CA.
CONCLUSION
Mahdavi, J., & Floyd, S. (1997). TCP-Friendly unicast rate-based flow control. Retrieved from http://www.psc. edu/networking/papers/tcp_friendly.html
An analysis of the TCP friendly rate control mechanism for UMTS has been presented. The TFRC mechanism gives the opportunity to estimate the appropriate transmission rate of the video data for avoiding congestion in the network. The three goals of this rate control could be stated as follows. First, the streaming rate does not cause any network instability (i.e., congestion collapse). Second, TFRC is assumed to be TCP Friendly, which means that any application that transmits data over a network presents friendly behavior towards the other flows that coexist in the network and especially towards the TCP flows that comprise the majority of flows in today’s networks. Third, it leads to the optimal performancethat is, it results in the highest possible throughput and lowest possible packet loss rate. Furthermore, an overview of video transmission over UMTS using real-time protocols such as RTP/RTCP has been presented.
REFERENCES Byers, J., Frumin, M., Horn, G., Luby, M., Mitzenmacher, M., Roetter, A., & Shaver, W. (2000). FLID-DL: Congestion control for layered multicast. Proceedings of the Interna-
Chen, M., & Zachor, A. (2004). Rate control for streaming video over wireless. IEEE INFOCOM, Hong Kong, China, (pp. 1181-1190). Chen, L., Low, S., & Doyle, J. (2005). Joint congestion control and media access control design for ad hoc wireless networks. IEEE INFOCOM, Miami, FL. Choe, H., & Low, S. (2003). Stabilized Vegas. IEEE INFOCOM, 22(1), 2290-2300. Floyd, S., Handley, M., Padhye, J., & Widmer, J. (2000). Equation-based congestion control for unicast applications. Proceedings of ACM SIGCOMM, Stockholm, Sweden, (pp. 43-56). Floyd, S., & Fall, K. (1999). Promoting the use of end-to-end congestion control in the Internet. IEEE/ACM Transactions on Networking, 7(4), 458-472. Fu, C. P., & Liew, S. C. (2003). TCP Veno: TCP enhancement for transmission over wireless access networks. IEEE Journal on Selected Areas in Communications, 21(2), 216-228. Handley, M., Floyd, S., Padhye, J., & Widmer, J. (2003). TCP Friendly Rate Control (TFRC). RFC, 3448. Li, X., Ammar, M., & Paul, S. (1999). Video multicast over the Internet. IEEE Network Magazine, 12(2), 46-60.
Rizzo, L. (2000). pgmcc: A TCP-friendly single-rate multicast congestion control scheme. Proceedings of ACM SIGCOMM, Stockholm, Sweden, (pp. 17-28). Sisalem, D., & Wolisz, A. (2000). LDA+ TCP-Friendly adaptation: A measurement and comparison study. Proceedings of the International Workshop on Network and Operating Systems Support for Digital Audio and Video (NOSSDAV), Chapel Hill, NC. Vicisiano, L., Rizzo, L., & Crowcroft, J. (1998). TCP-like congestion control for layered multicast data transfer. IEEE INFOCOM, San Francisco, CA, (pp. 996-1003). Vickers, J., Albuquerque, N., & Suda, T. (1998). Adaptive multicast of multi-layered video: Rate-based and creditbased approaches. IEEE INFOCOM, San Francisco, CA, (pp. 1073-1083). Widmer, J., & Handley, M. (2001). Extending equation-based congestion control to multicast applications. Proceedings of ACM SIGCOMM, San Diego, CA, (pp. 275-286).
A
Adaptive Transmission of Multimedia Data over UMTS
Xu, K., Tian, Y., & Ansari, N. (2005). Improving TCP performance in integrated wireless communications networks. Computer Networks, Science Direct, 47(2), 219-237.
Delay Jitter: The mean deviation (smoothed absolute value) of the difference in packet spacing at the receiver compared to the sender for a pair of packets.
Zhao, J., Kok, C., & Ahmad, I. (2004). MPEG-4 video transmission over wireless networks: A link level performance study. Wireless Networks, 10(2), 133-146.
Frame Rate: The rate of the frames, which are encoded by video encoder.
KEY TERMS Adaptive Real-Time Application: An application that has the capability to transmit multimedia data over heterogeneous networks and adapt media transmission to network changes.
Multimedia Data: Data that consist of various media types like text, audio, video, and animation. Packet Loss Rate: The fraction of the total transmitted packets that did not arrive at the receiver. RTP/RTCP: Protocol used for the transmission of multimedia data. The RTP performs the actual transmission, and the RTCP is the control and monitoring transmission.
Category: M-Business and M-Commerce
Addressing the Credibility of Mobile Applications Pankaj Kamthan Concordia University, Canada
INTRODUCTION Mobile access has opened new vistas for various sectors of society including businesses. The ability that anyone using (virtually) any device could be reached anytime and anywhere presents a tremendous commercial potential. Indeed, the number of mobile applications has seen a tremendous growth in the last few years. In retrospect, the fact that almost anyone can set up a mobile application claiming to offer products and services raises the question of credibility from a consumer’s viewpoint. The obligation of establishing credibility is essential for an organization’s reputation (Gibson, 2002) and for building consumers’ trust (Kamthan, 1999). If not addressed, there is a potential for lost consumer confidence, thus significantly reducing the advantages and opportunities the mobile Web as a medium offers. If a mobile application is not seen as credible, we face the inevitable consequence of a product, however functionally superior it might be, rendered socially isolated. The rest of the article is organized as follows. We first provide the motivational background necessary for later discussion. This is followed by introduction of a framework within which different types of credibility in the context of mobile applications can be systematically addressed and thereby improved. Next, challenges and directions for future research are outlined. Finally, concluding remarks are given.
BACKGROUND In this section, we present the fundamental concepts underlying credibility, and present the motivation and related work for addressing credibility within the context of mobile applications.
Basic Credibility Concepts For the purposes of this article, we will consider credibility to be synonymous to (and therefore interchangeable with) believability (Hovland, Janis, & Kelley, 1953). We follow the terminology of Fogg and Tseng (1999), and view credibility
and trust as being slightly different. Since trust indicates a positive belief about a person, object, or process, we do not consider credibility and trust to be synonymous. It has been pointed out in various studies (Fogg, 2003; Metzger, 2005) that credibility consists of two primary dimensions, namely trustworthiness and expertise of the source of some information. Trustworthiness is defined by the terms such as well-intentioned, truthful, unbiased, and so on. The trustworthiness dimension of credibility captures the perceived goodness or morality of the source. Expertise is defined by terms such as knowledgeable, experienced, competent, and so on. The expertise dimension of credibility captures the perceived knowledge and skill of the source. Together, they suggest that “highly credible” mobile applications will be perceived to have high levels of both trustworthiness and expertise. We note that trustworthiness and expertise are at such a high level of abstraction that direct treatment of any of them is difficult. Therefore, in order to improve credibility, we need to find quantifiable attributes that can improve each of these dimensions.
A Classification of Credibility The following taxonomy helps associating the concept of credibility with a specific user class in context of a mobile application. A user could consider a mobile application to be credible based upon direct interaction with the application (active credibility), or consider it to be credible in absence of any direct interaction but based on certain pre-determined notions (passive credibility). Based on the classification of credibility in computer use (Fogg & Tseng, 1999) and adapting them to the domain of mobile applications, we can decompose these further. There can be two types of active credibility: (1) surface credibility, which describes how much the user believes the mobile application is based on simple inspection; and (2) experienced credibility, which describes how much the user believes the mobile application is based on first-hand experience in the past. There can be two types of passive credibility: (1) presumed credibility, which describes how much the user believes the mobile application because of general assumptions that the user holds; and (2) reputed credibility, which describes how
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
A
Addressing the Credibility of Mobile Applications
much the user believes the mobile application because of a reference from a third party. Finally, credibility is not absolute with respect to users and with respect to the application itself (Metzger, Flanagin, Eyal, Lemus, & McCann, 2003). Also, credibility can be associated with a whole mobile application or a part of a mobile application. For example, a user may question the credibility of information on a specific product displayed in a mobile application. We contend that for a mobile application to be labeled non-credible, there must exist at least a part of it that is labeled non-credible based on the above classification by at least one user.
The Origins and Significance of the Problem of Mobile Credibility The credibility of mobile applications deserves special attention for the following reasons: •
•
•
Delivery Context: Mobile applications are different from the desktop or Web environments (Paavilainen, 2002) where context-awareness (Sadeh, Chan, Van, Kwon, & Takizawa, 2003) is a unique challenge. The delivery context in a changing environment of mobile markup languages, variations in user agents, and constrained capabilities of mobile devices presents unique challenges towards active credibility. Legal Context: Since the stakeholders of a mobile application need not be co-located (different jurisdictions in the same country or in different countries), the laws that govern the provider and the user may be different. Also, the possibilities of fraud such as computer domain name impersonation (commonly known as “pharming”) or user identity theft (commonly known as “phishing”) with little legal repercussions for the perpetrators is relatively high in a networked environment. These possibilities can impact negatively on presumed credibility. User Context: Users may deploy mobile devices with varying configurations, and in the event of problems with a mobile service, may first question the provider rather than the device that they own. In order for providers of mobile portals to deliver user-specific information and services, they need to know details about the user (such as profile information, location, and so on). This creates the classical dichotomy between personalization and privacy, and striking a balance between the two is a constant struggle for businesses (Kasanoff, 2002). The benefits of respecting one can adversely affect the other, thereby impacting their credibility in the view of their customers. Furthermore, the absence of a human component from non-proximity or “facelessness” of the provider can shake customer
confidence and create negative perceptions in a time of crisis such as denial of service or user agent crash. These instances can lead to a negative passive credibility.
Initiatives for Improving Mobile Credibility There have been initiatives to address the credibility of Web applications such as a user survey to identify the characteristics that users consider necessary for a Web application to be credible (Fogg et al., 2001) and a set of guidelines (Fogg, 2003) for addressing surface, experienced, presumed, and reputed credibility of Web applications. However, these efforts are limited by one or more of the following issues. The approach towards ensuring and/or evaluating credibility is not systematic, the proposed means for ensuring credibility is singular (only guidelines), and the issue of feasibility of the means is not addressed. Moreover, these guidelines are not specific to mobility, are not prioritized and the possibility that they can contradict each other is not considered, can be open to broad interpretation, and are stated at such a high level that they may be difficult to realize by a novice user.
ADDRESSING THE CREDIBILITY OF MOBILE APPLICATIONS In this section, we consider approaches for understanding and improving active credibility of mobile applications.
A Framework for Addressing Active Credibility of Mobile Applications To systematically address the active credibility of mobile applications, we take the following steps: 1. 2. 3.
View credibility as a qualitative aspect and address it indirectly via quantitative means. Select a theoretical basis for communication of information (semiotics), and place credibility in its setting. Address semiotic quality in a practical manner.
Based on this and using the primary dimensions that affect credibility, we propose a framework for active credibility of mobile applications (see Table 1). The external attributes (denoted by E) are extrinsic to the software product and are directly a user’s concern, while internal attributes (denoted by I) are intrinsic to the software product and are directly an engineer’s concern. Since not all attributes corresponding to a semiotic level are at the same echelon, the different tiers are denoted by “Tn.”
Addressing the Credibility of Mobile Applications
Table 1. A semiotic framework for active credibility of mobile applications Semiotic Level
Quality Attributes
Means for Credibility Assurance and Evaluation
Decision Support
“Expert” Knowledge (Principles, Guidelines, Patterns) Inspections Testing Metrics Tools
Feasibility
A
Credibility Social
Aesthetics, Legality, Privacy, Security, (Provider) Transparency [T5;E] Accessibility, Usability [T4;E]
Pragmatic
Interoperability, Portability, Reliability, Robustness [T3;E]
Semantic
Completeness and Validity [T2;I]
Syntactic
Correctness [T1;I]
We now describe each of the components of the framework in detail.
Semiotic Levels The first column of Table 1 addresses semiotic levels. Semiotics (Stamper, 1992) is concerned with the use of symbols to convey knowledge. From a semiotics perspective, a representation can be viewed on six interrelated levels: physical, empirical, syntactic, semantic, pragmatic, and social, each depending on the previous one in that order. The physical and empirical levels are concerned with the physical representation of signs in hardware and communication properties of signs, and are not of direct concern here. The syntactic level is responsible for the formal or structural relations between signs. The semantic level is responsible for the relationship of signs to what they stand for. The pragmatic level is responsible for the relation of signs to interpreters. The social level is responsible for the manifestation of social interaction with respect to signs.
Quality Attributes The second column of Table 1 draws the relationship between semiotic levels and corresponding quality attributes. Credibility belongs to the social level and depends on the layers beneath it. The external quality attributes legality, privacy, security, and (provider) transparency also at the social level depend upon the external quality attributes accessibility and usability at the pragmatic level, which in turn depend upon the external quality attributes interoperability, performance, portability, reliability, and robustness also at the pragmatic level. (We note here that although accessibility and usability do overlap in their design and implementation, they are not identical in their goals for their user groups.)
We discuss in some detail only the entries in the social level. Aesthetics is close to human senses and perception, and plays a crucial role in making a mobile application “salient” to its customers beyond simply the functionality it offers. It is critical that the mobile application be legal (e.g., is legal in the jurisdiction it operates and all components it makes use of are legal); takes steps to respect a user’s privacy (e.g., does not use or share user-supplied information outside the permitted realm); and be secure (e.g., in situations where financial transactions are made). The provider must take all steps to be transparent with respect to the user (e.g., not include misleading information such as the features of products or services offered, clearly label promotional content, make policies regarding returning/exchanging products open, and so on). The internal quality attributes for syntactic and semantic levels are inspired by Lindland, Sindre, and Sølvberg (1994) and Fenton and Pfleeger (1997). At the semantic level, we are only concerned with the conformance of the mobile application to the domain(s) it represents (that is, semantic correctness or completeness) and vice versa (that is, semantic validity). At the syntactic level the interest is in conformance with respect to the languages used to produce the mobile application (that is, syntactic correctness). The definitions of each of these attributes can vary in the literature, and therefore it is important that they be adopted and followed consistently. For example, the definition of usability varies significantly across ISO/IEC Standard 9126 and ISO Standard 9241 with respect to the perspective taken in their formulation.
Means for Credibility Assurance and Evaluation The third column of Table 1 lists (in no particular order, by no means complete, and not necessarily mutually exclusive) the means for assuring and evaluating active credibility.
Addressing the Credibility of Mobile Applications
•
•
•
•
“Expert” Body of Knowledge: The three types of knowledge that we are interested in are principles, guidelines, and patterns. Following the basic principles (Ghezzi, Jazayeri, & Mandrioli, 2003) underlying a mobile application enables a provider to improve quality attributes related to (T1-T3) of the framework. The guidelines encourage the use of conventions and good practice, and could also serve as a checklist with respect to which an application could be heuristically or otherwise evaluated. There are guidelines available for addressing accessibility (Chisholm, Vanderheiden, & Jacobs, 1999; Ahonen, 2003), security (McGraw & Felten, 1998), and usability (Bertini, Catarci, Kimani, & Dix, 2005) of mobile applications. However, guidelines tend to be more useful for those with an expert knowledge than for a novice to whom they may seem rather general to be of much practical use. Patterns are reusable entities of knowledge and experience aggregated by experts over years of “best practices” in solving recurring problems in a domain including that in mobile applications (Roth, 2001, 2002). They are relatively more structured compared to guidelines and provide better opportunities for sharing and reuse. There is, however, a lack of patterns that clearly address quality concerns in mobile applications. Also, there is a cost of adaptation of patterns to new contexts. Inspections: Inspections (Wiegers, 2002) are a rigorous form of auditing based upon peer review that can address quality concerns at both technical and social levels (T1-T5), and help improve the credibility of mobile applications. Inspections could, for example, decide what information is and is not considered “promotional,” help improve the labels used to provide cues to a user (say, in a navigation system), and assess the readability of documents. Still, inspections do involve an initial cost overhead from training each participant in the structured review process, and the logistics of checklists, forms, and reports. Testing: Some form of testing is usually an integral part of most development models of mobile applications (Nguyen, Johnson, & Hackett, 2003). There are test suites and test harnesses for many of the languages commonly used for representation of information in mobile applications. However, due to its very nature, testing addresses quality concerns of only some of the technical and social levels (T1, subset of T2, T3, T4, subset of T5). Therefore, testing complements but does not replace inspections. Accessibility or usability testing that requires hiring real users, infrastructure with video monitoring, and subsequent analysis of data can prove to be prohibitive for small-to-medium-size enterprises. Metrics: In a resource-constrained environment of mobile devices, efficient use of time and space is
•
critical. Metrics (Fenton & Pfleeger, 1997) provide a quantitative means for making qualitative judgments about quality concerns at technical levels. For example, metrics for a document or image size can help compare and make a choice between two designs, or metrics for structural complexity could help determine the number of steps required in navigation, which in turn could be used to estimate user effort. However, well-tested metrics for mobile applications are currently lacking. We also note that a dedicated use of metrics on a large scale usually requires tool support. Tools: Tools that have help improve quality concerns at technical and social levels. For example, tools can help engineers detect security breaches, report violations of accessibility or usability guidelines, find nonconformance to markup or scripting language syntax, suggest image sizes favorable to the small devices, or detect broken links. However, at times, tools cannot address some of the technical quality concerns (like complete semantic correctness of the application with respect to the application domain), as well as certain social quality concerns (like provider intent or user bias). Therefore, the use of tools as means for automatic quality assurance or evaluation should be kept in perspective.
Decision Support A mobile application project must take a variety of constraints into account: organizational constraints of time and resources (personnel, infrastructure, budget, and so on) and external forces (market value, competitors, and so on). These compel providers to make quality-related decisions that, apart from being sensitive to credibility, must also be feasible. For example, the provider of a mobile application should carry out intensive accessibility and usability evaluations, but ultimately that application must be delivered on a timely basis. Also, the impossibility of complete testing is well known. Indeed, the last column of Table 1 acknowledges that with respect to any assurance and/or evaluation, and includes feasibility as an all-encompassing consideration on the layers to make the framework practical. There are well-known techniques such as analytical hierarchy process (AHP) and quality function deployment (QFD) for carrying out feasibility analysis, and further discussion of this aspect is beyond the scope of this article.
Limitations of Addressing Credibility We note here that credibility, as is reflected by its primary dimensions, is a socio-cognitive concern that is not always amenable to a purely technological treatment. However, by decomposing it into quantifiable elements and approaching
Addressing the Credibility of Mobile Applications
them in a systematic and feasible manner, we can make improvements towards its establishment. We assert that the quality attributes we mention in pragmatic and social levels are necessary but make no claim of their sufficiency. Indeed, as we move from bottom to top, the framework gets less technically oriented and more human oriented. Therefore, finding sufficient conditions for establishing credibility is likely to be an open question, and it may be virtually impossible to provide complete guarantees for credibility.
FUTURE TRENDS In the previous section, we discussed active credibility; the issue of passive credibility poses special challenges and is a potential area of future research. We now briefly look at the case of reputed credibility. In case of Web applications, there have been two notable initiatives in the direction of addressing reputed credibility, namely WebTrust and TRUSTe. In response to the concerns related to for business-to-consumer electronic commerce and to increase consumer confidence, the American Institute of Certified Public Accountants (AICPA) and Canadian Institute of Chartered Accountants (CICA) have developed WebTrust Principles and Criteria and the related WebTrust seal of assurance. Independent and objective certified public accountant or chartered accountants, who are licensed by the AICPA or CICA, can provide assurance services to evaluate and test whether a particular Web application meets these principles and criteria. The TRUSTe program enables companies to develop privacy statements that reflect the information gathering and dissemination practices of their Web application. The program is equipped with the TRUSTe “trustmark” seal that takes users directly to a provider’s privacy statement. The trustmark is awarded only to those that adhere to TRUSTe’s established privacy principles and agree to comply with ongoing TRUSTe oversight and resolution process. Admittedly, not in the realm of pure academia, having similar quality assurance and evaluation programs for mobile applications, and perhaps even the use of ISO 9001:2000 as a basis for a certification, would be of interest. A natural extension of the preceding discussion on credibility could be in the context of the next generation of mobile applications such as semantic mobile applications (Alesso & Smith, 2002) and mobile Web services (Salmre, 2005). For example, ontological representation of information can present certain human-centric challenges (Kamthan & Pai, 2006) that need to be overcome for it to be a credible knowledge base. Finally, viewing a mobile application as an information system, it would of interest to draw connections between credibility and ethics (Johnson, 1997; Tavani, 2004).
CONCLUSION Although there have been significant advances towards enabling the technological infrastructure (Coyle, 2001) for mobile access in the past decade, there is much to be done in addressing the social challenges. Addressing credibility of mobile applications in a systematic manner is one step in that direction. The organizations that value credibility of their mobile applications need to take two aspects into consideration: (1) take a systematic approach to the development of the mobile applications, and (2) consider credibility as a first-class concern throughout the process. The former need to particularly include support for modeling a user’s environment (context, task, and device) (Gandon & Sadeh, 2004) and mobile user interface engineering. The latter implies that credibility is viewed as a mandatory non-functional requirement during the analysis phase and treated as a central design concern in the synthesis phase. In a user-centric approach to engineering, mobile applications belong to an ecosystem that includes both the people and the product. If the success of a mobile application is measured by use of its services, then establishing credibility with the users is critical for the providers. By making efforts towards improving the criteria that directly or indirectly affect credibility, the providers can meet user expectations and change the user perceptions in their favor.
REFERENCES Ahonen, M. (2003, September 19). Accessibility challenges with mobile lifelong learning tools and related collaboration. Proceedings of the Workshop on Ubiquitous and Mobile Computing for Educational Communities (UMOCEC 2003), Amsterdam, The Netherlands. Alesso, H. P., & Smith, C. F. (2002). The intelligent wireless Web. Boston: Addison-Wesley. Bertini, E., Catarci, T., Kimani, S., & Dix, A. (2005). A review of standard usability principles in the context of mobile computing. Studies in Communication Sciences, 1(5), 111-126. Chisholm, W., Vanderheiden, G., & Jacobs, I. (1999). Web content accessibility guidelines 1.0. W3C Recommendation, World Wide Web Consortium (W3C). Coyle, F. (2001). Wireless Web: A manager’s guide. Boston: Addison-Wesley. Fenton, N. E., & Pfleeger, S. L. (1997). Software metrics: A rigorous & practical approach. International Thomson Computer Press.
A
Addressing the Credibility of Mobile Applications
Fogg, B. J. (2003). Persuasive technology: Using computers to change what we think and do. San Francisco: Morgan Kaufmann. Fogg, B. J., Marshall, J., Laraki, O., Osipovich, A., Varma, C., Fang, N., et al. (2001, March 31-April 5). What makes Web sites credible?: A report on a large quantitative study. Proceedings of the ACM CHI 2001 Human Factors in Computing Systems Conference, Seattle, WA. Fogg, B. J., & Tseng, S. (1999, May 15-20). The elements of computer credibility. Proceedings of the ACM CHI 99 Conference on Human Factors in Computing Systems, Pittsburgh, PA. Gandon, F. L., & Sadeh, N. M. (2004, June 1-3). Contextawareness, privacy and mobile access: A Web semantic and multiagent approach. Proceedings of the 1st French-Speaking Conference on Mobility and Ubiquity Computing, Nice, France (pp. 123-130). Gibson, D. A. (2002). Communities and reputation on the Web. PhD Thesis, University of California, USA. Ghezzi, C., Jazayeri, M., & Mandrioli, D. (2003). Fundamentals of software engineering (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall. Hovland, C. I., Janis, I. L., & Kelley, J. J. (1953). Communication and persuasion. New Haven, CT: Yale University Press. Johnson, D. G. (1997). Ethics online. Communications of the ACM, 40(1), 60-65. Lindland, O. I., Sindre, G., & Sølvberg, A. (1994). Understanding quality in conceptual modeling. IEEE Software, 11(2), 42-49. Kamthan, P. (1999). E-commerce on the WWW: A matter of trust. Internet Related Technologies (IRT.ORG). Kamthan, P., & Pai, H.-I. (2006, May 21-24). Human-centric challenges in ontology engineering for the semantic Web: A perspective from patterns ontology. Proceedings of the 17th Annual Information Resources Management Association International Conference (IRMA 2006), Washington, DC. Kasanoff, B. (2002). Making it personal: How to profit from personalization without invading privacy. New York: John Wiley & Sons. McGraw, G., & Felten, E. W. (1998). Mobile code and security. IEEE Internet Computing, 2(6). Metzger, M. J. (2005, April 11-13). Understanding how Internet users make sense of credibility: A review of the state of our knowledge and recommendations for theory, policy,
0
and practice, Proceedings of the Internet Credibility and the User Symposium, Seattle, WA. Metzger, M. J., Flanagin, A. J., Eyal, K., Lemus, D., & McCann, R. (2003). Bringing the concept of credibility into the 21st century: Integrating perspectives on source, message, and media credibility in the contemporary media environment. Communication Yearbook, 27, 293-335. Nguyen, H. Q., Johnson, R., & Hackett, M. (2003). Testing applications on the Web: Test planning for mobile and Internet-based systems (2nd ed.). New York: John Wiley & Sons. Paavilainen, J. (2002). Mobile business strategies: Understanding the technologies and opportunities. Boston: Addison-Wesley. Roth, J. (2001, September 10). Patterns of mobile interaction. Proceedings of the 3rd International Workshop on Human Computer Interaction with Mobile Devices (Mobile HCI 2001), Lille, France. Roth, J. (2002). Patterns of mobile interaction. Personal and Ubiquitous Computing, 6(4), 282-289. Sadeh, N. M., Chan, T.-C., Van, L., Kwon, O., & Takizawa, K. (2003, June 9-12). A semantic Web environment for context-aware m-commerce. Proceedings of the 4th ACM Conference on Electronic Commerce, San Diego, CA (pp. 268-269). Salmre, I. (2005). Writing mobile code: Essential software engineering for building mobile applications. Boston: Addison-Wesley. Stamper, R. (1992, October 5-8). Signs, organizations, norms and information systems. Proceedings of the 3rd Australian Conference on Information Systems, Wollongong, Australia. Tavani, H. T. (2004). Ethics and technology: Ethical issues in an age of information and communication technology. New York: John Wiley & Sons. Wiegers, K. (2002). Peer reviews in software: A practical guide. Boston: Addison-Wesley.
KEY TERMS Delivery Context: A set of attributes that characterizes the capabilities of the access mechanism, the preferences of the user, and other aspects of the context into which a resource is to be delivered. Mobile Web Engineering: A discipline concerned with the establishment and use of sound scientific, engineering,
Addressing the Credibility of Mobile Applications
and management principles, and disciplined and systematic approaches to the successful development, deployment, and maintenance of high-quality mobile Web applications.
Semantic Web: An extension of the current Web that adds technological infrastructure for better knowledge representation, interpretation, and reasoning.
Personalization: A strategy that enables delivery that is customized to the user and user’s environment.
Semiotics: The field of study of signs and their representations.
Quality: The totality of features and characteristics of a product or a service that bear on its ability to satisfy stated or implied needs.
User Profile: A information container describing user needs, goals, and preferences.
A
Category: Human Factor
Adoption and Diffusion of M-Commerce Ranjan B. Kini Indiana University Northwest, USA Subir K. Bandyopadhyay Indiana University Northwest, USA
INTRODUCTION Mobile commerce (or in short, m-commerce) is currently at the stage where e-commerce was a decade ago. Many of the concerns consumers had regarding e-commerce (such as security, confidentiality, and reliability) are now directed towards m-commerce. To complicate the matter further, the lack of a standardized technology has made m-commerce grow in multiple directions in different parts of the world. Thus, the popularity of m-commerce-based services varies by country, by culture, and by individual user. For example, in Europe the most popular application is SMS (short message service) or text messaging, in Japan interactive games and picture exchange via NTT DoCoMo i-mode, and in North America e-mail via interactive pagers (such as RIM BlackBerry) and wireless application protocol-based (WAP-based) wireless data portals providing news, stock quotes, and weather information. It is safe to predict that these applications will take on different forms as the technologies mature, devices become more capable in form and functionality, and service providers become more innovative in their business models. It is true that m-commerce has witnessed spectacular growth across the globe. It is also encouraging that several factors are expected to accelerate the pace of adoption of m-commerce. Notable among these drivers is convergence in the voice/data industry, leaping improvements in related technology and standards, adoptive technology culture in many parts of the world, and governmental and regulatory initiatives. Despite the undisputed promise of m-commerce, there are several barriers that are slowing the pace of adoption of m-commerce. The major barriers include: (a) lack of good business models to generate revenues, (b) perception of lack of security, (c) short product lifecycle due to rapidly changing technology, (d) non-convergence of standards, (e) usability of devices, (f) limitation of bandwidth, and (g) cost. Many of the aforesaid were common to e-commerce also at its introduction and growth stage. We strongly believe it is worthwhile to investigate how e-commerce has been able to overcome these barriers so that we can incorporate some of the successful strategies to m-commerce. In our study, we will first compare and contrast e-commerce and
m-commerce with respect to a set of common criteria such as: (1) hardware requirement, (2) software requirement, (3) connection or access, and (4) content. In the process, we will identify the principal barriers to the development of m-commerce as outlined in the above list.
The Growth in E-Commerce Electronic commerce or e-commerce is the mode of commerce wherein the communication and transactions related to marketing, distributing, billing, communicating, and payment related to exchange of goods or services is conducted through the Internet, communication networks, and computers. Since the Department of Defense opened up the Internet for the public to access in 1991, there has been exponential growth in the number of Web sites, users on the Web, commerce through the Web, and now change of lifestyle through the Web (Pew, 2006). The chronology of events shows that as the Internet became easier and cheaper to use, and as the applications (such as e-mail and Web interaction) became necessary or useful to have, the rate of adoption of the Internet accelerated. In fact, the rate of adoption of the Internet surpassed all projections that were made based on the traditional technology adoption rates that were documented for electricity, automobile, radio, telephone, and television (Pew, 2006). Unfortunately, the over-enthusiastic media hyped up the growth rate to an unsustainable level, leading to unprecedented growth of investment in the Internet technologies and followed by a melt-down in the stock market. This shattered the confidence in Internet technologies in the investment market. Although there was a significant deceleration in IT investment, e-commerce has rebounded to a large extent since the dot.com bust. It has been growing at about 30% compound rate per year (Pew, 2006). In the last 10 years, the adoption of e-commerce has been extensively studied both by academicians as well as practitioners. During this period e-commerce and the scope of its definition also went through various iterations. For example, people may not buy a car on the Internet, but it is documented that 65% of car buyers have done extensive research on the Web about the car they eventually buy. Is this e-commerce? Should we restrict the e-commerce definition
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Adoption and Diffusion of M-Commerce
to financial exchange for goods or services? We have various such examples in the marketplace where extensive research about the product or service is conducted on the Internet, but the final purchase is made in the physical environment. Hence, although the number of consumer financial transactions has not grown to the level industry projected initially, there has been a significantly high rate of adoption of the activities supporting e-commerce. In addition, there has been a very high rate of adoption of business-to-business (B2B) commerce both in terms of financial and supporting transactions. In this article, we are interested in business-to-consumer (B2C) commerce. Hence, the comparison and contrast is made between ecommerce and m-commerce. All our discussion henceforth will be on B2C commerce using desktop and/or mobile technologies.
The Growth Potential of M-Commerce Mobile commerce is the model of commerce that performs transactions using a wireless device and data connection that result in the transfer of value in exchange for information, services, or goods. Mobile commerce is facilitated generally by mobile phones and newly developed handheld devices. It includes services such as banking, payment, ticketing, and other related services (DEVX, 2006; Kini & Thanarithiporn, 2005). Currently, most m-commerce activity is performed using mobile phones or handsets. This type of commerce is common in Asian countries led by Japan and South Korea. Industry observers are expecting that the United States will catch up soon, with mobile phones replacing existing devices such as ExxonMobil’s Speedpass (eMarketer, 2005; Kini & Thanarithiporn, 2005). Although the U.S. is lagging behind many countries in Asia and Europe in m-commerce, a UK-based research firm projects North American m-commerce users to total 12 million by 2009, with two-thirds of them using the devices to buy external items such as tickets and goods, and a third of them using it to make smaller transactions through vending machines (eMarketer, 2005). The firm also notes that there is a large potential number of the 95 million current American teens who are already making purchases on the Web that will adopt m-commerce. However, the study also remarks that generating widespread user interest in m-commerce and addressing security fears of mobile payment technologies and m-commerce services are critical in achieving a high level of adoption (eMarketer, 2005). While the Asia Pacific Research Group (APRG, 2006) projected in 2002 that global m-commerce would reach US$10 billion 2005, Juniper Research currently projects that the global mobile commerce market, comprising mobile entertainment downloads, ticket purchases, and point-of-sale (POS) transactions, will grow to $88 billion by 2009, largely
on the strength of micro-payments (e.g., vending machine type purchases). See eMarketer (2005) for more details. Today, a large percentage of mobile phone users use mobile phones to download ring tones and play games; hence content-based m-commerce is expected to make up a small percentage of m-commerce. One recent study, however, projects that in the future mobile phone users will move up the value chain from purchases that are used and enjoyed on the mobile phone to external items such as tickets, snacks, public transportation, newspapers, and magazines (eMarketer, 2005).
Diffusion Models of Technology Adoption There are many models that have been formulated and studied with regard to technology adoption, acceptance, diffusion, and continued adoption. These theories identify factors that are necessary to support different levels of adoption of information and communication technologies (ICTs). Notable among these models are the innovation-diffusion theory (Roger, 1995), technology acceptance model (or TAM) based on the theory of reasoned action (Davis, 1989; Fishbein & Ajzen, 1975), extended TAM2 model that incorporates social factors (Venkatesh & Davis, 2000), technology adoption model based on the theory of planned behavior (Ajzen & Fishbein, 1980), post acceptance model based on marketing and advertising concepts (Bhattacherjee, 2001), and SERVQUAL (Parasuraman, Berry, & Zeithaml, 1988) for service quality. These models have been extensively used to predict and evaluate online retail shopping and continued acceptance of ICTs. In addition, varieties of integrated models have been developed to measure the success of information systems, ICT, and Internet adoption and diffusion. Currently, many of these models are being tested in the context of mobile technology (primarily mobile phone services). The integration models mentioned above have been empirically tested in the e-commerce area. The models have been authenticated and proven to be extremely useful in predicting behavior of users of ICT and e-commerce. In the case of m-commerce, the results have been slightly inconsistent. Primarily these inconsistencies have been found because of the differing market maturity levels or the usage pattern of mobile devices. For example, in a South Korean study where mobile phones have been in use for quite some time, the results of testing an integrative m-commerce adoption model yielded different results for actual use than in a similar study conducted in Thailand where mobiles devices were introduced much later in the market. South Koreans were not influenced much by advertising, unlike Thai people in the initial adoption phase of m-commerce. Conversely, Thai people were not influenced by word-of-mouth to the extent South Koreans were influenced in the initial adoption (Thanarithiporn, 2005). According to Thanarithiporn
A
Adoption and Diffusion of M-Commerce
(2005), this is due to the fact South Koreans are at a more advanced level of adoption for ICTs. Furthermore, Thanarithiporn (2005) found that, unlike in South Korea where content availability had no influence in the continued use of mobile phones, it had a strong influence in Thailand on mobile usage rate. Also, in both countries self-efficacy had no influence one way or the other in the initial adoption of the mobile phone.
Key Factors that Affect the Adoption and Diffusion of E-Commerce and M-Commerce As expected, many factors influence the rate of adoption and diffusion of technological innovations. We reviewed the extant literature, as outlined above, to identify those factors. In particular, we were interested in a set of factors that have significant influence in the adoption and diffusion of both e-commerce and m-commerce. These include: (a) hardware requirement, (b) software requirement, (c) connection or accessibility, and (d) content. In the following paragraphs, we will outline how these factors have influenced the development of e-commerce, and are currently influencing the adoption and diffusion of m-commerce.
Hardware Requirement E-Commerce Computer users were used to the QWERTY keyboard (of typewriters), thus they easily adapted to the standardized desktop of the first personal computers (PCs) in the 1980s. The development of graphical user interface (GUI), mice, and various other multimedia-related accessories has made PCs and variations thereof easy to use. With the introduction of open architecture, the adoption and diffusion of PCs proliferated. The introduction of the Internet to the common public, and the introduction of the GUI browser immediately thereafter, allowed PC users to quickly adopt the Web browsers and demand applications in a hurry. The limitation of hardware at the user level was only restricted by the inherent rendering capability of a model based on the processors, configuration, and accessories that supported them. Since the Web and e-commerce server technologies that serve Internet documents or Web pages are also based on open architecture, limitations were similar to that of desktops. M-Commerce The hardware used for mobile devices are complex. The evolution of the hardware technology used in mobile devices is diverse because of the diversity in fundamental architecture. These architectures are based on diverse technology standards such as TDMA, CDMA, GPRS, GSM, CDMA/2000, WCDMA, and i-mode. In addition, these architectures have
gone through multiple generations of technology such as 1G (first generation – analog technology); 2G (second generation – digital technology, including 2.5G and 2.75G); and 3G, to meet the demands of customers in terms of bandwidth speed, network capabilities, application base, and corresponding price structures. The lack of uniform global standards and varied sizes and user interfaces to operate the devices has further disrupted the smoother adoption process. While the U.S. still suffers from a lack of uniform standard, Europe is moving towards uniformity through some variation of TDMA technology, and China is modifying CDMA technology to develop its own standard. Other countries are currently working towards a uniform standard based on a variation of base TDMA or CDMA technology (Keen & Mackintosh, 2001). The innovation in the changing standards, devices, applications, and cultural temperament have constantly maintained a turbulent environment in the adoption and diffusion of commerce through mobile devices. For example, if the device is WAP-enabled, then Web services can be delivered using standardized WML, CHTML, or J2ME development tools. But the WAP enabling has not given scale advantages because hardware standards have not converged, at least not in the U.S. where consumers use a multitude of devices such as Palm, different Web-enabled phones, and different pocket phones.
Software Requirement E-Commerce The standardization and open architecture of PCs, along with the high degree of penetration of PCs in the office and home environment, allowed for standardization of client devices. This allowed for the development of text browsers, and subsequently the development of the graphical interface through Web browsers. Apples, PCs, and other UNIX-based workstations were able to use the device-independent Web browsers, thus leading to rapid adoption and expansion in the usage of Web browsers. The low price of earlier browsers such as Mosaic and Netscape, and the distribution of Internet Explorer with the Windows Operating System by Microsoft allowed the diffusion of the browsing capability in almost every client in the market. Standardized browser software and interface, along with market dominant operating systems such as the Windows family of desktop operating systems and server platforms, facilitated the exponential growth of Internet users and applications. The availability, integration, and interoperability of application development tools, and the reliance on open systems concept and architecture, fueled further changes in the interactivity of the Web and indirectly boosted the commerce on the Web. The development of hardware-independent Java (by Sun Microsystems) and similarly featured tools allowed growth in the interactivity of the Web and application inte-
Adoption and Diffusion of M-Commerce
gration both at the front end and backend of the Web. The interoperability of Web applications to communicate with a wide variety of organizational systems initiated a concern for security of the data while in transit and storage. In the early stages of e-commerce, major credit card companies did not trust the methodologies that were used, although they allowed the transactions. Beginning in 1999, they started protecting the online customers just as they protected off-line customers (namely, a customer is only responsible for $50 if she reports the card stolen within 24 hours). The technology companies and financial service organizations collaboratively created and standardized methodologies for online secure transactions, and originated the concept of third-party certification of authority. This certification practice further strengthened the security of online commerce and established a strong basis for consumers to trust and online commerce to grow. M-Commerce Software for mobile technologies is dependent on the technology standard used and type of applications suitable for the mobile device. In most nations, like in the U.S., the use of mobile devices started with the use of analog cellular phones. These required proprietary software and proprietary networks. The digitization of handheld devices started with personal digital assistants (PDAs) for personal information management. The transformation of the PDA as a digital communication tool was made possible by private networks, operating systems, and applications developed by companies such as Palm. However, as Microsoft’s Windows CE (Compact Edition) and BlackBerry started offering email, information management tools, and Web surfing using micro-browsers, the growth in the use of handheld devices for Web applications started growing. The handheld industry responded with a variety of applications and made WAP a standard for applications development. Concurrently, the telecom industry brought out digital phones and devices that could offer voice, personal information management (PIM), and data applications. However, until now, operating systems, servers, and Web applications are not standardized in the handheld market. The diversity of server software and client operating systems, and the availability of applications have not made these devices interoperable. In addition, with each player offering its own network and original content or converted content (i.e., content originally developed for the desktop computers), the interest in commerce using mobile devices has not been too enthusiastic. Furthermore, the lack of common security standards has made mobile commerce adoption very slow.
Connection or Access E-Commerce In the United States, where telephone wire lines have been in existence for over 100 years, it was natural for the telecom companies to focus on offering Internet connectivity through the existing telephone network. In the early stages of pubic offering of the Internet, it was easy for people to adopt the Internet using their modem from a private network. As the Internet evolved into the World Wide Web, and innovation brought faster modems to the market, more Internet service providers (ISPs) started providing ramps to the Internet. When the Windows98 Operating System with its integrated Internet Explorer was introduced to the marketplace, the Internet adoption was growing in triple digits per year. The major infrastructural components were already in place. The telecom sector invested heavily into building the bandwidth and router network to meet the insatiable demand for Web surfing. Worldwide Internet adoption and use was growing exponentially. The ICT industry responded with innovative technologies, software and services using standardized PCs, modems, support for (Internet protocol suite) TCP/IP protocol of Internet, and highly competitive pricing. The etail industry subsequently started growing rapidly, and the financial service industry introduced innovative products and services while collaboratively designing secure electronic payment mechanisms with ICT industry players. The drop in pricing, availability of bandwidth, security, and quality of products and services bolstered the commerce activity on the Internet until the ‘dot.com bust’ of May 2000. Although the bust slowed the growth rate of e-commerce, in reality e-commerce continuously grew despite the bust. Support for e-commerce from the U.S. government to fuel the e-commerce growth through moratorium on taxes by two administrations considerably helped the diffusion of e-commerce. The concern about the security in e-commerce shown by laggards was eased by a variety of security and encryption tools, and the creation of the certification of authority concept by strong security services offered by companies such as Verisign, TRUSTe, and others. Lately, the demand for highly competitive broadband service availability, and the availability and delivery of media-rich content, has brought media and entertainment industry to the Web with greater force. These technological advances in the e-commerce sector have received increased attention, thus ensuring a strong global growth rate in ecommerce. M-Commerce In the mobile arena, customers may have been using analog cellular phones (1G) for a long of time. During the era of analog cellular phones, the common mobile commerce activity was the downloading of ring tones. This type of commerce activity is still quite prevalent in developing na
A
Adoption and Diffusion of M-Commerce
tions. In addition to this type of commerce, other types of commerce conducted using these devices are the same as the ones that can be performed using a standard desk phone, such as ordering tickets for an event, ordering catalog items, and similar tasks. With the introduction of digital devices (2G), mobile phones quite suddenly have become the lifeline for many transactions, such as e-mail, voicemail, and text messaging. With 2.5G, 2.75G, and now with 3G devices, more varied and complex applications such as photo transfers, interactive games, and videos have become the norm. The capabilities of these devices are determined by technical ability of the devices and the support of terrestrial tower structures by the vendors offering these services. In addition, the content availability and their desirability by the customers also determine the adoption of such services. The technology, standards, and competition have left U.S. vendors in the distance in rolling out new technology and services. While Asia’s (South Korea, Japan, and China) mobile penetration growth is three times that of the United States, Europe is closely behind Asia, with England (87%) and Finland (75%) achieving very high penetration rates (Shim, 2005). In the U.S., the major players in the telecom industry are collaborating to achieve the 3G-standard Universal Mobile Telecommunication System (UMTS) to provide penetration and support rollout of new technology and services. Several countries including South Korea were planning to offer a more advanced technology called the Digital multimedia broadband (DMB) or wire broadband (WiBro) by the end of 2006 (Shim, 2005). According to Shim (2005), it will take a while to obtain DMB cellular phone services in the U.S., since technical standards and logistical barriers will have to be overcome first. The private networks built by the wireless service providers through the customized devices will determine the access and speed available in the future in the United States. The investment in the network, along with the rollout of new technology and methods used to price the services, will be strong factors in building the capacity. Government policies are also vital in this respect. According to Shim (2005), the government commitment and push for IT strategy and long-term goals are among the most important factors to advance a country’s cellular mobile business, particularly for less-developed countries.
Content E-Commerce Identifying the most preferred method for delivery of any content has always been a thorny issue. In electronic commerce, the complete digital conversion of all media into technology mandated by the FCC by 2008 would be much easier (FCC, 2006). Voice, as well as radio and television signals, will be broadcast digitally. The Internet has built
capacity to deliver rich media content at high speed using the fiber network in the U.S. The convergence of devices such as TV monitors and PC monitors has already brought down the prices for such devices due to scale effects. The stumbling blocks to achieve a greater level of broadband adoption (from the current 53% in the U.S.) are pricing and quality of content (Pew, 2006). In e-commerce, content can be provided by anyone using standardized development tools and can be served on the standardized server software since most desktops can handle all the content delivered through the Web. The diffusion of such innovations is constrained by the pricing and the investment made by consumers at the client level. The industry has converged in standardizing hardware, software, and protocols. Globally as well as in the U.S., there is a clear trend to make the technology affordable throughout the world through the open systems concept. This has helped tremendously, especially in developing countries, in the adoption and diffusion of the Internet and generalized applications. M-Commerce In mobile commerce, the content such as data, text, audio, video, and video streaming can be delivered through the devices provided by service providers through their network infrastructure. As the service providers rollout new network technologies with greater capabilities to adapt to the new generation of hardware and software technologies, consumers can expect more media-rich content. Any content that is available in the e-commerce world will be specially modified for mobile delivery using specific development tools for WAP-enabled devices such as WML, CHTML, and J2ME. Depending on the type of device, the content will have to be delivered in device-specific configurationfor example, the content has to be delivered differently to a PocketPC, WAP-enabled mobile phone, and WAP-enabled PDAs. This type of dynamic configuration in the content delivery requires investment from service providers and/or value-added intermediaries. The special intermediaries provide enormous value-added services in converting the e-commerce content for different mobile devices and become consolidators of content and applications and essentially become data portals for mobile devices. The diversity of devices available in the market will require a significant amount of investments in the U.S. to offer it nationwide, unless it focuses only on high-population density regions to maximize returns.
CONCLUSION Based on the foregoing discussion, we can say that the introduction of e-commerce has been comparatively smoother than m-commerce. The development of the hardware capability (from PC to GUI to other multimedia-related accessories
Adoption and Diffusion of M-Commerce
such as printers, camera, etc.), the software capability (such as browsers, open operating systems, payment schemes, secure systems, etc.), better accessibility (such as phone lines, cables, etc.), and more varied content (such as voice, radio, and television signals) ensured a fast adoption and diffusion of e-commerce throughout the world. It is true that m-commerce also enjoys many advantages similar to e-commerce. For example, the mobile phonethe principal mode of m-commerceis witnessing a spectacular growth throughout the world. Unfortunately, unlike e-commerce, m-commerce does not enjoy an open architecture that can accommodate varied standards in hardware, software, connection technology, and the content. Several countries (such as Japan and South Korea) are further ahead of the U.S. in solving this issue of incompatible technologies. It is heartening to see a sincere effort in many countries, including the U.S., to achieve convergence in technologies so that m-commerce is able to grow true to its full potential.
REFERENCES Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall. APRG. (2006). Retrieved from http://www.aprg.com Bhattacherjee, A. (2001). Understanding information systems continuance: An expectation-confirmation model. MIS Quarterly, 25(3), 351-370. Cassidy, J. (2002). dot.con: The greatest story every sold. New York: Harper Collins. Davis, F. D. (1989). Perceived usefulness, perceived ease of use and user acceptance of information technology. MIS Quarterly, 13(2), 319-339. DEVX. (2006). Retrieved from http://www.devx.com/wireless/Door/11297
FCC. (2006). Retrieved from http://www.fcc.gov/cgb/consumerfacts/digitaltv.html Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Keen, P., & Mackintosh, R. (2001). The freedom economy: Gaining the m-commerce edge in the era of the wireless Internet. Berkeley, CA: Osborne/McGraw-Hill. Kini, R. B., & Thanarithiporn, S. (2004). M-commerce and e-commerce in ThailandA value space analysis. International Journal of Mobile Communications, 2(1), 22-37. Parasuraman, A., Berry, L. L., & Zeithaml, V. A. (1988). SERVQUAL: A multiple-item scale for measuring customer perceptions of service quality. Journal of Retailing, 64(1), 12-40. Pew. (2006). Retrieved from http://www.pewinternet.org Rogers, E. M. (1995). Diffusion of innovations. New York: The Free Press. Schifter, D. E., & Ajzen, I. (1985). Intention, perceived control, and weight loss: An application of the theory of planned behavior. Journal of Personality and Social Psychology, 49(3), 843-851. Shim, J. P. (2005). Korea’s lead in mobile cellular and DMB phone services. Communications of the Association for Information Systems, 15, 555-566. Thanarithiporn, S. (2004). A modified technology acceptance model for analyzing the determinants affecting initial and post intention to adopt mobile technology in Thailand. Unpublished dissertation, Bangkok University, Thailand. Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186-204.
eMarketer. (2005). Mobile marketing and m-commerce: Global spending and trends. eMarketer, (February 1).
A
38 Category: Human Factor
Adoption of M-Commerce Devices by Consumers Humphry Hung Hong Kong Polytechnic University, Hong Kong Vincent Cho Hong Kong Polytechnic University, Hong Kong
INTRODUCTION The Internet has undoubtedly introduced a significant wave of changes. The increased electronic transmission capacity and technology further paves a superhighway towards unrestricted communication networks (Chircu & Kauffman, 2000; Cowles, Kiecker, & Little, 2002). It is estimated that by 2007, the total number of Internet users in the world will be over 1.4 billion and the percentage of wireless users is projected to take up about 57% of the vast number (Magura, 2003). Most people anticipate that the next-generation commerce will emerge from traditional commerce to PC-based e-commerce, and eventually to mobile commerce (Ellis-Chadwick, McHardy, & Wiesnhofer, 2000, Miller, 2002, Watson, Pitt, Berthon, & Zinkhan, 2002). Mobile commerce (m-commerce) is an extension, rather than a complete replacement, of PC-based electronic commerce. It allows users to interact with other users or businesses in a wireless mode, anytime and anywhere (Balasubramanian, Peterson, & Jarvenpaa, 2002; Samuelsson & Dholakia, 2003). It is very likely that PC-based e-commerce will still prevail for a relatively long period of time in spite of the trend that more and more people will choose to adopt m-commerce for their purchases (Miller, 2002). The focus of our article is on the consumers’ adoption of m-commerce devices (MCDs), which are equipment and technologies that facilitate users to make use of m-commerce. MCDs include mobile phones, personal digital assistants (PDA), portable computer notebooks, Bluetooth, WAP, and other facilities that can have access to the wireless networks. We expect that the heading towards a world of mobile networks and wireless devices, which will present a new perspective of time and space, is definitely on its way. Several basic questions about m-commerce devices will be addressed in this article. First, why should consumers adopt MCDs? What will be the influencing factors for consideration? Are these MCDs easy to use and proven to be useful? Second, how do the MCDs compare with the devices for other types of commerce, such as e-commerce or traditional mail order? Consumers will only adopt MCDs when there are some potential significant advantages when comparing to old devices for other types of commerce. There is still a
lack of comprehensive framework within which the adoption of MCDs can be evaluated. Traditional viewpoints regarding this issue, especially those that are based on technology acceptance models, will need to be revisited and revised when consumers are considering such an adoption. In this article, we propose a framework for identifying the various influencing factors of the adoption of MCD, as well as the antecedents of these influencing factors. Because of the need of the standardization of the application, interface, and inter-connectivity of all hardware and software relevant to the adoption and usage of MCDs, our proposed framework will have some global implications (Zwass, 1996). Our conceptual framework can, therefore, make significant contributions to a more in-depth understanding in the spread and acceptability of m-commerce through knowing why and how relevant MCDs are adopted. While using technology acceptance models (TAMs) as our primary reference, we also incorporate the important implications of an options model into our basic framework of analyzing consumers’ adoption of MCDs. Based on our theoretical framework, we identify four influencing factorsmerits, maturity, maneuverability, and mentalitywhich we consider to be relevant to the decision of consumers in adopting MCDs. We also identify two generic antecedents of these influencing factorsmobility and matching. We plan to investigate the extent of influence of these influencing factors and their antecedents, which will affect consumers’ adoption decisions of MCDs. Figure 1 is a graphical representation of our conceptual model of the adoption of MCDs by consumers.
INFLUENCING FACTORS BASED ON TECHNOLOGY ACCEPTANCE MODEL The technology acceptance model is an information systems theory that models how users come to accept and use a new technology, with reference to two major considerations, perceived usefulness and perceived ease of use (Venkatesh & Davis, 2000). The former is about the degree to which a person believes that using a particular system will make
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Adoption of M-Commerce Devices by Consumers
Figure 1. A conceptual model of the adoption of m-commerce devices
A
Features of M-commerce Devices
Standardization and common interface
Antecedents
Technology Acceptance Model (Focusing on the advantages of new technologies)
Independent variables
Dependent Variable
Free to move
Matching
Mobility
Options Model
Recognition by Peer groups
Mentality
Easy to Use Usefulness
Maneuverability
Significant improvement
Merits
Maturity of Technology
(Focusing on the comparison between old and new technologies)
Maturity
Adoption of M-commerce Devices By Consumers
his or her life easier, for instance, by enhancing his or her job performance or reducing the workload, while the latter is the degree to which a person believes that it is not difficult to actually use a particular system (Venkatesh & Davis, 2000). With reference to TAM, we consider whether the adoption of MCDs will bring advantages to consumers. We identify two Ms, maneuverability and mentality, for relating the acceptability of MCDs to users. The first influencing factor, maneuverability, is related to the perceived usefulness in the adoption of MCDs and the degree to which a person can make the best use of such MCDs. Consumers will tend to adopt devices that are user friendly and do not require some intensive training of adoption (Prasanna et al., 1994). The second influencing factor, mentality, is concerned with the match between the new technology and consumers’ own mindsets, as well as the appropriate recognition of their peer groups (Bessen, 1999; Venkatesh & Davis, 2000). General acceptance by the consumers, especially by their peer groups, will be very important to consumers
when they consider using MCDs for matching the devices of other people.
Influencing Factors Based on Options Model While mainstream literature on the adoption of new technologies is primarily based on the technology acceptance model, we consider that, in the context of m-commerce, we also need to think about some other aspects. The options model demonstrates that a new technology with a moderate expected improvement in performance can experience substantial delays in adoption and price distortions even in a competitive market (Bessen, 1999; Sheasley, 2000). Rather than adopting a new technology that demonstrates only marginal improvement, consumers have the option of not adopting until the new technology, in terms of performance, is substantially better than the old technology. Consumers contemplating the adoption of a new technology are, of course, aware of the possibility of 39
Adoption of M-Commerce Devices by Consumers
Figure 2. A diagrammatic representation of the role of the four influencing factors of MCDs
Future MCD
Options Model: Comparing future and existing MCD
Technology Acceptance Model: Focus on the existing MCD
Maturity Maneuverability
Existing MCD
Matching
Merits
Old device for traditional commerce
sequential improvement. They consider not only the current technical level of the new technology, but also their expectations of possible upgrades and changes in the future of the new technology (Sheasley, 2000). With regards to the options model, we consider the comparison between MCDs and devices for other types of commerce, and in particular, the comparative advantages of MCDs to consumers. Based on the options model, we identify two Ms, merits and maturity, in relation to the comparison. We identify the third influencing factor, merits, which is about the degree to which a buyer believes that the MCD can provide significant improvement in the purchase process. Handheld mobile devices, such as PDAs and other enhanced alphanumeric communicators, have supplemented mobile telephones, thus expanding the range of MCDs available for m-commerce transactions. With the abilities to be connected to digital communication networks, MCDs are considered to be in possession of important comparative advantage of mobility. The fourth influencing factor, maturity, is the possibility that the technology of the MCD is mature enough so that there will not be any possible significant improvements at a later stage. While academic researchers and business practitioners recognize that the electronic market will penetrate and replace a traditional type of commerce, there are still some reservations that will likely cause the early adopters of new technologies some problems in terms of the obso40
lescence of devices (Samuelsson & Dholakia, 2003). Most consumers will prefer adopting MCDs with more mature technologies so that there is no need for a high level of subsequent upgrading of devices. In essence, the option model focuses on the comparison between existing and old MCDs, while TAM places emphasis on the generic attributes and utility of MCDs. Figure 2 shows the inter-relationship among the four influencing factors. Based on the four factors that we have identified, we propose the followings: Proposition 1: Maneuverability, mentality, merits, and maturity are the influencing factors when consumers consider adopting MCDs for purchases.
GENERIC ATTRIBUTES OF MCD In addition to the identification of the influencing factors of the adoption of MCDs, we also consider their antecedents, which are related to the very basic and essential characteristics of MCDs. We start our analysis by considering two generic attributes of MCDs, mobility and matching. Mobility is the most fundamental aspect of m-commerce because the name m-commerce arises from the mobile nature of the wireless environment that supports mobile electronic transactions (Coursaris, Hassanein, & Head, 2003). Mobile wireless de-
Adoption of M-Commerce Devices by Consumers
vices, such as mobile phones, PDAs, and portable computer notebooks, can have the ability to help users gain access to the Internet. Based on these wireless devices, m-commerce is a natural extension of e-commerce but can provide some additional advantages of mobility for consumers. Mobility is a major prerequisite for the adoption of MCDs. It is an antecedent of the influencing factors of the adoption of MCDs because people will consider adopting a wireless connection because it can allow significant improvement compared with traditional device (i.e., merits), and is perceived to be useful and convenient (i.e., maneuverability). Matching describes the need for the standardized and common interface of MCDs (Coursaris et al., 2003). The unique characteristic of m-commerce very often requires both ends of this new type of commerce to have a common interface. M-commerce applications have the challenging task of discovering services in a dynamically changing environment. Effective mechanisms need to be in place for the interface between various types of MCDs. Matching is an important antecedent of the influencing factors of consumers’ adoption of MCDs because the need for standardization (i.e., matching) is important for m-commerce technology, which allows for the connection of MCDs with the wireless networks and the connections among different MCDs. This standardized interface (i.e., matching) also reflects that the MCD is mature (i.e., maturity). Moreover, the standardized interface (matching) will also help to promote the universal acceptance of MCDs by people (i.e., mentality). Based on these arguments, we develop the second proposition: Proposition 2: The generic attributes of m-commerce, mobility and matching, are the antecedents of the influencing factors when consumers adopt MCDs for purchases.
RESEARCH IMPLICATIONS Based on our conceptual framework, we identify the various influencing factors (i.e., 4 Ms) which can affect consumers’ decisions about the adoption of MCDs in their purchases. It is possible to collect data on whether consumers will consider the adoption of MCDs, and at the same time, researchers can also investigate the reasons why they adopt or do not adopt MCDs, in terms of timing, opportunities, changing trends, and applications. In our conceptual framework, the dependent variable is the intention of consumers to adopt MCDs. We identify four Ms as the primary influencing factors of the adoption of new technologies in m-commerce (maneuverability, mentality, merits, and maturity). These are independent variables in our framework. We also identify the antecedents of these influencing factors, mobility and matching. First, maneuverability will be measured by the usability of the MCD. Mentality can be evaluated by the perceived
peer groups’ acceptance of MCDs. Merits can be measured by the comparative advantages of the MCD in relation to the old devices for other types of commerce. Maturity can be assessed by the perception that the relevant MCD can or cannot be upgraded. The first antecedent, mobility, can be measured by the extent of access to wireless networks. Matching can be measured by the degree that MCDs can be compatible with each other. In addition to the primary independent variables, we suggest measuring some important control or moderating variables, such as price, and completing a demographic profile such as sex, age, and education levels, as well as occupations and incomes of consumers.
CONCLUSION AND IMPLICATIONS In our proposed model, we are exploring new insights and new adoption behavior in the ubiquitous world of m-commerce, which we believe is not yet fully understood by most marketers and scholars (Stevens & McElhill, 2000; Struss, El-Ansary, & Frost, 2003). Our proposed model will be of interest to academics in the IT field, who may be keen to know how they can perform further relevant research in m-commerce. Our proposed framework represents a theory-driven examination of the adoption of MCDs by consumers in their purchase processes. The powerful tool of m-commerce can allow for faster and easier response to market demand, and at the same time consumers can obtain relevant information as well as purchasing goods and services at any time and anywhere as they prefer. It is expected that our proposed framework can provide important guidelines for pointing the way towards some relevant research on the significance of the adoption of MCDs. Our conceptual framework contributes to literature by ascertaining the most significant independent variables from among all those key variables that we have identified based on our literature review, which can determine which and how new technologies are likely to be adopted in mcommerce.
REFERENCES Balasubramanian, S., Peterson, R.A., & Jarvenpaa, S.L. (2002). Exploring the implications of m-commerce for markets and marketing. Academy of Marketing Science Journal, 30(4), 348-361. Bessen, J. (1999). Real options and the adoption of new technologies. Retrieved from http://www.researchoninnovation.org/online.htm#realopt
41
A
Adoption of M-Commerce Devices by Consumers
Bisbal, J., Lawless, D., Wu, B., & Grimson, J. (1999). Legacy information system migration: A brief review of problems, solutions and research issues. IEEE Software, 16, 103-111. Chircu, A., & Kauffman, R. (2000). Reintermediation strategies in business-to-business electronic commerce. International Journal of Electronic Commerce, 4(4), 7-42. Coursaris, C., Hassanein, K., & Head, M. (2003). M-commerce in Canada: An interaction framework for wireless privacy. Canadian Journal of Administrative Sciences, 20(1), 54-73. Cowles, D.L., Kiecker, P., & Little, M.W. (2002). Using key informant insights as a foundation for e-retailing theory development. Journal of Business Research, 55, 629-636. Ellis-Chadwick, F., McHardy, P., & Wiesnhofer, H. (2000). Online customer relationships in the European financial services sector: A cross-country investigation. Journal of Financial Services Marketing, 6(4), 333-345. Magura, B. (2003). What hooks m-commerce customers? MIT Sloan Management Review, 44(3), 9-10. Miller, A.I. (2002). Einstein, Picasso: Space, time, and the beauty that causes havoc. New York: Basic Books. Samuelsson, M., & Dholakia, N. (2003). Assessing the market potential of network-enabled 3G m-business services. In S. Nansi (Ed.), Wireless communications and mobile commerce. Hershey, PA: Idea Group Publishing. Sheasley, W.D. (2000). Taking an options approach to new technology development. Research Technology Management, 43(6), 37-43. Stevens, G.R., & McElhill, F. (2000). A qualitative study and model of the use of e-mail in organizations. Electronic Networking Applications and Policy, 10(4), 271-283. Struss, J., El-Ansary, A., & Frost, R. (2003). E-marketing (3rd ed.). Englewood Cliffs, NJ: Prentice Hall. Venkatesh, V., & Davis, F.D. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186-204.
42
Watson, R.T., Pitt, L.F., Berthon, P., & Zinkhan, G.M. (2002). U-commerce: Extending the universe of marketing. Journal of the Academy of Marketing Science, 30(4), 329-343. Zwass, V. (1996). Electronic commerce: Structures and issues. International Journal of Electronic Commerce, 1(1), 3-23.
KEY TERMS Maneuverability: The perceived usefulness in the adoption of MCDs and the degree to which a person can make the best use of such MCDs; one of the influencing factors when consumers consider adopting MCDs for purchases. Matching: The need for the standardized and common interface of MCDs. Maturity: The possibility that the technology of the MCD is mature enough so that there will not be any possible significant improvements at a later stage; one of the influencing factors when consumers consider adopting MCDs for purchases. MCD: M-commerce device. Mentality: The match between the new technology and consumers’ own mindsets, as well as the appropriate recognition of their peer groups; one of the influencing factors when consumers consider adopting MCDs for purchases. Merits: The degree to which a buyer believes that MCDs can provide significant improvement in the purchase process; one of the influencing factors when consumers consider adopting MCDs for purchases. Options Model: A model that proposes that consumers have the option of not adopting until the new technology, in terms of performance, is substantially better than the old technology, and as a result of such options, a new technology with a moderate expected improvement in performance can experience substantial delays in adoption and price distortions even in a competitive market. Technology Acceptance Model (TAM): An information systems theory that models how users come to accept and use a new technology, with reference to two major considerations, perceived usefulness and perceived ease of use.
Category: M-Business and M-Commerce
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce Michele Ruta Politecnico di Bari, Italy Tommaso Di Noia Politecnico di Bari, Italy Eugenio Di Sciascio Politecnico di Bari, Italy Francesco Maria Donini Università della Tuscia, Italy Giacomo Piscitelli Politecnico di Bari, Italy
INTRODUCTION
STATE OF THE ART
New mobile architectures allow for stable networked links from almost everywhere, and more and more people make use of information resources for work and business purposes on mobile systems. Although technological improvements in the standardization processes proceed rapidly, many challenges, mostly aimed at the deployment of value-added services on mobile platforms, are still unsolved. In particular the evolution of wireless-enabled handheld devices and their capillary diffusion have increased the need for more sophisticated service discovery protocols (SDPs). Here we present an approach, which improves Bluetooth SDP, to provide m-commerce resources to the users within a piconet, extending the basic service discovery with semantic capabilities. In particular we exploit and enhance the SDP in order to identify generic resources rather than only services. We have integrated a “semantic layer” within the application level of the standard Bluetooth stack in order to enable a simple interchange of semantically annotated information between a mobile client performing a query and a server exposing available resources. We adopt a simple piconet configuration where a stable networked zone server, equipped with a Bluetooth interface, collects requests from mobile clients and hosts a semantic facilitator to match requests with available resources. Both requests and resources are expressed as semantically annotated descriptions, so that a semantic distance can be computed as part of the ranking function, to choose the most promising resources for a given request.
Usually, resource discovery protocols involve a requester, a lookup or directory server and finally a resource provider. Most common SDPs, as service location protocol (SLP), Jini, UPnP (Universal Plug aNd Play), Salutation or UDDI (universal description discovery and integration), include registration and lookup of resources as well as matching mechanisms (Barbeau, 2000). All these systems generally work in a similar manner. Basically a client issues a query to a directory server or to a specific resource provider. The request may explicitly contain a resource name with one or more attributes. The lookup server—or directly the resource provider—attempts to match the query pattern with resource descriptions stored in its database, then it replies to the client with discovered resources identification and location (Liu, Zhang, Li, Zhu, & Zhang, 2002). These discovery architectures are based on some common assumptions about network infrastructure under the application layer in the protocol stack. In particular, current SDPs usually require a continuous and robust network connectivity, which may not be the case in wireless contexts, and especially in the ad-hoc ones. In fact in such environments, network consistence varies continuously and temporary disconnections occur frequently, so bringing to a substantial decrease traditional SDP performances (Chakraborty, Perich, Avancha, & Joshi, 2001). Actually, there are several issues that restrain the expansion of advanced wireless applications, among them, the variability of scenarios. An ad-hoc environment is based on short-range, low power technologies like Bluetooth
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
A
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
(Bluetooth, 1999), which grant the peer-to-peer interaction among hosts. In such a mobile infrastructure there could be one or more devices providing and using resources but, as a MANET is a very unpredictable environment, a flexible resource search system is needed to overcome difficulties due to the host mobility. Furthermore, existing mobile resource discovery methods use simple string-matching, which is largely inefficient in advanced scenarios as the ones related to electronic commerce. In fact, in these cases there is the need to submit articulate requests to the system to obtain adequate responses (Chakraborty & Chen, 2000). With specific reference to the SDP in the Bluetooth stack, it is based on a 128-bit universally unique identifier (UUID); each numeric ID is associated to a single service class. In other words, Bluetooth SDP is code-based and consequently it can handle only exact matches. Yet, if we want to search and retrieve resources whose description cannot be classified within a rigid schema (e.g., the description of goods in a shopping mall), a more powerful discovery architecture is needed (Avancha, Joshi, & Finin, 2002). SDP should be able to cope with non-exact matches (Chakraborty & Chen, 2000), and to provide a ranked list of discovered resources, computing a distance between each retrieved resource and the request after a matchmaking process. To achieve these goals, we exploit both theoretical approach and technologies of semantic Web vision and adapt them to small ad-hoc networks based on the Bluetooth technology (Ruta, Di Noia, Di Sciascio, Donini, & Piscitelli, 2005). In a semantic-enabled Web—what is known as the semantic Web vision—each available resource should be annotated using RDF (RDF Primer, 2004), with respect to an OWL ontology (Antoniou & van Harmelen, 2003). There is a close relation between the OWL-DL subset of OWL and description logics (DLs) (Baader, Calvanese, McGuinness, Nardi, & Patel-Schneider, 2002) semantics, which allows the use of DLs-based reasoners in order to infer new information from the one available in the annotation itself. In the rest of the article we will refer to DIG (Bechhofer, 2003) instead of OWL-DL because it is less verbose and more compact: a good characteristic in an ad-hoc scenario. DIG can be seen as a syntactic variant of OWL-DL.
THE PROPOSED APPROACH In what follows we outline our framework and we sketch the rationale behind it. We adopt a mobile commerce context as reference scenario. In our mobile environment, a user contacts via Bluetooth a zone resource provider (from now on hotspot) and submits her semantically annotated request in DIG formalism. We assume the zone server—which classifies resource contents by means of an OWL ontology—has previously identified
shopping malls willing to promote their goods and it has already collected semantically annotated descriptions of goods. Each resource in the m-marketplace owns an URI and exposes its OWL description. The hotspot is endowed with a MatchMaker [in our system we adapt the MAMAS-tng reasoner (Di Noia, Di Sciascio, Donini, & Mongiello, 2004)], which carries out the matchmaking process between each compatible offered resource and the requested one measuring a “semantic distance.” The provided result is a list of discovered resources matching the user demand, ranked according to their degree of correspondence to the demand itself. By integrating a semantic layer within the OSI Bluetooth stack at service discovery level, the management of both syntactic and semantic discovery of resources becomes possible. Hence, the Bluetooth standard is enriched by new functionalities, which allow to maintain a backward compatibility (handheld device connectivity), but also to add the support to matchmaking of semantically annotated resources. To implement matchmaking and ontology support features, we have introduced a semantic service discovery functionality into the stack, slightly modifying the existing Bluetooth discovery protocol. Recall that SDP uses a simple request/response method for data exchange between SDP client and SDP server (Gryazin, 2002). We associated unused classes of 128-bit UUIDs in the original Bluetooth standard to mark each specific ontology and we call this identifier OUUID (ontology universally unique identifier). In this way, we can perform a preliminary exclusion of supply descriptions that do not refer to the same ontology of the request (Chakraborty, Perich, Avancha, & Joshi, 2001). With OUUID matching we do not identify a single service, but directly the context of resources we are looking for, which can be seen as a class of similar services. Each resource semantically annotated is stored within the hotspot as resource record. A 32-bit identifier is uniquely associated to a semantic resource record within the hotspot, which we call SemanticResourceRecordHandle. Each resource record contains general information about a single semantic enabled resource and it entirely consists of a list of resource attributes. In addition to the OUUID attribute, there are ResourceName, ResourceDescription, and a variable number of ResourceUtilityAttr_i attributes (in our current implementation 2 of them). ResourceName is a text string containing a human-readable name for the resource, the second one is a text string including the resource description expressed in DIG formalism and the last ones are numeric values used according to specific applications. In general, they can be associated to context-aware attributes of a resource (Lee & Helal, 2003), as for example its price or the physical distance it has from the hotspot (expressed in metres or in terms of needed time to get to the resource). We use them as parameters of the overall utility function that computes matchmaking results.
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
Table 1. List of PDU IDs with corresponding descriptions PDU ID
Description
0x00
Reserved
0x01
SDP_ErrorResponse
0x02
SDP_ServiceSearchRequest
0x03
SDP_ServiceSearchResponse
0x04
SDP_ServiceAttributeRequest
0x05
SDP_ServiceAttributeResponse
0x06
SDP_ServiceSearchAttributeRequest
0x07
SDP_ServiceSearchAttributeResponse
0x08
SDP_OntologySearchRequest
0x09
SDP_OntologySearchResponse
0x0A
SDP_SemanticServiceSearchRequest
0x0B
SDP_SemanticServiceSearchResponse
0x0C-0xFF
Reserved
Table 2. SDP_OntologySearchRequest PDU parameters PDU ID
1.
The user searches for a specific ontology identifier by submitting one or more OUUIDR she manages by means of her client application
-
0x08
OntologySearchPattern ContinuationState
Table 3. SDP_OntologySearchResponse PDU parameters PDU ID 0x09
2. 3. 4.
In particular, to allow the representation and the identification of a semantic resource description we introduced in the data representation of the original Bluetooth standard two new data element type descriptor: OUUID and DIG text string. The first one is associated to the type descriptor value 9 whereas to the second one corresponds the type descriptor value 10 (both reserved in the original standard). We will associate 1, 2, 4 byte as valid size for the first one and 5, 6, 7 for the DIG text string. Since the communication is referred to the peer layers of the protocol stack, each transaction is represented by one request Protocol Data Unit (PDU) and another PDU as response. If the SDP request needs more than a single PDU (this case is frequent enough if we use semantic service discovery) the SDP server generates a partial response and the SDP client waits for the next part of the complete answer. By adding two SDP features SDP_OntologySearch (request and response) and SDP_SemanticServiceSearch (request and response) to the original standard (exploiting not used PDU ID) we inserted together with the original SDP capabilities further semantic-enabled resource search functions (see Table 1). The transaction between service requester and hotspot starts after ad-hoc network creation. When a user becomes a member of a MANET, she is able to ask for a specific service/resource (by submitting a semantic-based description). The generic steps, up to response providing, for a service request are detailed in the following:
parameters
5.
6.
parameters -
TotalOntologyCount OntologyRetrievedPattern ContinuationState
The hotspot selects OUUIDs matching each OUUIDR and replies to the client The user sends a service request (R) to the hotspot The hotspot extracts descriptions of each resource cached within the hotspot itself, which is classified with the previously selected OUUIDR The hotspot performs the matchmaking process between R and selected resources it shares. Taking into account the matchmaking results, all the resources are ranked with respect to R The hotspot replies to the user.
It is important to remark that basically all the previous steps are based on the original SDP in Bluetooth. No modifications are made to the original structure of transactions, but simply we differently use the SDP framework. In what follows we outline the structure of the SDP PDUs we added within the original framework to allow semantic resource discovery. The first one is the SDP_OntologySearchRequest PDU. Their parameters are shown in Table 2. The OntologySearchPattern is a data element sequence where each element in the sequence is a OUUID. The sequence must contain at least 1 and at most 12 OUUIDs, as in the original standard. The list of OUUIDs is an ontology search pattern. The ContinuationState parameter maintains the same purpose of the original Bluetooth (Bluetooth, 1999). The SDP_OntologySearchResponse PDU is generated by the previous PDU. Their parameters are reported in Table 3. The TotalOntologyCount is an integer containing the number of ontology identifiers matching the requested ontology pattern. Whereas the OntologyRetrievedPattern is a data element sequence where each element in the sequence is a OUUID matching at least one sent with the OntologySearchPattern. If no OUUID matches the pattern,
A
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
Table 4. SDP_SemanticServiceSearchRequest PDU parameters PDU ID
0x0A
parameters -
SemanticResourceDescription ContextAwareParam1 ContextAwareParam2 MaximumResourceRecordCount ContinuationState
the TotalOntologyCount is set to 0 and the OntologyRetrievedPattern contains only a specific OUUID able to allow the browsing by the client of all the OUUIDs managed by the hotspot (see the following ontology browsing mechanism for further details). Hence the pattern sequence contains at least 1 and at most 12 OUUIDs. The SDP_SemanticServiceSearchRequest PDU follows previous PDU. Their parameters are shown in Table 4. The SemanticResourceDescription is a data element text string in DIG formalism representing the resource we are searching for; ContextAwareParam1 and ContextAwareParam2 are data element unsigned integers. In our case study, which models an m-marketplace in an airport terminal, we use them respectively to indicate a reference price for the resource and the hour of the scheduled departure of the flight. Since a generic client interacting with a hotspot is in its range, using the above PDU parameter she can impose—among others—a proximity criterion in the resource discovery policy. The SDP_SemanticServiceSearchResponse PDU is generated by the previous PDU. Their parameters are reported in Table 5. The SemanticResourceRecordHandleList includes a list of resource record handles. Each of the handles in the list refers to a resource record potentially matching the request. Note that this list of service record handles does not contain header fields, but only the 32-bit record handles. Hence, it does not have the data element format. The list of handles is arranged according to the relevance order of resources, excluding resources not compatible with the request. The other parameters maintain the same purpose of the original Bluetooth (Bluetooth, 1999). In all the previous cases, the error handling is managed with the same mechanisms and techniques of Bluetooth standard (Bluetooth, 1999). Notice that each resource retrieval session starts after settling between client and server the same ontology identifier (OUUID). Nevertheless if a client does not support any ontology or if the supported ontology is not managed by the hotspot, it is desirable to discover what kind of merchandise class (and then what OUUIDs) are handled by the zone server
Table 5. SDP_SemanticServiceSearchResponse PDU parameters PDU ID 0x0B
parameters -
TotalResourceRecordCount CurrentResourceRecordCount SemanticResourceRecordHandleList ContinuationState
without any a priori information about resources. For this purpose we use the service browsing feature (Bluetooth, 1999) in a slightly different fashion with respect to the original Bluetooth standard, so calling this mechanism ontology browsing. It is based on an attribute shared by all semantic enabled resource classes, the BrowseSemanticGroupList attribute which contains a list of OUUIDs. Each of them represents the browse group a resource may be associated with for browsing. Browse groups are organized in a hierarchical fashion, hence when a client desires to browse a hotspot merchandise class, she can create an ontology search pattern containing the OUUID that represents the root browse semantic group. All resources that may be browsed at the top level are made members of the root browse semantic group by having the root browse group OUUID as a value within the BrowseSemanticGroupList attribute. Generally a hotspot supports relatively few merchandise classes, hence all of their resources will be placed in the root browse group. However, the resources exposed by a provider may be organised in a browse group hierarchy, by defining additional browse groups below the root browse group. Having determined the goods category and the corresponding reference ontology, the client can also download a DIG version of it from the hotspot as .jar file [such a file extension—among other things—also allows a total compatibility with the Connected Limited Device Configuration (CLDC) technology]. Also notice that since the proposed approach is fully compliant with semantic Web technologies, the user exploits the same semantic enabled descriptions she may use in other Semantic Web compliant systems (e.g., in the Web site of a shopping mall). That is, there is no need for different customized resource descriptions and modelling, if the user employs different applications either on the Web or in mobile systems. The syntax and formal semantics of the descriptions is unique with respect to the reference ontology and can be shared among different environments. In e-commerce scenarios, the match between demand and supply involves not only the description of the good but also data-oriented properties. It would be quite strange to have a commercial transaction without taking into account
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
price, quantity, and availability, among others. The demander usually specifies how much she is willing to pay, how many items she wants to buy, and the delivery date. Hence, the overall match value depends not only on the distance between the (semantic-enabled) description of the demand and of the supply. It has to take into account the description distance with the difference of (the one asked by the demander and the other proposed by the seller), quantity, and delivery date. The overall utility function combines all these values to give a global value representing the match degree. Also notice that, in m-commerce applications, in addition to “commercial” parameters also context-aware variables should influence matching results. For example, in our airport case study, we consider the price difference but also the physical distance between requester and seller to weigh the match degree. The distance becomes an interesting value since a user has a temporal deadline for shopping: the scheduled hour of her flight. Hence, a resource might be chosen also according to its proximity to the user. We will express this distance in terms of time to elapse for reaching the shop where a resource is, leaving from the hotspot area. In such a manner the hotspot will exclude resources not reachable by the user while she is waiting for boarding and it will assign to resources unlikely reachable (farther) a weight smaller than one assigned to easily reachable ones. The above approach can be further extended to other data-type properties. The utility function we used depends on: • • • • •
pD : price specified by the demander pO : price specified by the supplier tD : time interval available to the client tO : time to reach the supplier and come back, leaving from the hotspot area s_match: score computed during the semantic matchmaking process, computed through rankPotential (Di Noia, Di Sciascio, Donini, & Mongiello, 2004) algorithm. t D − tO
RUNNING EXAMPLE A simple example can clarify the rationale of our setting. Here we will present a case study analogous to the one presented in Avancha, Joshi, and Finin (2002), and we face it by means of our approach. Let us suppose a user is in a duty free area of an airport, she is waiting for her flight to come back home and she is equipped with a wireless-enabled PDA. She forgot to buy a present for her beloved little nephew and now she wants to purchase it from one of the airport gift stores. In particular she is searching for a learning toy strictly suitable for a kid (she dislikes a child toy or a baby toy) and possibly the toy should not have any electric power supply. Clearly this request is too complex to be expressed by means of standard UUID Bluetooth SDP mechanism. In addition, non-exact matches between resource request and offered ones is highly probable and the on/off matching system provided by the original standard in this case could be largely inefficient. Hence both the semantic resource request and offered ones can be expressed in a DIG statement exploiting DL semantics and encapsulated in an SDP PDU. The hotspot equipped with MAMAS reasoner collects the request and initially selects supplies expressed by means of the same ontology shared with the requester. Hence a primary selection of suitable resources is performed. In addition, the matchmaker carries out the matchmaking process between each offered resource in the m-marketplace and the requested one measuring a “semantic distance” (Colucci, Di Noia, Di Sciascio, Donini, & Mongiello, 2005). Finally the matchmaking results are ranked and returned to the user. A subset of the ontology used as a reference in the examples is reported in Figure 1. For the sake of simplicity, only the class hierarchy and disjoint relations are represented. Let us suppose that after the hotspot selects supplies, its knowledge base is populated with the following individuals whose description is represented using DL formalism: •
Alice_in_wonderland. Price 20$. 5 min from the hotspot:
(1)
•
Barbie_car. Price 80$. 10 min from the hotspot:
Notice that pD is weighted by a (1+α) factor. The idea behind this weight is that, usually, the demander is willing to pay up to some more than what she originally specified on condition that she finds the requested item, or something very similar. In the tests we carried out, we find α=0.1 and β=10 are values in accordance with user preferences. These values seem to be in some accordance with experience, but they could be changed according to different specific considerations.
•
classic_guitar. Price 90$. 17 min from the hotspot:
•
shape_order. Price 40$. 15 min from the hotspot:
s _ match u ( s _ match, pD , pO , t D , tO ) = + 2
tanh
3
+
(1 + )pD − pO 6(1 + )pD
•
book ∀has_genre.fantasy
car ∀suggested_for.girl ∀has_power_supply.battery
musical_instrument ∀suitable_for.kid (£ 0 has_power_ supply) educational_tool ∀suitable_for.child ∀stimulates_to_learn. shape_and_color
Playstation. Price 160$. 28 min from the hotspot: video_game ∀has_power_supply.DC
A
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
Figure 1. The simple toy store ontology used as reference in the example
•
Winnie_the_pooh. Price 30$. 15 min from the hotspot: teddy_bear ∀suitable_for.baby
On the other hand, the request D submitted to the system by the user can be formalized in DL syntax as follows: learning_toy ∀suggested_for.boy ∀suitable_for.kid (£ 0 has_power_supply)
In addition she imposes a reference price of 200$ (pD=200) as well as the scheduled departure time as within 30 minutes (tD=30). In Table 6 matchmaking results are presented. The second column shows whether each retrieved resource is compatible
or not with request D and, in case, the rankPotential computed result. In the fourth column, matchmaking results are also expressed in a relative form between 0 and 1 to allow a more immediate semantic comparison among requests and different resources and to put in a direct correspondence various rank values. Finally in the last column results of the overall utility function application are shown. Notice that the semantic distance of the individual classic_guitar from D is the smaller one; then the system will recommend to the user this resource first. Hence the ranked list returned by the hotspot is a strict indication for the user about best available resources in the airport duty free piconet in order of relevance with respect to the request. Nevertheless
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
Table 6. Matchmaking results
A
demand – supply
compatibility (y/n)
score
s_match
u(٠)
D - Alice_in_wonderland
n
-
-
-
D - Barbie_car
y
7
0.364
0.609
D - classic_guitar
y
3
0.727
0.748
D - shape_order
n
-
-
-
D - Playstation
y
5
0.546
0.378
D - Winnie_the_pooh
n
-
-
-
a user can choose or not a resource according to her personal preferences and her initial purposes. After having selected the best resource, the server of the chosen virtual shop will receive a connection request from the user PDA with its connection parameters and in this manner the transaction may start. The user can provide her credit card credentials, so that when she reaches the store, her gift will be already packed. This final part of the application is not yet implemented, but it is trivially achievable exploiting the above SDP infrastructure.
CONCLUSION AND FUTURE WORK In this article we have presented an advanced semantic-enabled resource discovery protocol for m-commerce applications. The proposed approach aims to completely recycle the basic functionalities of the original Bluetooth service discovery protocol by simply adding semantic capabilities to the classic SDP ones and without introducing any change in the regular communication work of the standard. A matchmaking algorithm is used to measure the semantic similarity among demand and resource descriptions. Future trends of the proposed framework aim to create a more advanced DSS to help a user in a generic m-marketplace. Under investigation is the support to creation of P2P small communities of mobile hosts where goods and resources are advertised and opinions about shopping are exchanged (Avancha, D’Souza, Perich, Joshi, & Yesha, 2003). If a user decides to “open” her shopping trolley sharing information she owns (purchased goods, discounts, opinion about specific vendors or products) the system will insert her in a buyer mobile community where she can exchange information with other users. Another future activity focuses on strict control of the good advertising. In an m-marketplace, the system will send to various potential buyers best proposals about their interests.
We intend to implement a mechanism to advertise goods or services in a more direct and personalized fashion. From this point of view, an additional feature of the system is oriented to the user profiling extraction and management (Prestes, Carvalho, Paes, Lucena, & Endler, 2004; Ruta, Di Noia, Di Sciascio, Donini, & Piscitelli, 2005; von Hessling, Kleemann, & Sinner, 2004). Without imposing any explicit profile submission to the user, the system could collect her preferences by means of previously submitted requests (Ruta, Di Noia, Di Sciascio, Donini, & Piscitelli, 2005); that is, by means of the “history” of the user in the m-marketplace.
REFERENCES Antoniou, G., & van Harmelen, F. (2003). Web ontology language: OWL. In Handbook on Ontologies in Information Systems. Avancha, S., D’Souza, P., Perich, F., Joshi, A., & Yesha, Y. (2003). P2P m-commerce in pervasive environments. ACM SIGecom Exchanges, 3(4), 1-9. Avancha, S., Joshi, A., & Finin, T. (2002). Enhanced service discovery in Bluetooth. IEEE Computer, 35(6), 96-99. Baader, F., Calvanese, D., McGuinness, D., Nardi, D., & Patel-Schneider, P. (2002). The description logic handbook. Cambridge: Cambridge University Press. Barbeau, M. (2000). Service discovery protocols for ad hoc networking. Workshop on Ad-hoc Communications (CASCON ’00). Bechhofer, S. (2003). The DIG description logic interface: DIG/1.1. Retrieved from http://dlweb.man.ac.uk/ dig/2003/02/interface.pdf. Bluetooth specification document. (1999). Retrieved from http://www.bluetooth.com.
Advanced Resource Discovery Protocol for Semantic-Enabled M-Commerce
Chakraborty, D., & Chen, H. (2000). Service discovery in the future for mobile commerce. ACM Crossroads, 7(2), 18-24. Chakraborty, D., Perich, F., Avancha, S., & Joshi, A. (2001). Dreggie: Semantic service discovery for m-commerce applications. In Workshop on Reliable and Secure Applications in Mobile Environment. Colucci, S., Di Noia, T., Di Sciascio, E., Donini, F. M., & Mongiello, M. (2005). Concept abduction and contraction for semantic-based discovery of matches and negotiation spaces in an e-marketplace. Electronic Commerce Research and Applications, 4(4), 345-361. Di Noia, T., Di Sciascio, E., Donini, F. M., & Mongiello, M. (2004). A system for principled matchmaking in an electronic marketplace. International Journal of Electronic Commerce, 8(4), 9-37. Gryazin, E. (2002). Service discovery in Bluetooth. Retrieved from http://www.hpl.hp.com/techreports/2002/HPL-2002233.pdf. Lee, C., & Helal, S. (2003). Context attributes: An approach to enable context awareness for service discovery. In Symposium on Applications and the Internet (SAINT ’03) (pp. 22-30). Liu, J., Zhang, Q., Li, B., Zhu, W., & Zhang, J. (2002). A unified framework for resource discovery and QoS-aware provider selection in ad hoc networks. ACM Mobile Computing and Communications Review, 6(1), 13-21. Prestes, R., Carvalho, G., Paes, R., Lucena, C., & Endler, M. (2004). Applying ontologies in open mobile systems. In Workshop on Building Software for Pervasive Computing OOPSLA’04. RDF Primer-W3C Recommendation. (2004, February 10). Retrieved from http://www.w3.org/TR/rdf-primer/ Ruta, M., Di Noia, T., Di Sciascio, E., Donini, F.M., & Piscitelli, G. (2005). Semantic based collaborative P2P in ubiquitous computing. In IEEE/WIC/ACM International Conference Web Intelligence 2005 (WI ’05) (pp. 143-149). von Hessling, A., Kleemann, T., & Sinner, A. (2004). Semantic user profiles and their applications in a mobile environment. In Artificial Intelligence in Mobile Systems 2004.
0
KEY TERMS Description Logics (DLs): A family of logic formalisms for knowledge representation. Basic syntax elements are concept names, role names, and individuals. Intuitively, concepts stand for sets of objects, and roles link objects in different concepts. Individuals are used for special named elements belonging to concepts. Basic elements can be combined using constructors to form concept and role expressions, and each DL has its own distinct set of constructors. DLbased systems are equipped with reasoning services: logical problems whose solution can make explicit knowledge that was implicit in the assertions. M-Marketplace: Virtual environment where demands and supplies (submitted or offered by users equipped with mobile devices) encounter each other. Ontology: An explicit and formal description referred to concepts of a specific domain (classes) and to relationships among them (roles or properties). Piconet: Bluetooth-based short-range wireless personal area network. A Bluetooth piconet can host up to eight mobile devices. More piconets form a scatternet. Service Discovery Protocol (SDP): It identifies the application layer of an OSI protocol stack and manages the automatic detection of devices with joined services. Semantically Annotated Resource: any kind of good, tangible or intangible (e.g., a document, an image, a product or a service) endowed of a description that refers to a shared ontology. Semantic Matchmaking: The process of searching the space of possible matches between a request and several resources to find those best matching the request, according to given semantic criteria. It assumes that both the request and the resources are annotated according to a shared ontology.
Category: Wireless Networking
Anycast-Based Mobility
A
István Dudás Budapest University of Technology and Economics, Hungary László Bokor Budapest University of Technology and Economics, Hungary Sándor Imre Budapest University of Technology and Economics, Hungary
INTRODUCTION We have entered the new millennium with two great inventions, the Internet and mobile telecommunication, and a remarkable trend of network evolution toward convergence of these two achievements. It is an evident step to combine the advantages of the Internet and the mobile communication methods together in addition to converge the voice and data into a common packet-based and heterogeneous network infrastructure. To provide interworking, the future systems have to be based on a universal and widespread network protocol, such as Internet protocol (IP) which is capable of connecting the various wired and wireless networks (Macker, Park, & Corson, 2001). However, the current version of IP has problems in mobile wireless networks; the address range is limited, IPv4 is not suitable to efficiently manage mobility, support real-time services, security, and other enhanced features. The next version, IPv6 fixes the problems and also adds many improvements to IPv4, such as extended address space, routing, quality of service, security (IPSec), network autoconfiguration and integrated mobility support (Mobile IPv6). Today’s IP communication is mainly based on unicast (one-to-one) delivery mode. However it is not the only method in use: other delivery possibilities, such as broadcast (one-to-all), multicast (one-to-many) and anycast (one-toone-of-many) are available. Partridge, Mendez, and Milliken (1993) proposed the host anycasting service for the first time in RFC 1546. The basic idea behind the anycast networking paradigm is to separate the service identifier from the physical host, and enable the service to act as a logical entity of the network. This idea of anycasting can be achieved in different layers (e.g., network and application layers) and they have both strengths and weaknesses as well. We focus on network-layer anycasting in this article, where a node sends a packet to an anycast address and the network will deliver the packet to at least one, and preferably only one of the competent hosts. This approach makes anycasting a kind of group communication in that a group of hosts are specified for a service represented by an anycast address and
underlying routing algorithms are supposed to find out the appropriate destination for an anycast destined packet.
OvERvIEW OF IPv6 ANYCASTING RFC 1546 introduced an experimental anycast address for IPv4, but in this case the anycast addresses were distinguishable from unicast addresses. IPv6 adopted the paradigm of anycasting as one of the basic and explicitly included services of IP and introduced the new anycast address besides the unicast and multicast addresses (Deering & Hinden, 1998). IPv6 anycast addresses were designed to allow reaching a single interface out of a group of interfaces. The destination node receiving the sent packets is the “nearest” node. The distance is dependent on the metric of the underlying routing protocol. In case of IPv6, an anycast address is defined as a unicast address assigned to more than one interface, so anycast addresses can not be distinguished from unicast addresses: they both share the same address space. Therefore the beginning part of any IPv6 anycast address is the network prefix. The longest P prefix identifies the topological region in which all interfaces are belonging to that anycast address reside. In the region identified by P, each member of the anycast membership must be handled as a separate entry of the routing system. Based on the length of P, IPv6 anycast can be categorized into two types: subnet anycast and global anycast. Hashimoto, Ata, Kitamura, and Murata (2005) summarized all that issues and defined the main terminology of IPv6 anycasting (Figure 1). Hinden and Deering (2003) declared some restrictions concerning the further usage of the anycast addressing paradigm. The main purpose for setting these limitations was to keep the usage of anycast addresses under control until enough experience has been gathered in order to fit this new scheme to the existing structure of the Internet. These restrictions are now being eased that research could find appropriate solution for them (Abley, 2005). The biggest concern that had to be dealt with was routing since anycast packets (packets with an anycast address in the destination
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Anycast-Based Mobility
Figure 1. IPv6 anycast terminology basics
field) might be forwarded to domains with different prefixes, as anycast receivers might be distributed all over the Internet. As a result a scalable and stable routing solution for anycasting is necessary.
Routing Protocols for IPv6 Anycasting The current IPv6 standards do not define the anycast routing protocol, although the routing is one of the most important elements of network-layer anycasting. There is a quite small amount of literature about practical IPv6 anycasting. Park and Macker (1999) proposed and evaluated anycast extensions of link-state routing algorithm and distance-vector routing algorithm. Xuan, Jia,, Zhao, and Zhu (2000) proposed and compared several routing algorithms for anycast. Eunsoo Shim (2004) proposed an application load sensitive anycast routing method (ALSAR) and analyzed the existing routing algorithms in his PhD thesis. Doi, Ata, Kitamura, and Murata (2004) summarized the problems and possible solutions regarding the current specifications for IPv6 anycasting and proposed an anycast routing architecture based on seed nodes, gradual deployment and the similarities to multicasting. Based on their work, Matsunaga, Ata, Kitamura,and Murata (2005) designed and implemented three IPv6 anycast routing protocols (AOSPF—anycast open shortest path first, ARIP—anycast routing information protocol and PIASM—protocol independent anycast - sparse mode) based on existing multicast protocols. The recent studies are focusing on subnet anycast routing protocols since they offer various possibilities for research while global anycast routing still faces scalability problems to be solved. The recently introduced anycast routing protocols all share a common ground as they are all based on multicast routing protocols because of the similarities of the two addressing schemes.
Unfortunately it does not fit the scope of this document to introduce each anycast routing protocol one-by-one although it is important to present the main idea that lies beneath all these protocols. The principal task to be performed is to discover all the anycast capable routers and nodes in the network: this can happen by flooding (as in case of AOSPF) or discovery methods (e.g., PIA-SM). The next, and maybe the most important step, is to maintain an up-to-date anycast routing table so all possible receivers could be reached in case of need. The easiest way to keep the routing entries up-todate is to maintain a so-called Anycast Group Membership (Figure 1) where the anycast hosts can sign in or out when joining or leaving a certain anycast group designated by its anycast address.
APPLICATIONS OF ANYCASTING Since the introduction of IPv6 anycast only a few applications have emerged using these addresses. It is mainly because the flexibility of the anycasting paradigm has not yet been widespread in the public. An excellent survey of the IPv6 anycast characteristics and applications was made by Weber and Cheng, 2004; Doi, Ata, Kitamura, and Murata, 2004; Matsunaga, Ata, Kitamura, and Murata (2005), where the authors describe many advantages and possible applications of anycasting. These applications can be classified into the following main types.
Main Application Schemas The most popularly known application of anycast technology is helping the communicating nodes in selection of service providing servers. In the server selection approach the client host can choose one of many functionally identical servers.
Anycast-Based Mobility
The anycast server location and selection method could be a simple and transparent technique since the same address can be used from anywhere in the network, and the anycast routing would automatically choose the best destination for the client. Anycast addresses can also be useful in discovering and locating services. In case of service discovery, the clients just need to know only one address: they can communicate with an optimal (e.g., minimum delay) host selected from the anycast group and easily discover the closest provider. This is especially beneficial in case of dynamically and frequently changing environments such as mobile ad-hoc systems. Services based on this characteristic can be acquired easily and optimally by the mobile clients through network-layer anycasting.
Application Scenarios The most important advantage of network-layer anycasting is its ability to provide a simple mechanism where the anycast initiator (Figure 1) can receive a specific service without exact information about the server nodes and networks. Moreover the whole procedure is totally transparent: the clients do not need to know whether the server’s address is unicast or anycast, because anycast addresses are syntactically indistinguishable from unicast addresses. Only servers have additional knowledge about their explicitly configured anycast addresses. The main application schemas and the application scenarios below are demonstrating the possibilities of the anycasting communication paradigm. With the help of IPv6 anycasting local information services (e.g., emergency calls) can be given by getting each node to communicate with the appropriate server to the node’s actual location. This kind of application is very useful in a mobile environment where nodes move from one network to another while resorting a given service. By assigning a well-known anycast address to widespread applications, we can achieve host auto-configuration. The clients can use these services without knowing the appropriate unicast address of the server. The clients can utilize these applications everywhere only by specifying the service’s well-known anycast address. For example, DNS resolvers no longer have to be configured with the unicast IP addresses for every host in every network if a standardized anycast address is built in the hardware or software, end users can get the service without configuration. Improving the system reliability is another good example of IPv6 anycasting. Anycast communication grants multiple numbers of hosts with the same address and by increasing the number of hosts load balancing, service redundancy and DoS attack avoidance can be achieved based on the routing mechanism where anycast requests are fairly forwarded. In a widely distributed environment (like a peer-to-peer architecture) services can construct a logical topology above
the physical network. This logical topology can be based on anycast addresses. When a client wants to participate, it specifies the anycast address of the logical level in order to join in the logical network. In such a way, one of the participating nodes will become the gate of the logical network determined by the underlying anycast routing protocol. As we can see, there are some real promising application scenarios of IPv6 anycasting. However there is only one standardized anycast application these days, called dynamic home agent address discovery. In Mobile IPv6 the home agents (HA) have an anycast address, since the HA may change while the mobile terminal is not attached to its home network. Therefore a mobile node should use the anycast address of the home agents to reach one HA out of the set of home agents on its home link.
EMERGING APPLICATION: ANYCAST-BASED MICROMOBILITY In Mobile IPv6 every mobile node (MN) is identified by its home address (HA), totally independently of where it is located in the network. When a MN is away from its home network (HN), it gets a new care-of-address (CoA). The IPv6 packets sent to the mobile node’s HA will be routed to the mobile node’s new CoA (Johnson, Perkins, & Arkko, 2004). Although Mobile IPv6 is capable of handling global mobility of users, it has shortcomings in supporting low latency and packet loss—required by real time multimedia services—during handover. To improve handover performance, the movement of a mobile node inside a subnet has to be dealt locally, by hiding intra-domain movements. As a result, the number of signaling messages reduced and the handover performance improved. Inside such a local subnet—called micromobility domain—the terminal receives a temporal IP address, which is valid throughout the subnet, and can be used a temporal CoA for the HA while the mobile terminal is located in the micromobility domain. Inside the micromobility domain, micromobility protocols are responsible for the proper routing of packets intended to the mobile hosts (Saha, Mukherjee, Misra, & Chakraborty, 2004). Leaving the micromobility domain, Mobile IPv6 provides global mobility management. We have developed a new type of anycast application based on the main characteristics and the new research achievements of IPv6 anycasting in order to provide micromobility support in a standard IPv6 environment. In our proposed scheme, anycast addresses are used to identify mobile IPv6 hosts entering a micromobility domain while the underlying anycast routing protocol is used to maintain the anycast address routing information exchange. As a result the care-of-address obtained if the mobile terminal moves into a micromobility area is an anycast address. According to our proposal an anycast address identifies a single mobile
A
Anycast-Based Mobility
Figure 2. Entering a foreign micromobility domain
node. Therefore IP packets sent to the CoA of the mobile terminal have no chance to reach another “nearest” mobile node, since in this sense anycast addresses identifying mobile nodes are unique. The mobile node with a unique anycast care-of-address matches the correspondent anycast responder (CAR) in anycasting terminology. Also it has to be noted that in case of anycast address-based mobility there is no need for a peer unicast address since the CoA obtained is unique. The reason why unique anycast address is used instead of unicast address is the fact that anycast addresses are valid in the whole micromobility domain. Therefore the same anycast address can not be assigned to a second mobile node in a given micromobility domain. The mobile node with a unique anycast address forms a virtual group. The members of this virtual group are the possible positions of the mobile node in the micromobility domain (that equals the validity area of the anycast address defined by the anycast P prefix) and the “nearest” mobile equipment is at the actual position of the mobile node. Therefore the mobile node remains reachable at any time (Figure 2). The purpose of using anycast address as an identifier for mobile nodes is that routing and handover management can be simplified with the help of changing the routing metrics. With the proper selection of the P prefix, the size of the virtual anycast group (VAG) can be adjusted easily. The virtual anycast group equals anycast group membership (AGM), while the virtual copies of the mobile node
match the anycast responders. The operation of the anycast addressing-based mobility has to be investigated in case of different scenarios. In the first scenario the mobile terminal leaves its current domain (e.g., its home network) and enters (1) another local administrative mobility domain (a new micromobility domain), as seen in Figure 2. In such case the mobile node first of all obtains (2)—with the help of IPv6 address autoconfiguration method—a unique anycast address that is valid in the whole area due to the properly set P prefix of the anycast address. As a result, the source address can be a unique anycast address since the source of a packet can be identified unequivocally. After getting the unique anycast care-of-address, the mobile node has to build the binding towards its home agent; therefore a binding procedure (3) is started by sending a binding update message. Next the mobile terminal has to initiate its membership in the virtual anycast group (VAG) of the new micromobility domain by having its anycast CoA (4). On receiving an anycast group membership report message the anycast access-router starts to propagate the new routing information by creating special routing information messages and sending it towards its adjacent routers. Based upon the underlying anycast routing protocol, each router in the new micromobility domain will get an entry in their routing table on how to reach the mobile terminal. Since each routing entry has a timeout period, thus the mobile node should send the membership report mes-
Anycast-Based Mobility
Figure 3. Moving in a given micromobility domain
sage periodically to maintain its routing entry. The updating time of the routing entry should be defined according to the refresh interval of the routing entries. In the second scenario (Figure 3), the mobile node moves in a given micromobility area (1). At the new wireless point of attachment the mobile terminal has to notify the new access router about its new location. This updating process can be done, for example, with the help of data packets of an active communication. In this case the new access router notices that packets with the anycast address in the source address field are being sent over one of its interfaces (2) (the access router checks the direction where it receives the anycastsourced packets). According to the anycast routing protocol the access router has an entry in its routing table regarding this source anycast address. Therefore the router modifies the entry regarding the anycast address of the mobile node so that the new entry forwards the packets towards their new destination (the interface from which it has received the packet with the anycast address in the source address field), the actual location of the mobile terminal. The access router also has to initiate anycast routing information exchange (3). Our approach gives a unique viewpoint on applications of IPv6 anycasting: introduces a new solution for micromobility management based on the IPv6 anycast addresses. The proposed method fits to the Mobile IPv6 standard and works efficiently in micromobility environment, while reducing the volume of control messages during the mobile operation and resulting in more seamless handover. The procedure can be
A
realized without new protocol stacks, because the method is based only on the built-in features of IPv6 standard. The anycast-based micromobility can work on any mobility-supporting IPv6 system.
FUTURE TRENDS In this article we have tried to give you an overview of the main issues that are being tackled by the ongoing research. The focus of the recent research is to construct an anycast routing protocol that is capable of handling large amounts of anycast hosts with reasonably low overhead generated by the routing system. At the moment there are multiple candidate protocols that could fulfill all the requirements set for the anycast routing protocol, while the standardization of these protocols is on the way. It is also important to take a look on the trends that can be found among the applications. One can easily see that various applications could benefit from the properties of anycast addressing, therefore more and more applications emerge for exploit these possibilities. First of all, it should be highlighted that application of the anycast addressing scheme is closely related to introduction of IPv6 into today’s network, therefore until the usage of IPv6 gets more widespread, the scope of anycasting is also limited. In accordance with the present trends the vision for the anycasting looks bright, since as soon as there will be
Anycast-Based Mobility
standardized routing protocols more and more application will be able to use the advanced services of the anycast addressing. In our view micromobility management could be one of the driving applications using anycast addresses.
Eunsoo, S. (2004). Mobility management in the wireless Internet. PhD Thesis, Columbia University.
CONCLUSION
Hinden, R., & Deering, S. (2003). .IP Version 6 Addressing Architecture. IETF RFC 3513.
Our aim in this article was to present an overview of the usage of anycast addressing paradigm and also show a possible new usage of the anycast address introduced in IPv6. The proposed anycast-based micromobility scheme is fairly simple: the mobile node after joining a foreign network obtains a unique anycast care-of-address that is valid until the mobile terminal stays inside the micromobility area, no matter if the mobile node moves around. Our method uses the services of anycast routing protocols that are capable of routing the traffic towards the “nearest” node from the set of nodes having the same anycast address. Currently none of the existing anycast routing protocols have been widely adopted, due to the lack of standardization.
REFERENCES Abley, J. (2005). Anycast addressing in IPv6. draft-jableyv6-anycast-clarify-00.txt Deering, S., & Hinden, R. (1998), Internet Protocol Version 6 (IPv6). IETF RFC 2460. Doi, S., Ata, S., Kitamura, H., & Murata M. (2004). IPv6 anycast for simple and effective service-oriented communications. IEEE Communications Magazine, 163-171. Doi, S., Ata, S., Kitamura, H., & Murata, M. (2005). Design, implementation and evaluation of routing protocols for IPv6 anycast communication. In 19th International Conference on Advanced Information Networking and Applications AINA’05 (Vol. 2, pp. 833-838). Taiwan.
Hashimoto, M., Ata, S., Kitamura, H., & Murata, M. (2005). IPv6 anycast terminolgy definition. draft-doi-ipv6-anycastfunc-term-03.txt
Johnson, D., Perkins, C., & Arkko, J. (2004). Mobility support in IPv6. IETF RFC 3775. Macker, J. P., Park, V. D., & Corson, S. M. (2001). Mobile and wireless Internet services: Putting the pieces together. IEEE Communications Magazine, 39(6), 148-155 Matsunaga, S., Ata, S., Kitamura, H., & Murata, M. (2005). Applications of IPv6 Anycasting. draft-ata-ipv6-anycastapp-01.txt. Matsunaga, S., Ata, S., Kitamura, H., & Murata, M. (2005). Design and implementation of IPv6 anycast routing protocol: PIA-SM. In 19th International Conference on Advanced Information Networking and Applications (AINA’05), (Vol. 2, pp. 839-844). Taiwan. Partridge, C., Mendez, T., & Milliken, W. (1993). Host anycasting service. IETF RFC 1546. Saha, D., Mukherjee, A., Misra, I.S., & Chakraborty, M. (2004). Mobility support in IP: A survey of related protocols. IEEE Network, 18(6), 34-40. Park, V. D., & Macker, J .P. (1999). Anycast routing for mobile networking. In MILCOM ’99 Conference Proceedings. Weber, S., & Cheng, L. (2004). A survey of anycast in IPv6 Networks. IEEE Communications Magazine, 127-133 Xuan, D., Jia, W., Zhao, W., & Zhu, H. (2000). A routing protocol for anycast messages. IEEE Transactions on Parallel and Distributed Systems, 11(6), 571-588
Category: Converging Technology
Applications Suitability on PvC Environments Andres Flores University of Comahue, Argentina Macario Polo Usaola Universidad de Castilla-La Mancha, Spain
INTRODUCTION Pervasive computing (PvC) environments should support the continuity of users’ daily tasks across dynamic changes of operative contexts. Pervasive or ubiquitous computing implies computation becoming part of the environment. Many different protocols and operating systems, as well as a variety of heterogeneous computing devices, are interrelated to allow accessing information anywhere, anytime in a secure manner (Weiser, 1991; Singh, Puradkar, & Lee, 2005; Ranganathan & Campbell, 2003). According to the initial considerations by Weiser (1991), a PvC environment should provide the feeling of an enhanced natural human environment, which makes the computers themselves vanish into the background. Such a disappearance should be fundamentally a consequence not of technology but of human psychology, since whenever people learn something sufficiently well, they cease to be aware of it. This means that the user’s relationship to computation changes to an implicit human-computer interaction. Instead of thinking in terms of doing explicit tasks “on the computer”—creating documents, sending e-mail, and so on—on PvC environments individuals may behave as they normally do: moving around, using objects, seeing and talking to each other. The environment is in charge of facilitating these actions, and individuals may come to expect certain services which allow the feeling of “continuity” on their daily tasks (Wang & Garlan, 2000). Users should be allowed to change their computational tasks between different operative contexts, and this could imply the use of many mobile devices that help moving around into the environment. As a result, the underlying resources to run the required applications may change from wide memory space, disk capacity, and computational power, to lower magnitudes. Such situations could make a required service or application inappropriate in the new context, with a likely necessity of supplying a proper adjustment. However, users should not perceive the surrounding environment as something that constraints their working/living activities. There should be a continuous provision of proper services or applications. Hence the environment must be provided with a mechanism for dynamic applications suitability (Flores & Polo, 2006).
PERvASIvE COMPUTING ENvIRONMENTS In the field of PvC there is still a misuse of some related concepts, since often PvC is used interchangeably with ubiquitous computing and mobile computing. However, nowadays consistent definitions are identified in the literature as follows (Singh et al., 2005). Mobile computing is about elevating computing services and making them available on mobile devices using the wireless infrastructure. It focuses on reducing the size of the devices so that they can be carried anywhere or by providing access to computing capacity through high-speed networks. However, there are some limitations. The computing model does not change considerably as we move, since the devices cannot seamlessly and flexibly obtain information about the context in which the computing takes place and adjust it accordingly. The only way to accommodate the needs and possibilities of changing environments is to have users manually control and configure the applications while they move—a task most users do not want to perform. PvC deals with acquiring context knowledge from the environment and dynamically building computing models dependent on context. That is, providing dynamic, proactive, and context-aware services to the user. It is invisible to human users and yet provides useful computing services (Singh et al., 2005). Three main aspects must be properly understood (Banavar & Bernstein, 2002). First is the way people view mobile computing devices and use them within their environments to perform tasks. A device is a portal into an application/data space, not a repository of custom software managed by the user. Second is the way applications are created and deployed to enable such tasks to be performed. An application is a means by which a user performs a task, not a piece of software that is written to exploit a device’s capabilities. And third is the environment and how it is enhanced by the emergence and ubiquity of new information and functionality. The computing environment is the user’s information-enhanced physical surroundings, not a virtual space that exists to store and run software. Ubiquitous computing uses the advances in mobile computing and PvC to present a global computing environment where seamless and invisible access to computing resources
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
A
Applications Suitability on PvC Environments
Figure 1. Vision of an enhanced physical environment by ubiquitous computing
is provided to the user. It aims to provide PvC environments to a human user as s/he moves from one location to another. Thus, it is created by sharing knowledge and information between PvC environments (Singh et al., 2005). Figure 1 shows the vision a user may have of a physical environment that is enhanced by ubiquitous computing. Some approaches for PvC are concerned with interconnecting protocols from different hardware artifacts and devices, or solving problems of intermittent network connections and fluctuation on bandwidth. Therefore, their applications are quite general or low level, yet mainly related to communication tools which still requires a big effort for a user to accomplish a working task. Other approaches are focused on solving problems of prohibited access to information or even to a closed or restricted environment. If we consider that the environment is populated with an enormous amount of users, each intending accesses to different hardware and software resources, the security concerns increase proportionally (Kallio, Niemelä, & Latvakoski, 2004).
On the other side, there are approaches particularly concerned with providing higher level services more related to users tasks, in order to help them reduce the working effort (Roman, Ziebart, & Campbell, 2003; Becker & Schiele, 2003; Chakraborty, Joshi, Yesha, & Finin, 2006; Gaia Project, 2006; Aura Project, 2006). Most of them have been conceptualized with some sort of self-adjusted applications or by applications relying on basic services provided by the underlying platform (e.g., CORBA). No matter how users need a transparent delivery of functionality, so they could have a sense of continued presence of the environment. Therefore, any unavailability of a required service implies that a user understand that the underlying environment cannot provide all that is needed, thus destroying the aspiration of transparency.
SUITABILITY FOR PERvASIvE APPLICATIONS Functionality on a PvC environment is usually shaped as a set of aggregated components that are distributed among different computing devices. On changes of availability of a given device, the involved component behavior still needs to be accessible in the appropriate form according to the updated technical situation. This generally makes users be involved on a dependency with the underlying environment and increases the complexity of its internal mechanisms (Iribarne, Troya, & Vallecillo, 2003; Warboys et al., 2005). Applications composed of dynamically replaceable components imply the need of an appropriate integration process according to component-based software development (CBSD) (Cechich, Piattini, & Vallecillo, 2003; Flores, Augusto, Polo, & Varea, 2004). For this, an application model may provide the specification of a required functionality in the form of the aggregation of component models, as can be seen in Figure 2. A component model provides a definition to
Figure 2. Connection of models and components to integrate an application
Application Model
Required Component Model 1
Component Model 1.1
Component 1.1
Required Component Model 2
Component Model 2.1
Component 2.1
Required Component Model M
Component Model M.1
Component M.1
Applications Suitability on PvC Environments
instantiate a component and its composition aspects through standard interactions and unambiguous interfaces (Cechich et al., 2003; Iribarne et al., 2003; Warboys et al., 2005). In order to assure the adequacy of a given component with respect to an application model, there is a need to evaluate its component model. Hence we present an assessment procedure which can be applied both on a development stage and also at runtime. The latter becomes necessary when the current technical situation makes unsuitable a given component demanding that a surrogate be provided. The assessment procedure compares functional aspects from components against the specification provided by the application model, which is component oriented. Besides analyzing component services at a syntactic level, its behavior is also inspected, thus embracing semantic aspects. The latter is done by abstracting out the black box functionality hidden on components in the form of assertions, and also exposing its likely interactions by means of the protocol of use, which describes the expected order of use for its services (Flores & Polo, 2005)—also called choreography (Iribarne et al., 2003). So far we have been experimenting with the addition of metadata for comparing behavioral aspects from components. Metadata has been used in several approaches as a technique to easy verification procedures (Cechich et al., 2003; Cechich & Polo, 2005; Orso et al., 2001). By adding meta-methods we may then retrieve detailed information concerning assertions and the protocol of use, which somehow implies a component adaptation, particularly referred to as the instrumentation mechanism (Flores & Polo, 2005). The assessment procedure is described by means of a set of conditions which must be satisfied according to certain thresholds. Different techniques are applied to achieve the required evaluations. Compatibility on both assertions and the usage protocol is carried out by generating Abstract Syntax Trees and applying some updated algorithms, which were originally developed to detect similar pieces of code
(clones) on existing programs. Such compatibility analysis is based on the following consideration. Post-conditions (for example) on services from two similar components necessarily should relate to a similar structure and semantic. Hence, they could be thought of as one being a clone of the other (Flores & Polo, 2005). All such techniques applied on our assessment procedure allow the accomplishment of a consistent mechanism to assure a fair component integration. As PvC environments imply many challenges, the whole integration process is based on considering all aspects concerning reliability. We may make use of a simple example to illustrate the way a functionality is composed from distributed disparate components and understand how the assessment procedure may help to assure the suitability of a certain involved component. Suppose we represent a PvC environment for a museum, which includes a tour guide application for proposing different paths according to the user’s dynamic choices. When the user enters the museum, s/he may carry a computing device (a PDA or a smart phone), and an automatic detection is done in order to identify and connect the device to the environment. As the user walks by each art piece (e.g., painting or sculpture), descriptions and information of particular interest to the user are displayed on the PDA or spoken through the phone. Figure 3 shows a likely scenario of the presented case study. A related application could allow creating an album with images of some art pieces the user has visited. To obtain the images the user will probably have to pay a fee, for example, when s/he intends to leave the museum. The album organizer application—maybe downloaded into a user’s notebook recognized by the environment—may allow creating a sort of document with images and some notes written by the user. Notes could be stored on separated text files and bind to the document by means of hypertext links. Thus every time the user needs to write or edit a note, a proper editor is provided.
Figure 3. PvC environment for a museum
A
Applications Suitability on PvC Environments
Figure 4. Distributed components for the album organizer application
The user may also be allowed to print a selection of pages of the document, or even send the created album by e-mail in case s/he does not have a device to carry those files. If we focus on the album organizer application, we might analyze the potential required components. There could be an album organizer component to represent the main logic of the application. This component could have an ad-hoc sophisticated album visual editor or it could be a Web-style editor in which is additionally required a generic Web browser. The visual editor also depends on the actual used device. In order to make notes, different components could be used as a simple sort of Notepad, or other replacements like WordPad or similar applications according to the underlying software platform. For sending e-mail, applications like Outlook or Eudora could be used, and to provide a printer service, different kinds of printers and ad hoc wireless sensors should be available. Another component is concerned with the database for images and descriptions of art pieces. Figure 4 shows a diagram with the likely comprised components and devices for the album organizer application. Suppose a user needs to write a note by using a notebook which runs a Linux platform. One available text editor is KEdit. The environment then evaluates this component so to ensure it is appropriate to fulfill the task. Following can be seen the interfaces of both the KEdit component and the required component model named TextEditor. In order to assure the compatibility with the surrogate, the assessment procedure is applied to analyze if the degree of similarity raises a proper level. Since implementation alternatives are fairly important, we have properly explored some of them. For example both Microsoft .NET and Enterprise Java Beans allow the addition of metadata to components. Information about components—their structure and the state of their corresponding instances—can be retrieved at runtime by using reflection on both environments as well. Particularly .NET includes the possibility of adding attributes, which is a special class
0
intended to provide additional information about some design element as a class, a module, a method, a field, and so on (Flores & Polo, 2005). Hence, we have implemented on the .Net technology the current state of our approach (Flores & Polo, 2006). Though simple, this prototype gives us rewarding data on possibilities to make our proposals concrete. All the applied techniques are selected according to our goal of achieving consistent mechanisms to assure a fair component integration. As we proceed with our work, reliability is mainly considered, since we focus the whole integration process for those challenging systems as PvC environments.
FUTURE TRENDS Our work continues by completing the coverage of functional aspects for components, describing the components replacement mechanism and then focusing on the non-functional aspects, particularly on quality of service. For this we are analyzing the use and extending the schemas from Iribarne et al. (2003) which provide a consistent format for specifications of components by means of XML. The approach covers all of the aspects from components: functional, non-functional, and commercial. As some authors have pointed out (Chen, Finin, & Joshi, 2005; Ranganathan & Campbell, 2003), temporal aspects could also be helpful to achieve a more accurate component integration process. This may give the chance to analyze whether a component can fit the requirements when time conditions are also included in the set of updated requirements upon a change on the user’s context of operation. Selection of appropriate methods, techniques, and languages must be accurately accomplished upon the concern of a reliable mechanism. This is the emphasis of our next development in this area.
Applications Suitability on PvC Environments
REFERENCES Aura Project. (2006). Aura Project Web site. Retrieved from http://www-2.cs.cmu.edu/ aura/ Banavar, G., & Bernstein, A. (2002). Issues and challenges in ubiquitous computing: Software infrastructure and design challenges for ubiquitous computing applications. Communications of the ACM, 45(12). Becker, C., & Schiele, G. (2003). Middleware and application adaptation requirements and their support in pervasive computing. Proceedings of IEEE ICDCSW (pp. 98-103), Providence, RI. Cechich, A., & Polo, M. (2005). COTS component testing through aspect-based metadata. In S. Beydeda & V. Gruhn (Eds.), Building quality into components—testing and debugging. Berlin: Springer-Verlag.
Orso, A., Harrold, M.J., Rosenblum, D., Rothermel, G., Do, H., & Sofía, M.L. (2001). Using component metacontent to support the regression testing of component-based software. Proceedings of IEEE ICSM, Florence, Italy, (pp. 716-725). Ranganathan, A., & Campbell, R. (2003). An infrastructure for context-awareness based on first order logic. Personal and Ubiquitous Computing, 7, 353-364. Roman, M., Ziebart, B., & Campbell, R. (2003). Dynamic application composition: Customizing the behavior of an active space. Proceedings of IEEE PerCom. Singh, S., Puradkar, S., & Lee, Y. (2005). Ubiquitous computing: Connecting pervasive computing through semantic Web. Journal of ISeB. Berlin: Springer-Verlag.
Cechich, A., Piattini, M., & Vallecillo, A. (Eds.). (2003). Component-based software quality: Methods and techniques (LNCS 2693). Berlin: Springer-Verlag.
Wang, Z., & Garlan, D. (2000). Task-driven computing. Technical Report No. CMU-CS-00-154, School of Computer Science, Carnegie Mellon University, USA. Retrieved from http://reports-archive.adm.cs.cmu.edu/anon/2000/abstracts/00-154.html
Chakraborty, D., Joshi, A., Yesha, Y., & Finin, T. (2006). Toward distributed service discovery pervasive computing environments. IEEE Transactions on Mobile Computing, 5(2).
Warboys, B., Snowdon, B., Greenwood, R. M., Seet, W., Robertson, I., Morrison, R., Balasubramaniam, D., Kirby, G., & Mickan, K. (2005). An active-architecture approach to COTS integration. IEEE Software, 22(4), 20-27.
Chen, H., Finin, T., & Joshi, A. (2004). Semantic Web in the context broker architecture. Proceedings of IEEE PerCom, Orlando, FL, (pp. 277-286).
Weiser, M. (1991). The computer for the 21st century. Scientific American, 265(3), 94-104. Retrieved from http://www. ubiq.com/hypertext/weiser/SciAmDraft3.html
Flores, A., & Polo, M. (2005, June 27-30). Dynamic component assessment on PvC environments. Proceedings of IEEE ISCC, Cartagena, Spain, (pp. 955-960).
KEY TERMS
Flores, A., & Polo, M. (2006, May 23). An approach for applications suitability on pervasive environments. Proceedings of IWUC (held at ICEIS). Paphos, Cyprus: INSTICC Press. Flores, A., Augusto, J. C., Polo, M., & Varea, M. (2004, October 10-13). Towards context-aware testing for semantic interoperability on PvC environments. Proceedings of IEEE SMC, The Hague, The Netherlands, (pp. 1136-1141). Gaia Project. (2006). Gaia Project Web site. Retrieved from http://www.w3.org/2001/sw Iribarne, L., Troya, J., & Vallecillo, A. (2003). A trading service for COTS components. The Computer Journal, 47(3). Kallio, P., Niemelä, E., & Latvakoski, J. (2004). UbiSoft—pervasive software. Research Notes, 2238. Finland: VTT Electronics. Retrieved from www.vtt.fi/inf/pdf/tiedotteet/2004/T2238.pdf
Component-Based Software Development (CBSD): A development paradigm where software systems are developed from the assembly or integration of software components. Component Model: A specification that describes how to instantiate or build a software component, and gives guidelines for its binding to other software components by means of standard interactions or communication patterns and unambiguous interfaces. Mobile Computing: Small wireless devices that can be carried anywhere, allowing a computing capacity through wireless networks. Also known as nomadic computing. Pervasive Computing (PvC): Enhancement of the physical surroundings by providing and adapting mobile computing according to the user’s needs. Also known as ambient intelligent. Software Component: A unit of independent deployment that is ready “off-the-shelf” (OTS), from a commercial source (COTS) or reused from another system (in-house or legacy
A
Applications Suitability on PvC Environments
systems). It is usually self-contained, enclosing a collection of cooperating and tightly cohesive objects, thus providing a significant aggregate of functionality. It is used “as it is found” rather than being modified, may possibly execute independently, and can be integrated with other components to achieve a required bigger system functionality.
and provides voice, data, games, and video applications. The most familiar is the mobile or cell phone, then Palm Pilot and its handheld descendent, the PDA (personal digital assistant), a great evolution because of its large amount of new applications. Laptop or notebook and tablet PCs are also well-known wireless devices.
Ubiquitous Computing: Provides PvC environments to a human user as s/he moves from one location to another. It allows sharing knowledge and information between PvC environments. Also known as Global Computing.
Wireless Network: Offers mobility and elimination of unsightly cables, by the use of radio waves and/or microwaves to maintain communication channels between computers. It is an alternative to wired networking, which relies on copper and/or fiber optic cabling between network devices. Popular wireless local area networking (WLAN) products conform to the 802.11 “Wi-Fi” standards.
Wireless Device: A small computer that is reduced in size and in computing power, can be carried everywhere,
Category: Converging Technology
A Bio-Inspired Approach for the Next Generation of Cellular Systems Mostafa El-Said Grand Valley State University, USA
INTRODUCTION In the current 3G systems and the upcoming 4G wireless systems, missing neighbor pilot refers to the condition of receiving a high-level pilot signal from a Base Station (BS) that is not listed in the mobile receiver’s neighbor list (LCC International, 2004; Agilent Technologies, 2005). This pilot signal interferes with the existing ongoing call, causing the call to be possibly dropped and increasing the handoff call dropping probability. Figure 1 describes the missing pilot scenario where BS1 provides the highest pilot signal compared to BS1 and BS2’s signals. Unfortunately, this pilot is not listed in the mobile user’s active list. The horizontal and vertical handoff algorithms are based on continuous measurements made by the user equipment (UE) on the Primary Scrambling Code of the Common Pilot Channel (CPICH). In 3G systems, UE attempts to measure the quality of all received CPICH pilots using the Ec/Io and picks a dominant one from a cellular system (Chiung & Wu, 2001; El-Said, Kumar, & Elmaghraby, 2003). The UE interacts with any of the available radio access networks based on its memorization to the neighboring BSs. As the UE moves throughout the network, the serving BS must constantly update it with neighbor lists, which tell the UE which CPICH pilots it should be measuring for handoff purposes. In 4G systems, CPICH pilots would be generated from any wire-
less system including the 3G systems (Bhashyam, Sayeed, & Aazhang, 2000). Due to the complex heterogeneity of the 4G radio access network environment, the UE is expected to suffer from various carrier interoperability problems. Among these problems, the missing neighbor pilot is considered to be the most dangerous one that faces the 4G industry. The wireless industry responded to this problem by using an inefficient traditional solution relying on using antenna downtilt such as given in Figure 2. This solution requires shifting the antenna’s radiation pattern using a mechanical adjustment, which is very expensive for the cellular carrier. In addition, this solution is permanent and is not adaptive to the cellular network status (Agilent Technologies, 2005; Metawave, 2005). Therefore, a self-managing solution approach is necessary to solve this critical problem. Whisnant, Kalbarczyk, and Iyer (2003) introduced a system model for dynamically reconfiguring application software. Their model relies on considering the application’s static structure and run-time behaviors to construct a workable version of reconfiguration software application. Self-managing applications are hard to test and validate because they increase systems complexity (Clancy, 2002). The ability to reconfigure a software application requires the ability to deploy a dynamically hardware infrastructure in systems in general and in cellular systems in particular (Jann, Browning, & Burugula, 2003).
Figure 1. Missing pilot scenario
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
B
A Bio-Inspired Approach for the Next Generation of Cellular Systems
Figure 2. Missing pilot solution: Antenna downtilt
Missing Pilot Tower (MP)
BS3 BS1
BS2
Konstantinou, Florissi, and Yemini (2002) presented an architecture called NESTOR to replace the current network management systems with another automated and softwarecontrolled approach. The proposed system is inherently a rulebased management system that controls change propagation across model objects. Vincent and May (2005) presented a decentralized service discovery approach in mobile ad hoc networks. The proposed mechanism relies on distributing information about available services to the network neighborhood nodes using the analogy of an electrostatic field. Service requests are issued by any neighbor node and routed to the neighbor with the highest potential. The autonomic computing system is a concept focused on adaptation to different situations caused by multiple systems or devices. The IBM Corporation recently initiated a public trail of its Autonomic Toolkit, which consists of multiple tools that can be used to create the framework of an autonomic management system. In this article, an autonomic engine system setting at the cellular base station nodes is developed to detect the missing neighbor (Ganek & Corbi, 2003; Haas, Droz, & Stiller, 2003; Melcher & Mitchell, 2004). The autonomic engine receives continuous feedback and performs adjustments to the cell system’s neighboring set by requiring the UE to provide signal measurements to the serving BS tower (Long, 2001). In this article, I decided to use this toolkit to build an autonomic rule-based solution to detect the existence of any missing pilot. The major advantage of using the IBM autonomic toolkit is providing a common system infrastructure for processing and classifying the RF data from multiple sources regardless of its original sources. This is a significant step towards creating a transparent autonomic high-speed physical layer in 4G systems.
PROPOSED SOLUTION The proposed AMS relies on designing an autonomic highspeed physical layer in the smart UE and the BS node. At the UE side, continuous CPICH pilot measurements will be recorded and forwarded to the serving BS node via its radio interface. At the BS node, a scalable self-managing autonomic engine is developed using IBM’s autonomic computing toolkit to facilitate the mobile handset’s vertical/horizontal handover such as shown in Figure 3. The proposed engine is cable of interfacing the UE handset with different wireless technologies and detects the missing pilot if it is existed. The autonomic engine relies on a generic log adapter (GLA), which is used to handle any raw measurements log file data and covert it into a standard format that can be understood by the autonomic manager. Without GLA, separate log adapters would have be coded for any system that the autonomic manager interfaced with. The BS node will then lump all of the raw data logs together and forward them to the Generic Log Adapter for data classification and restructuring to the common base event format. Once the GLA has parsed a record in real time to common base event format, the autonomic manager will see the record and process it and take any action necessary by notifying the BS node to make adjustments to avoid the missing pilot and enhance the UE devices’ quality of service.
PERFORMANCE MEASUREMENTS AND KEY FINDINGS To test the applicability of the proposed solution, we decided to use the system’s response time, AS’s service rate for callers experiencing missing pilot problem, and the performance
A Bio-Inspired Approach for the Next Generation of Cellular Systems
Figure 3. Autonomic base station architecture
B
RF Pilot Signals Measurements
Application Plane
Generic Log Adapter (GLA)
Autonomic Plane
Common Base Event Format (CBE) Wireless Autonomic Manger (WAM) MP Detection
MP Correction
Table 1. Summary of the system performance analysis Log File Size in (# Records)
System Response Time in (Sec)
Processing Rate by the Base Station in (Records/Sec)
Trial Experiment 1
985
145
6.793103448
Trial Experiment 2
338
95
3.557894737
Trial Experiment 3
281
67
4.194029851
Trial Experiment 4
149
33
4.515151515
Average Processing Rate by the Base Station in (Records/Sec)
4.765044888
Base Station Service Rate For callers experiencing missing pilot problem (Records/Sec)
5.3
Performance Gain
1.112266542
gain as performance metrics. Also, we developed a Java class to simulate the output of a UE in a heterogeneous RF access network. Table 1 summarizes the simulation results for four simulation experiments with different log files size. The results shown in Table 1 comply with the design requirements for the current 3G system. This is illustrated in the following simple example.
DESIGN REQUIREMENTS FOR 3G SYSTEMS •
•
The 3G cell tower’s coverage area is divided into three sectors, with each sector having (8 traffic channel * 40 call/channel = 320 voice traffic per sector) and (2 control channels * 40 callers/channels = 80 control traffic per sector). The overlapped area between towers (handoff zone) occupies 1/3 of the sector size and serves (1/3 of 320
A Bio-Inspired Approach for the Next Generation of Cellular Systems
•
= 106 callers (new callers and/or exciting ones)). If we consider having the UE report its status to the tower every 5 seconds, we could potentially generate 21.2 records in 1 second. It is practical to assume that 25% of the 21.2 reports/ second accounts for those callers that may suffer from the missing pilot problemthat is, the tower’s service rate for missing neighbor pilot callers is 21.2/4 = 5.3 records/second. This is the threshold level used by the tower to accommodate those callers suffering from the missing pilot problem.
CONCLUSION In this article, we have developed an autonomic engine system setting at the cellular base station (BS) nodes to detect the missing neighbor. The autonomic engine receives continuous feedback and performs adjustments to the cell system’s neighboring set by requiring the user equipment (UE) to provide signal measurements to the serving BS tower. The obtained results show that the proposed solution is able to detect the missing pilot problem in any heterogeneous RF environment.
ANALYSIS OF THE RESULTS
REFERENCES
•
Agilent Technologies. (2005). Retrieved October 2, 2005, from http://we.home.agilent.com
•
• • •
Response time is the time taken by the BS to process, parse the incoming log file and detect the missing neighbor pilot. It is equal to (145, 95, 67, and 33) for the four experiment trials. All values are in seconds. Processing rate by the base station is defined as the total number of incoming records divided by the response time in (records/second). It is equal to (6.7, 3.5, 4.1, and 4.5) for the four experiment trials. The UE reports a missing pilot problem with an average rate of 4.7 records/second. The base station’s service rate for callers experiencing the missing pilot problem = 5.3 records/second. The performance gain is defined as:
Base Station Service Rate For callers experienci ng missing pilot problem in (Records/S ec) Average Processing Rate by the Base Station in (Records/S ec)
= 5.3/4.7=1.1 •
Here it is obvious that the service rate (5.3 records/ second) is greater than the UE’s reporting rate to the base station node (4.7 records/second). Therefore, the above results prove that the proposed solution does not overload the processing capabilities of the BS nodes and can be scaled up to handle a large volume of data.
FUTURE TRENDS An effective solution for the interoperability issues in 4G wireless systems must rely on an adaptive and self-managing network infrastructure. Therefore, the proposed approach in this article can be scaled to maintain continuous user connectivity, better quality of service, improved robustness, and higher cost-effectiveness for network deployment.
Bhashyam, S., Sayeed, A., & Aazhang, B. (2000). Timeselective signaling and reception for communication over multipath fading channels. IEEE Transaction on Communications, 48(1), 83-94. Chiung, J., & Wu, S. (2001). Intelligent handoff for mobile wireless Internet. Journal of Mobile Networks and Applications, 6, 67-79. Clancy, D. (2002). NASA challenges in autonomic computing. Almaden Institute 2002, IBM Almaden Research Center, San Jose, CA. El-Said, M., Kumar, A., & Elmaghraby, A. (2003). Pilot pollution interference cancellation in CDMA systems. Special Issue of Wiley Journal: Wireless Communication and Mobile Computing on Ultra Broadband Wireless Communications for the Future, 3(6), 743-757. Ganek, A., & Corbi, T. (2003). The dawning of the autonomic computing era. IBM Systems Journal, l42(1), 5-19. Haas, R., Droz, P., & Stiller, B. (2003). Autonomic service deployment in networks. IBM Systems Journal, 42(1), 150-164. Jann, L., Browning, A., & Burugula, R. (2003). Dynamic reconfiguration: Basic building blocks for autonomic computing on IBM pSeries servers. IBM Systems Journal, 42(1), 29-37. Konstantinou, A., Florissi, D., & Yemini, Y. (2002). Towards self-configuring networks. Proceedings of the DARPA Active Networks Conference and Exposition (pp. 143-156). LCC International. (2004). Retrieved December 10, 2004, from http://www.hitech-news.com/30112001-MoeLLC. htm
A Bio-Inspired Approach for the Next Generation of Cellular Systems
Lenders, V., May, M., & Plattner, B. (2005). Service discovery in mobile ad hoc networks: A field theoretic approach. Special Issue of Pervasive and Mobile Computing, 1, 343-370.
Candidate Set: Depicts those base stations that are in transition into or out of the active set, depending on their power level compared to the threshold level.
Long, C. (2001). IP network design. New York: McGrawHill Osborne Media.
Missing Neighbor Pilot: The condition of receiving a high-level pilot signal from a base station (BS) that is not listed in the mobile receiver’s neighbor list.
Melcher, B., & Mitchell, B. (2004). Towards an autonomic framework: Self-configuring network services and developing autonomic applications. Intel Technology Journal, 8(4), 279-290. Metawave. (2005). Retrieved November 10, 2005, from http://www.metawave.com Whisnant, Z., Kalbarczyk, T., & Iyer, R. (2003). A system model for dynamically reconfigurable software. IBM Systems Journal, 42(1), 45-59.
KEY TERMS
Neighbor Set: Represents the nearby serving base stations to a mobile receiver. The mobile receiver downloads an updated neighbor list from the current serving base station. Each base station or base station sector has a unique neighbor list. Policy-Based Management: A method of managing system behavior or resources by setting “policies” (often in the form of “if-then” rules) that the system interprets. Virtual Active Set: Includes those base stations (BSs) that are engaged in a live communication link with the mobile user; they generally do not exceed three base stations at a time.
Adaptive Algorithm: Can “learn” and change its behavior by comparing the results of its actions with the goals that it is designed to achieve. Autonomic Computing: An approach to self-managed computing systems with a minimum of human interference. The term derives from the body’s autonomic nervous system, which controls key functions without conscious awareness or involvement.
B
Category: Mobile Software Engineering
Brain Computer Interfacing Diego Liberati Italian National Research Council, Italy
INTRODUCTION In the near future, mobile computing will benefit from more direct interfacing between a computer and its human operator, aiming at easing the control while keeping the human more free for other tasks related to displacement. Among the technologies enabling such improvement, a special place will be held by brain computer interfacing (BCI), recently listed among the 10 emerging technologies that will change the world by the MIT Technology Review on January 19, 2004. The intention to perform a task may be in fact directly detected from analyzing brain waves: an example of such capability has been for instance already shown trough artificial neural networks in Babiloni et al. (2000), thus allowing the switch of a bit of information in order to start building the control of a direct interaction with the computer.
BACKGROUND Our interaction with the world is mediated through sensory-motor systems, allowing us both to acquire information from our surroundings and manipulate what is useful at our reach. Human-computer interaction ergonomically takes into account the psycho-physiological properties of such interaction to make our interactions with computers increasingly easy. Computers are in fact nowadays smaller and smaller without significant loss of power needed for everyday use, like writing an article like this one on a train going to a meeting, while checking e-mail and talking (via voice) to colleagues and friends. Now, the center of processing outside information and producing intention to act consequently is well known to be our brain. The capability to directly wire neurons on electronic circuits is not (yet?) within our reach, while interesting experiments of compatibility and communication capabilities are indeed promising at least in vitro. At the other extreme, it is not hard to measure non-invasively the electromagnetic field produced by brain function by positioning small electrodes over the skull, as in the standard clinical procedure of electroencephalography. Obviously, taking from outside a far-field outside measure is quite different than directly measuring the firing of every single motor neuron of interest: a sort of summing of all the
brain activity will be captured at different percentages. Nonetheless, it is well known that among such a messy amount of signal, when repeating a task it is not hard to enhance the very signal related to task, while reducing—via synchronized averaging—the overwhelming contribution of all the other neurons not related to the task of interest. On this premise, Deecke, Grozinger, and Kornhuber (1976) have been able to study the so-called event-related brain potential, naming the onset of a neural activation preceding the task, in addition to the neural responses to the task itself. Statistical pattern recognition and classification has been shown to improve such event-related detection by Gevins, Morgan, Bressler, Doyle, and Cutillo (1986). A method to detect such preparatory potential on a single event basis (and then not needing to average hundreds of repetitions, as said before) was developed and applied some 20 years ago by Cerutti, Chiarenza, Liberati, Mascellani, and Pavesi (1988). One extension of the same parametric identification approach is that developed by del Millan et al. (2001) at the European Union Joint Research Center of Ispra, Italy, while a Bayesian inference approach has been complementary proposed by Roberts and Penny (2000).
BRAIN COMPUTER INTERFACING Autoregressive with exogenous input parametric identification (Cerutti et al., 1988) is able to increase by some 20 dB the worst signal-to-noise ratio of the event-related potential with respect to the overwhelming background brain activity. Moreover, it provides a reduced set of parameters that can be used as features to perform post-processing, should it be needed. A more sophisticated, though more computing-demanding, time variant approach based on an optimal so-called Kalman filer has been developed by Liberati, Bertolini, and Colombo (1991a). The joint performance of more than one task has also been shown to evoke more specific brain potential (Liberati, Bedarida, Brandazza, & Cerutti, 1991b). Multivariable joint analysis of covariance (Gevins et al., 1989), as well as of total and partial coherence among brain field recordings at different locations (Liberati, Cursi, Locatelli, Comi, & Cerutti, 1997), has also improved the capability of discriminating the single potential related to a particular task (Liberati, 1991a).
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Brain Computer Interfacing
Artificial neural networks (Liberati, 1991b; Pfurtscheller, Flotzinger, Mohl, & Peltoranta, 1992; Babiloni et al., 2000) offered the first approach to the problem of so-called artificial intelligence, whose other methods of either soft computing, like the fuzzy sets made popular by Lofthi Zadeh, and the rule extraction like the one proposed by Muselli and Liberati (2002), are keen to be important post-processing tools for extracting real commands from the identified parameters. In particular, the rule extraction approach proposed by Muselli and Liberati (2000) has the nice properties of processing the huge amount of data in a very fast quadratic time (and even in terms of binary operations), yielding both pruning of the redundant variables for discrimination (like not necessary recording points or time windows) and an understandable set of rules relating the residual variable of interest: this is thus quite useful in learning the BCI approach. When the space of the feature is then confined to the few really salient, the Piece-Wise Affine identification recently developed (Ferrari-Trecate, Muselli, Liberati, & Morari, 2003) and also applied in a similar context for instance to hormone pulse or sleep apnea detection (Ferrari-Trecate & Liberati, 2002), is keen to be a good tool to help refine the detection of such mental decision on the basis of the multivariate parametric identification of a multiple set of dynamic biometric signals. Here the idea is to cluster recorded data or features obtained by the described preprocessing in such a way to identify automatically both approximate linear relations among them in each region of interest and the boundaries of such regions themselves, thus allowing quite precise identification of the time of the searched switching.
FUTURE TRENDS Integration of more easily recorded signals is even more promising for automatic processing of the intention of interacting with the computer, both in a context of more assisted performance in everyday life, as well as in helping to vicariate lost functions because of handicaps.
CONCLUSION The task is challenging, though at a first glance it would even appear not so complex: it wants to discriminate at least a bit of information (like opening and closing a switch), and then sequentially, it would be possible to compose a word of any length. The point is that every single bit of intention should be identified with the highest accuracy, in order to avoid too many redundancies, demanding time while offering safety.
REFERENCES Babiloni, F., Carducci, F., Cerutti, S., Liberati, D., Rossini, P., Urbano, A., & Babiloni, C. (2000). Comparison between human and ANN detection of Laplacian-derived electroencephalographic activity related to unilateral voluntary movements. Comput Biomed Res, 33, 59-74. Cerutti, S., Chiarenza, G., Liberati, D., Mascellani, P., & Pavesi, G. (1988). A parametric method of identification of the single trial event-related potentials in the brain. IEEE Transactions of Biomedical Engineering, 35(9), 701. Deecke, L., Grözinger, B., & Kornhuber, H. (1976). Voluntary finger movements in man: Cerebral potentials and theory. Biological Cybernetics, 23, 99-119. del Millan et al. (2001). Brain computer interfacing. In D. Liberati (Ed.), Biosys: Information and control technology in health and medical systems. Milan: ANIPLA. Ferrari-Trecate, G., & Liberati, D. (2002). Representing logic and dynamics: The role of piecewise affine models in the biomedical field. Proceedings of the EMSTB Math Modeling and Computing in Biology and Medicine Conference, Milan. Ferrari–Trecate, G., Muselli, M., Liberati, D., & Morari, M. (2003). A clustering technique for the identification of piecewise affine systems. Automatica, 39, 205-217. Gevins, A., Morgan, N., Bressler, S., Doyle, J., & Cutillo, B. (1986). Improved event-related potential estimation using statistical pattern classification. Electroenceph. Clin. Neurophysiol, 64, 177. Gevins, A., Bressler, S. L., Morgan, N. H., Cutillo, B., White, R. M., Greer, D. S., & Illes, J. (1989). Event-related covariances during a bimanual visuomotor task: Methods and analysis of stimulus and response-locked data. Electroenceph. Clin. Neurophysiol, 74, 58. Liberati, D. (1991a), Total and partial coherence analysis of evoked brain potentials. Proceedings of the 4th International Symposium on Biomedical Engineering, Peniscola, Spain, (pp. 101-102). Liberati, D. (1991b). A neural network for single sweep brain evoked potential detection and recognition. Proceedings of the 4th International Symposium on Biomedical Engineering, Peniscola, Spain, (pp. 143-144). Liberati, D., Bertolini, L., & Colombo, D. C. (1991a). Parametric method for the detection of inter and intra-sweep variability in VEP’s processing. Med Biol Eng Comput, 29, 159-166.
B
Brain Computer Interfacing
Liberati, D., Bedarida, L., Brandazza, P., & Cerutti, S. (1991b). A model for the cortico-cortical neural interaction in multisensory evoked potentials. IEEE T Bio-Med Eng, 38(9), 879-890. Liberati, D., Cursi, M., Locatelli, T., Comi, G., & Cerutti, S. (1997). Total and partial coherence of spontaneous and evoked EEG by means of multi-variable autoregressive processing. Med Biol Eng Comput, 35(2), 124-130. Muselli, M., & Liberati, D. (2000). Training digital circuits with hamming clustering. IEEE T Circuits I, 47, 513-527.
Bayesian Statistics: Named after its developer, Bayes, it takes into account conditional probabilities in order to describe variable relationships. Brain Computer Interfacing: Ability, even if only partial, of outside controlling by detecting intention through brain wave analysis. Event-Related Potential: Electro-magnetic brain activity related to a specific event: it may be evoked from the outside, or self-ongoing.
Muselli, M., & Liberati, D. (2002). Binary rule generation via hamming clustering. IEEE T Knowl Data En, 14(6), 1258-1268.
Fuzzy Logic: Set theory mainly developed and made popular by Lofthi Zadeh at the University of California at Berkeley where belonging of elements is not crisply attributed to only one disjoint subset.
Pfurtscheller, G., Flotzinger, D., Mohl, W., & Peltoranta, M. (1992). Prediction of the side of hand movements from single-trial multi-channel EEG data using neural networks. Electroenceph. Clin Neurophysiol, 82, 313.
Parametric Identification: Black box mathematical modeling of an input-output relationship via simple, even linear equations depending on a few parameters, whose values do identify the system dynamics.
Roberts, S. J., & Penny, W. D. (2000). Real-time braincomputing interfacing: A preliminary study using Bayesian learning. Medical & Biological Engineering & Computing, 38(1), 56-61.
Piecewise Affine Identification: Linearization of a nonlinear function, automatically partitioning data in subsets whose switching identifies state commuting in a hybrid dynamic-logical process.
KEY TERMS
Rule Induction: Inference from data of “if…then…else” rules describing the logical relationships among data.
Artificial Neural Networks: Non-linear black box inputoutput relationships built on a regular structure of simple elements, loosely inspired to the natural neural system, even in learning by example.
0
Category: Service Computing
Bridging Together Mobile and Service-Oriented B Computing Loreno Oliveira Federal University of Campina Grande, Brazil Emerson Loureiro Federal University of Campina Grande, Brazil Hyggo Almeida Federal University of Campina Grande, Brazil Angelo Perkusich Federal University of Campina Grande, Brazil
INTRODUCTION The growing popularity of powerful mobile devices, such as modern cellular phones, smart phones, and PDAs, is enabling pervasive computing (Weiser, 1991) as the new paradigm for creating and interacting with computational systems. Pervasive computing is characterized by the interaction of mobile devices with embedded devices dispersed across smart spaces, and with other mobile devices on behalf of users. The interaction between user devices and smart spaces occurs primarily through services advertised on those environments. For instance, airports may offer a notification service, where the system registers the user flight at the check-in and keeps the user informed, for example, by means of messages, about flight schedule or any other relevant information. In the context of smart spaces, service-oriented computing (Papazoglou & Georgakopoulos, 2003), in short SOC, stands out as the effective choice for advertising services to mobile devices (Zhu, Mutka, & Ni, 2005; Bellur & Narendra, 2005). SOC is a computing paradigm that has in services the essential elements for building applications. SOC is designed and deployed through service-oriented architectures (SOAs) and their applications. SOAs address the flexibility for dynamic binding of services, which applications need to locate and execute a given operation in a pervasive computing environment. This feature is especially important due to the dynamics of smart spaces, where resources may exist anywhere and applications running on mobile clients must be able to find out and use them at runtime. In this article, we discuss several issues on bridging mobile devices and service-oriented computing in the context of smart spaces. Since smart spaces make extensive use of services for interacting with personal mobile devices, they become the ideal scenario for discussing the issues for this integration. A brief introduction on SOC and SOA is also
presented, as well as the main architectural approaches for creating SOC environments aimed at the use of resourceconstrained mobile devices.
BACKGROUND SOC is a distributed computing paradigm whose building blocks are distributed services. Services are self-contained software modules performing only pre-defined sets of tasks. SOC is implemented through the deployment of any software infrastructure that obeys its key features. Such features include loose coupling, implementation neutrality, and granularity, among others (Huhns & Singh, 2005). In this context, SOAs are software architectures complying with SOC features. According to the basic model of SOA, service providers advertise service interfaces. Through such interfaces, providers hide from service clients the complexity behind using different and complex kinds of resources, such as databanks, specialized hardware (e.g., sensor networks), or even combinations of other services. Service providers announce their services in service registries. Clients can then query these registries about needed services. If the registry knows some provider of the required service, a reference for that provider is returned to the client, which uses this reference for contacting the service provider. Therefore, services must be described and published using some machine-understandable notation. Different technologies may be used for conceiving SOAs such as grid services, Web services, and Jini, which follow the SOC concepts. Each SOA technology defines its own standard machineries for (1) service description, (2) message format, (3) message exchange protocol, and (4) service location.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Bridging Together Mobile and Service-Oriented Computing
In the context of pervasive computing, services are the essential elements of smart spaces. Services are used for interacting with mobile devices and therefore delivering personalized services for people. Owning to the great benefits that arise with the SOC paradigm, such as interoperability, dynamic service discovery, and reusability, there is a strong and increasing interest in making mobile devices capable of providing and consuming services over wireless networks (Chen, Zhang, & Zhou, 2005; Kalasapur, Kumar, & Shirazi, 2006; Kilanioti, Sotiropoulou, & Hadjiefthymiades, 2005). The dynamic discovery and invocation of services are essential to mobile applications, where the user context may change dynamically, making different kinds of services, or service implementations, adequate at different moments and places. However, bridging mobile devices and SOAs requires analysis of some design issues, along with the fixing of diverse problems related to using resources and protocols primarily aimed at wired use, as discussed in the next sections.
INTEGRATING MOBILE DEvICES AND SOAS Devices may assume three different roles in a SOA: service provider, service consumer, or service registry. In what follows, we examine the most representative high-level scenarios of how mobile devices work in each situation.
Consuming Services The idea is to make available, in a wired infrastructure, a set of services that can be discovered and used by mobile devices. In this context, different designs can be adopted for bridging mobile devices and service providers. Two major architectural configurations can be derived and adapted to different contexts (Duda, Aleksy, & Butter, 2005): direct communication and proxy aided communication. In Figure 1 we illustrate the use of direct communication. Figure 1. Direct communication between mobile client and SOA infrastructure
Query
Service Registry
Publish
Service Provider Interact Interact
In this approach, applications running at the devices directly contact service registries and service providers. This approach assumes the usage of fat clients with considerable processing, storage, and networking capabilities. This is necessary because mobile clients need to run applications coupled with SOA-defined protocols, which may not be suited for usage by resource-constrained devices. However, most portable devices are rather resource-constrained devices. Thus, considering running on mobile devices applications with significant requirements of processing and memory footprint reduces the range of possible client devices. This issue leads us to the next approach, proxy-aided communication, illustrated in Figure 2. In this architectural variation, a proxy is introduced between the mobile device and the SOA infrastructure, playing the role of mobile device proxy in the wired network. This proxy interacts via SOA-defined protocols with registries and service providers, and may perform a series of content adaptations, returning to mobile devices results using lightweight protocols and data formats. This approach has several advantages over the previous one. The proxy may act as a cache, storing data of previous service invocations as well as any client relevant information, such as bookmarks and profiles. Proxies may also help client devices by transforming complex data into lightweight formats that could be rapidly delivered through wireless channels and processed by resource-constrained devices.
Advertising Services In a general way, mobile devices have two choices for advertising services (Loureiro et al., 2006): the push-based approach and the pull-based approach. In the first one, illustrated in Figure 3, service providers periodically send the descriptions of the services to be advertised directly to potential clients, even if they are not interested in such services (1). Clients update local registries with information about available services (2), and if some service is needed, clients query their own registries about available providers (3). In the pull-based approach, clients only receive the description of services when they require a service discovery. This process can be performed in two ways, either through centralized or distributed registries. In the former, illustrated in Figure 4, service descriptions are published in central servers (1), which maintain entries about available services (2). Clients then query this centralized registry in order to discover the services they need (3). In the distributed registry approach, illustrated in Figure 5, the advertisement is performed in a registry contained in each provider (1). Therefore, once a client needs to discover a service, it will have to query all the available hosts in the environment (2) until discovering some service provider for the needed service (3).
Bridging Together Mobile and Service-Oriented Computing
Figure 2. Proxy intermediating communication between mobile client and SOA Query (SOA protocols)
B
Service Registry Publish
Interact/Query (Lightweight protocols)
Service Provider Proxy Interact (SOA protocols)
Figure 3. Push-based approach
Service “X”??
Service Registry
Service Registry
Service Registry
Figure 4. Pull-based approach with centralized registry Service “X”??
Service Registry
ISSUES ON INTEGRATING MOBILE DEvICES AND SOAs Regardless of using mobile devices for either consuming or advertising services in SOAs, both mobility and the limitations of these devices are raised as the major issues for this
Service “Y”??
integration. Designing and deploying effective services aimed at mobile devices requires careful analysis of diverse issues related to this kind of service provisioning. Next, we depict several issues that arise when dealing with mobile devices in SOAs. This list is not exhaustive, but rather representative of the dimension of parameters that should be balanced when designing services for mobile use.
Bridging Together Mobile and Service-Oriented Computing
Figure 5. Pull-based approach with distributed registries Service “X”
Service Registry
Service “Z”??
Service “Z”
No, Sorry
Service “Y”
Service Registry
Yes!!
Service “Z”??
Suitability of Protocols and Data Formats SOAs are primarily targeted at wired infrastructures. Conversely, small mobile devices are known by their welldocumented limitations. Thus, protocols and formats used in conventional SOAs may be inadequate for use with resource-constrained wireless devices (Pilioura, Tsalgatidou, & Hadjiefthymiades, 2002; Kilanioti et al., 2005). For instance, UDDI and SOAP are, respectively, standard protocols for service discovery and messaging in Web services-based SOAs. When using UDDI for service discovery, multiple costly network round trips are needed. In the same manner, SOAP messages are too large and require considerable memory footprint and CPU power for being parsed. Hence, these two protocols impact directly in the autonomy of battery-enabled devices.
Disconnected and Connected Services In the scope of smart spaces, where disconnections are the norm rather than the exception, we can identify two kinds of services (Chen et at., 2005): disconnected and connected services. The first ones execute by caching the inputs of users in the local device. Once network connectivity is detected, the service performs some sort of synchronization. Services for messaging (e.g., e-mail and instant messages) and field research (e.g., gathering of data related to the selling of a specific product in different supermarkets) are some examples of services that can be implemented as disconnected ones.
Service Registry
Connected services, on the other hand, are those that can only execute when network connectivity is available in the device. Some examples of connected services include price checking, ordering, and shipment tracking. Note, however, that these services could certainly be implemented as disconnected services, although their users will generally need the information when demanded, neither before nor later. Therefore, there is no precise categorization of what kind of services would be connected or disconnected, as this decision is made by the system designer.
User Interface User interfaces of small portable devices are rather limited in terms of screen size/resolution and input devices, normally touch screens or small built-in keyboards. This characteristic favors services that require low interaction to complete transactions (Pilioura et al., 2002). Services requiring many steps of data input, such as long forms, tend to: stress users, due to the use of non-comfortable input devices; reduce device autonomy, due to the extra time for typing data; and increase the cost of data transfer, due to larger amounts of data being transferred. A possible alternative for reducing data typing by clients is the use of context-aware services (Patterson, Muntz, & Pancake, 2003). Context-aware services may reduce data input operations of mobile devices by inferring, or gathering through sensors, information about a user’s current state and needs.
Bridging Together Mobile and Service-Oriented Computing
Frequent Temporary Disconnections Temporary disconnections between mobile device and service provider are common due to user mobility. Thus, both client applications and service implementations must consider the design of mechanisms for dealing with frequent disconnections. Different kinds of services require distinct solutions for dealing with disconnections. For instance, e-business applications need machineries for controlling state of transactions and data synchronization between mobile devices and service providers (Sairamesh, Goh, Stanoi, Padmanabhan, & Li, 2004). Conversely, streaming service requires seamless reestablishment and transference of sessions between access points as the user moves (Cui, Nahrstedt, & Xu, 2004).
Security and Privacy Normally, mobile devices are not shared among different users. Enterprises may benefit from this characteristic for authenticating employees, for instance. That is, the system knows the user and his/her access and execution rights based on profiles stored in his/her mobile devices. However, in commercial applications targeted at a large number of unknown users, this generates a need of anonymity and privacy of consumers. This authentication process could cause problems, for example in case of device thefts, because the device is authenticated and not the user (Tatli, Stegemann, & Lucks, 2005). Security also has special relevance when coping with wireless networks (Grosche & Knospe, 2002). When using wireless interfaces for information exchange, mobile devices allow any device in range, equipped with the same wireless technology, to receive the transferred data. At application layer, service providers must protect themselves from opening the system to untrusted clients, while clients must protect themselves from exchanging personal information with service providers that can use user data for purposes different than the ones implicit in the service definition.
Device Heterogeneity and Content Adaptation Modern mobile devices are quite different in terms of display sizes, resolutions, and color capabilities. This requires services to offer data suitable for the display of different sorts of devices. Mobile devices also differ in terms of processing capabilities and wireless technologies, which makes harder the task of releasing adequate data and helper applications to quite different devices. Therefore, platform-neutral data formats stand out as the ideal choice for serving heterogeneous sets of client devices. Another possible approach consists of using on-demand data
adaptation. Service providers may store only one kind of best-suited data format and transform the data, for example, using a computational grid (Hingne, Joshi, Finin, Kargupta, & Houstis, 2003), when necessary to transfer the data to client devices. Moreover, dynamic changes of conditions may also require dynamic content adaptation in order to maintain pre-defined QoS threshold values. For instance, users watching streamed video may prefer to dynamically reduce video quality due to temporary network congestion, therefore adapting video data, and to maintain a continuous playback instead of maintaining quality and experiencing constant playback freezing (Cui et al., 2004).
Consuming Services As discussed before, system architects can choose between two major approaches for accessing services of SOAs from mobile devices: direct communication and proxy-aided communication. The two approaches have some features and limitations that should be addressed in order to deploy functional services. Direct communication suffers from the limitations of mobile devices and relates to other discussions presented in this article, such as adequacy of protocols and data formats for mobile devices and user interface. If, on the one hand, proxy-aided communication seems to be the solution for problems of the previous approach, on the other hand it also brings its own issues. Probably the most noted is that proxies are single points of failures. Furthermore, some challenges related to wired SOAs are also applicable to both approaches discussed. Service discovery and execution need to be automated to bring transparency and pervasiveness to the service usage. Moreover, especially in the context of smart spaces, services need to be personalized according to the current user profile, needs, and context. Achieving this goal may require describing the semantics of services, as well as modeling and capturing the context of the user (Chen, Finin, & Joshi, 2003).
Advertising Services A number of issues and technical challenges are associated with this scenario. The push-based approach tends to consume a lot of bandwidth of wireless links according to the number of devices in range, which implies a bigger burden over mobile devices. Using centralized registries creates a single point of failure. If the registry becomes unreachable, it will not be possible to advertise and discover services. In the same manner, the discovery process is the main problem with the approach of distributed registries, as it needs to be well designed in order to allow clients to query all the hosts in the environment. Regardless of using centralized or distributed registries, another issue rises with mobility of service providers. When
B
Bridging Together Mobile and Service-Oriented Computing
service providers move between access points, a new address is obtained. This changing of address makes service providers inaccessible by clients that query the registry where it published its services. Mechanisms for updating the registry references must be provided in order for services to continue to be offered to their requestors.
FUTURE TRENDS The broad list of issues presented in this article gives suggestions about future directions for integrating SOC and mobile devices. Each item depicted in the previous section is already an area of intensive research. Despite this, both SOC and mobile computing still lack really functional and mature solutions for the problems presented. In particular, the fields of context-aware services and security stand out as present and future hot research fields. Besides, the evolution itself of mobile devices towards instruments with improved processing and networking power, as well as better user interfaces, will reduce the complexity of diverse challenges presented in this article.
CONCLUSION In this article we have discussed several issues related to the integration of mobile devices and SOC. We have presented the most representative architectural designs for integrating mobile devices to SOAs, both as service providers and service consumers. While providing means for effective integration of mobile devices and service providers, SOC has been leveraging fields such as mobile commerce and pervasive computing. Nonetheless, several issues remain open, requiring extra efforts for designing and deploying truly functional services.
REFERENCES Bellur, U., & Narendra, N. C. (2005). Towards service orientation in pervasive computing systems. Proceedings of the International Conference on Information Technology: Coding and Computing (ITCC’05) (Vol. II, pp. 289-295). Chen, H., Finin, T., & Joshi, A. (2003). An ontology for context-aware pervasive computing environments. The Knowledge Engineering Review, 18(3), 197-207. Chen, M., Zhang, D., & Zhou, L. (2005). Providing Web services to mobile users: The architecture design of an m-service portal. International Journal of Mobile Communications, 3(1), 1-18.
Cui, Y., Nahrstedt, K., & Xu, D. (2004). Seamless user-level handoff in ubiquitous multimedia service delivery. Multimedia Tools Applications, 22(2), 137-170. Duda, I., Aleksy, M., & Butter, T. (2005). Architectures for mobile device integration into service-oriented architectures. Proceedings of the 4th International Conference on Mobile Business (ICBM’05) (pp. 193-198). Grosche, S.S., & Knospe, H. (2002). Secure mobile commerce. Electronics & Communication Engineering Journal, 14(5), 228-238. Hingne, V., Joshi, A., Finin, T., Kargupta, H., & Houstis, E. (2003). Towards a pervasive grid. Proceedings of the 17th International Parallel and Distributed Processing Symposium (IPDPS’03) (p. 207.2). Huhns, M.N., & Singh, M.P. (2005). Service-oriented computing: Key concepts and principles. IEEE Internet Computing, 9(1), 75-81. Kalasapur, S., Kumar, M., & Shirazi, B. (2006). Evaluating service oriented architectures (SOA) in pervasive computing. Proceedings of the 4th IEEE International Conference on Pervasive Computing and Communications (PERCOMP’06) (pp. 276-285). Kilanioti, I., Sotiropoulou, G., & Hadjiefthymiades, S. (2005). A client/intercept based system for optimized wireless access to Web services. Proceedings of the 16th International Workshop on Database and Expert Systems Applications (DEXA’05) (pp. 101-105). Loureiro, E., Bublitz, F., Oliveira, L., Barbosa, N., Perkusich, A., Almeida, H., & Ferreira, G. (2007). Service provision for pervasive computing environments. In D. Taniar (Ed.), Encyclopedia of mobile computing and commerce. Hershey, PA: Idea Group Reference. Papazoglou, M. P., & Georgakopoulos, D. (2003). Serviceoriented computing: Introduction. Communications of the ACM, 46(10), 24-28. Patterson, C. A., Muntz, R. R., & Pancake, C. M. (2003). Challenges in location-aware computing. IEEE Pervasive Computing, 2(2), 80-89. Pilioura, T., Tsalgatidou, A., & Hadjiefthymiades, S. (2002). Scenarios of using Web services in m-commerce. ACM SIGecom Exchanges, 3(4), 28-36. Sairamesh, J., Goh, S., Stanoi, I., Padmanabhan, S., & Li, C. S. (2004). Disconnected processes, mechanisms and architecture for mobile e-business. Mobile Networks and Applications, 9(6), 651-662. Tatli, E. I., Stegemann, D., & Lucks, S. (2005). Security challenges of location-aware mobile business. Proceedings
Bridging Together Mobile and Service-Oriented Computing
of the 2nd IEEE International Workshop on Mobile Commerce and Services (WMCS’05) (pp. 84-95). Weiser, M. (1991). The computer for the 21 century. Scientific American, 265(3), 66-75. st
Zhu, F., Mutka, M. W., & Ni, L. M. (2005). Service discovery in pervasive computing environments. IEEE Pervasive Computing, 4(4), 81-90.
KEY TERMS Grid Service: A kind of Web service. Grid services extend the notion of Web services through the adding of concepts such as statefull services. Jini: Java-based technology for implementing SOAs. Jini provides an infrastructure for delivering services in a network
Mobile Device: Any low-sized portable device used to interact with other mobile devices and resources from smart spaces. Examples of mobile devices are cellular phones, smart phones, PDAs, notebooks, and tablet PCs. Proxy: A network entity that acts on behalf of another entity. A proxy’s role varies since data relays to the provision of value-added services, such as on-demand data adaptation. Streaming Service: One of a number of services that transmit some sort of real-time data flow. Examples of streaming services include audio streaming or digital video broadcast (DVB). Web Service: Popular technology for implementing SOAs built over Web technologies, such as XML, SOAP, and HTTP.
B
Category: Human Factor
Browser-Less Surfing and Mobile Internet Access Gregory John Fleet University of New Brunswick at Saint John, Canada Jeffery G. Reid xwave Saint John, Canada
INTRODUCTION Lately, we have seen the use of a number of new technologies (such as Javascript, XML, and RSS) used to show how Web content can be delivered to users without a traditional browser application (e.g., Microsoft Explorer). In parallel, a growing number of PC applications, whose main job previously was to manage local resources, now are adding Internet connectivity to enhance their role and use (e.g., while iTunes started as a media player for playing and managing compressed audio files, it now includes Web access to download and purchase music, video, podcasts, television shows, and movies). While most attempts at providing Internet access on mobile devices (whether wireless phones or personal digital assistants) have sought to bring the traditional browser, or a mobile version of the browser, to these smaller devices, they have been far from successful (and a far cry from the richer experience provided by browsers on the PC using standard input and control devices of keyboards and a mouse). Next, we will highlight a number of recent trends to show how these physical and use-case constraints can be significantly diminished.
BACKGROUND Mobile telephony and mobile computing continue to display unprecedented growth worldwide. Zee News (2005) reports that in some parts of the world, such as India, mobile phones are now more popular than traditional landline phones. Since 2000, many developed countries have spent large amounts of money on the installation and deployment of wireless communication infrastructure (Kunz & Gaddah, 2005). And this growth trend is not confined to the mobile phone handset market. It is also being experienced across other mobile devices. In fact, in 2004 more mobile phones shipped than both automobiles and personal computers (PCs) combined, making them the fastest adopted consumer product of all time (Clarke & Flaherty, 2005). Further, Wiberg (2005) points out that this increase in mobile device usage spans across business and non-business usage. Therefore, this growth is
not simply due to increased consumer demand; businesses are continually seeing new value in equipping employees with mobile computing and communication devices. There has also been steady growth in the use of the Internet, as well as in the nature of Internet usage. The size of the Internet, measured in terms of the number of users, is more than 800 million users (Global Reach, 2004). While the majority of the users are English, other languages are experiencing significant growth in the number of users, and this growth is expected to continue, given the large numbers of non-English-speaking populations. Some of the drivers for the increase in Internet usage include the growth in Web-enabled applications and the availability of high-speed, always-on Internet (Bink, 2004). Kunz and Gaddah (2005) identify two broad technological developments that are converging to enable mobile computing (the use of the Internet through mobile devices). The first of these technological developments is the accessibility to the Internet regardless of location, as evidenced by the growth in wireless hotspots. Now users can connect to the Internet from various locations and access Internet content without being connected to a physical local area network (LAN) connection or other type of landline connection. The second technological development is the drive to reduce the size of computer hardware (Kunz & Gaddah, 2005). This size reduction increases the portability of these devices, leading to the mobile nature of the devices as well as the desire to connect these devices to the Internet. Unfortunately, being able to provide Internet access to mobile devices has not ensured a quality Web experience. The next section will profile the current mobile Web experience.
USER EXPERIENCE OF WEB ON MOBILE DEvICES The Web Browser on a PC Let us start with the typical experience of the Web. The most common way to navigate the Internet is through the
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Browser-Less Surfing and Mobile Internet Access
use of a browser, a software application that allows the user to locate and display Web pages (Webopedia, 2006). On the personal computer (PC), there are a variety of browsers available, including Microsoft Internet Explorer, Mozilla Firefox, Opera, Netscape, Apple Safari, and Konqueror (Wikipedia, 2006a). A cross-section of definitions from the Web outlines the basic functionality of these browsers (http://www.google. com/search?hl=en&lr=&q=define%3A+web+browser&bt nG=Search); the Web browser is a graphical interface (i.e., icons, buttons, menu options) that: • •
interprets HTML files (resources, services) from Web servers, and formats them into Web pages; and provides the ability to both view and interact with Web content (including download and upload of media content).
Yet most modern browsers also include additional functionality, assisting with the management of the tool’s functionality and the content to which they provide access. This functionality includes: • •
• • •
Bookmarking: The ability to save and manage Web addresses. Cookies and Form-Filling: The ability of the browser to pre-fill form fields (e.g., address or contact information), or provide the Web server with identifying information in order to customize the content received from the server. Searching: The ability to conduct a Web or local file search. History: The automatic cataloging of previously visited Web sites. Display Modification: The ability to customize the way Web content is displayed (e.g., size of text, types of media files that can be viewed, etc.)
It is also important to note that in the typical use of a Web browser, the user searches for information on the Web, often starting with a broad search and successively narrowing that search to meet his or her information goal (i.e., to go from the general to the specific).
The Web Browser on a Mobile Device For mobile devices, such as cell phones or personal digital assistants (PDAs), the Web browser application is often referred to as a microbrowser (also minibrowser and mobile browser; see Wikipedia, 2006b). The difference between a full browser and the microbrowser is that the code in the microbrowser application has been optimized to accommodate the smaller screens, memory, and bandwidth limitations of mobile devices. In addition, the Web servers often communicate
with these microbrowsers using variations on the standard HTML (hypertext markup language), again to accommodate the screen, memory, and bandwidth restrictions. Internet usage on mobile devices poses a number of challenges that are different than those found on a traditional computing device such as a PC (Becker, 2005). As mentioned previously, the computing power (processor and memory configuration), the transmission bandwidth, and screen size on the mobile device are really just a fraction of what users have available to them on a PC. More importantly, the limitations in screen size and physical interface often require users to restrict the activities they might otherwise seek to accomplish on the Web. The physical restrictions (that being the telephone keypad and four-way scroll and navigation keys) can be quite significant. On a PC, we have a full-sized QWERTY keyboard and mouse interface for entering searches and addresses, or navigating Web pages. On the mobile device, in particular the cell phone, these physical input and control devices are replaced with a keypad designed for dialing phone numbers (not entering text strings), and horizontal/vertical navigation keys that significantly slow simple scrolling and selection of content.1 In user studies, Chen, Xie, Ma, and Zhang (2005) report that users, when browsing the Web on a phone, handheld computer, or personal digital assistant, spend the majority of their time scrolling the screen to locate and select the content of interest. Despite these real challenges, Nugent (2005) expects that the need for mobile Web browsing will increase, and people will want these devices to stay small, weigh less, cost less, run cooler and longer on one charge, but continue to do more than today’s devices. Lawton (2001) believes that meeting these needs will require faster wireless connections, larger displays, as well as new usage paradigms and/or content that fits these smaller devices. This is the environment mobile users are operating in today. A user can either struggle with a small screen and content that does not fit within that screen, or lug around a larger device that has an adequately sized screen but more limited connectivity options.
Technology and Service Barriers There are a number of technological hurdles that need to be overcome for widespread adoption of mobile Internet usage. Chan and Fang (2005) identify a number of technological barriers, which range from connectivity and bandwidth issues to the lack of standards and broad use of proprietary tools and languages. Kuniavsky (2006) also notes the numerous and often complex relationships that exist between the multiple service, application, and technology providers currently needed to deliver mobile computing to the user, and how none of these players is wholly responsible for the resulting user experience.
B
Browser-Less Surfing and Mobile Internet Access
Identification of the challenges associated with mobile Internet usage is only part of the problem. After the issues have been identified, developing solutions to deal with the problems is the next step in the process. Many manufacturers are presently making moves to deal with some of these identified issues. One common move is to increase the screen size of the mobile device. One example of this trend is the Sony Ericsson P800 SmartPhone, which has increased the size of its display area while attempting to maintain the overall size of the device itself. Another more recent example is the Sony Ericsson P910i, with its larger screen, miniature QWERTY keyboard, and pen-based interface. Another design approach to deal with limited screen space is to focus on the content rather than the size of the screen, as is attempted with standards such as WAP, WML, HDML, as well as services such as i-mode (Chen et al., 2005). What is needed, though, is the development of Web content and mobile applications that can be viewed, navigated, and controlled from small devices (Nayak, 2005), because, at this time, consumers find the small screen display and small buttons on these devices difficult to use. Chan and Fang (2005) believe that these technologies need to mature, and until that time, the mobile Internet will be geared toward applications requiring limited bandwidth, short exchange of data and text, and simple functionality. Therefore, using smaller mobile devices to perform tasks similar to those carried out on a traditional computing platform poses challenges for users and manufacturers alike.
FUTURE TRENDS Interestingly, there are a number of new technologies and trends that might suggest an evolution of the mobile computing Web experience. This evolution comes from a number of different places. In this section of the article, a few specific trends are highlighted in order to demonstrate this potential.
Web-Enabled Desktop Clients In recent years, we are seeing more and more desktop clients (or applications) reaching beyond the processes of the PC and the content on the local drive to networked and Internet resources and content. Apple’s iTunes media player was first released as a desktop application for playing music files from one’s hard drive. Since that first release, it has grown to not only allow streaming of music libraries over networks, but now has a built-in Web browser tied to one of the most successful online music stores today. There are additional examples of traditional desktop clients that have added Web connectivity to their functional specifications; these include desktop applications with built-
0
in version checking; address book applications that communicate with LDAP servers; Google desktop™, providing the ability to search and find information not only on your local drives, but also on e-mail and Web servers; and cataloguing programs that match your own library of CDs, books, or DVDs with online databases such as Internet Movie Database (imd.com) or Amazon.com. I suspect this trend is only just beginning, and we will continue to see additional examples as software companies add both Internet connectivity and imbedded browsers into desktop applications in order to add new and unique value for users.
Webtop Clients At the same time, there are also some exciting examples of Web applications (or services) that only require a standard Web browser. Web services have been around since the beginning of the Web, but what differentiates these newer Web applications is their attention to usability and responsiveness, resulting in a Webtop client that responds and behaves in ways similar to a desktop client. For example, Flickr.com allows individuals to upload photos from their cameras and hard drives to the Flickr Web servers. Then, in desktop-client fashion, they allow us to arrange the order with simple drag-and-drop, or name, edit, and tag photo labels by directly clicking on the titles within the Web browser. Other examples include the Web services of Google, MyYahoo, and MyMSN, as well as the excellent services from 37signals.com (Basecamp, Backpack, Writeboard, Ta-Da List, and the growing number of services built using the Ruby on Rails development environment). In all these examples, the responsiveness of these Web clients is quite impressive, mimicking the behavior of their desktop counterparts.
The Changing Mental Model of Web Access These examples of desktop and Webtop clients demonstrate a blurring of the lines between Web browsing or surfing and running local applications. I believe this is a good thing, since it suggests that Web connectivity is not limited to what is accomplished and viewed through a browser. And for mobile devices, this lack of distinction should also be a good thing—allowing users to think about Internet content separate from traditional browsers. This could also suggest that users might adapt their user model of expecting Internet content (especially on small devices) to be only through an Internet browser. Seeing the Internet on mobile devices as separate from a browser is only one (significant) step in producing a better user experience. It is also important to recognize the other constraints that limit a quality Internet experience
Browser-Less Surfing and Mobile Internet Access
on these small devices: the constrained visual and physical interface. Reproducing Web sites onto small screens, at best, requires the ability to visualize content beyond the screen, and, at worst, produces a frustrating, unacceptable experience.
SOLvING THE vISUAL LIMITS OF MOBILE DEvICES A variety of technologies (XML, ATOM, Javascript, WebKit) have been used of late to create a number of useful Web services. One of the most common is RSS feeds, where the user can subscribe to the content found on a Web server. The current implementation of RSS satisfies two user goals: to filter Web content to only those topics of interest, and to provide real-time notification of updates to the Web site. Therefore, RSS provides a technology to allow users to browse the Web in a more focused manner, providing personalized views of self-selected content. Another example of viewing self-selected Web content is found with Yahoo’s Konfabulator (also known as the Yahoo! Widget Engine—see http://widgets.yahoo.com). This desktop client is a real-time Javascript compiler that can execute small Javascript files (called widgets) to accomplish whatever task they have been programmed to accomplish. The result is a small, windowless, (and in the case of scripts that communicate with Web servers) browser-less view of live Web content. Figure 1 shows an example using the Weather widget, which displays live weather conditions and the five-day forecast for a particular location configured by the user. Additional Web information is available with mouse-over or clicks on the widget. The latest operating system from Apple (known as Tiger or 10.4) has also added similar functionality for displaying Figure 1. The Weather widget, using Yahoo’s Widget Engine, showing current and forecasted weather for Palo Alto, California
self-selected Web content. More appropriately named, these Dashboard™ widgets organize and/or present Web content in a way that is easy to read. Presently, thousands of widgets are available for download, whether from Yahoo’s Web site or Apple.com (as of January 2006, there were more than 4,000 widgets available for download). What is interesting about widgets is the fact that most are designed to present their Web content in a fairly constrained visual space, separate from large resource requirements or visual real estate needs. In other words, these widgets provide what could be a perfect example of self-selected, rich Internet content for small screens. Another interesting example comes from some software developers in Japan. They have demonstrated the ability to create a Dashboard™ widget of a full Web browser, only miniaturized to dimensions that could easily work on a standard PDA-size screen (see http://hmdt-web.net/shiira/ mini/en). Therefore, with a high-quality display such as that found on today’s mobile devices and iPods, it is quite easy to imagine using this miniaturized view of Web pages to surf the Web, especially on Web sites where the format and layout is familiar. Now we turn to the problem of the physical (input and control) interface on mobile devices. If you have this view of a miniaturized Web page, how do you move around and select the buttons and links on the page? Using a four-way scroll key might work for the limited content in most Widgets, but is a very poor substitute for a keyboard and mouse when browsing a full (though miniaturized) Web page.
Solving the Physical Interface of Mobile Devices A keyboard and mouse is not just the standard input and control device for Web surfing, but provides a rich interaction for control and text input. A telephone keypad and four-way scroll key does not even come close to that user experience, therefore making the Web experience on mobile devices very constrained indeed. Yet we have an excellent example of one specific mobile device that has, over the past four years, shown that navigation of large hierarchies of data can be very quickly accomplished with only a finger or thumb. The iPod music player provides both a simple and extremely intuitive interface for moving through and selecting from vast playlists, photos, folders, and files—using the scroll wheel or click wheel2 design. The click wheel interface is currently only available on Apple’s iPod music and video players, but there has been much discussion of the possibility of this interface being used on other small devices, such as cell phones or PDAs (see Shortflip, 2006; Baig, 2005). After experiencing how easy it is to use an iPod to navigate and select music, photos, or videos, it is not difficult to imagine the same physical interface being
B
Browser-Less Surfing and Mobile Internet Access
used to select phone numbers from an address book, and even select and navigate content on a Web page. More recently, there has been discussion of some patents filed at the U.S. Patent Office site (see http://tinyurl. com/8zxuv). The patent application document demonstrates a device where the whole front is a display screen, and the control interface comes from a touch-activated, touch-sensitive click wheel that is available from the visual display wherever the user touches the screen. Therefore, the combination of a large screen, with a touch-activated telephone keypad and/or click wheel, provides a compelling possibility for a high-fidelity Web experience on a mobile device.
CONCLUSION In this article, we have argued that today’s mobile devices provide a poor user experience when presenting Internet and Web content. These devices have two major physical constraints: a visual display that is too small to present the typical browser-based view of the Web, and a physical interface (telephone keypad and four-way scroll key with center select) that is not easily adapted to text entry or interface control. Users, therefore, are unable to reproduce the familiar encyclopedic browsing of Web content on these miniature visual and physical interfaces. At the same time, a number of new trends has been highlighted, demonstrating the possibility of producing a much stronger and more compelling user experience of the Web on small mobile devices.
REFERENCES Baig, E. C. (2005). New iTunes phone a snazzy device. USA Today.com. Retrieved February 1, 2006, from http://www. usatoday.com/tech/columnist/edwardbaig/2005-09-07itunes-phone_x.htm Becker, S. A. (2005). Web usability. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 5, pp. 3074-3078). Hershey, PA: Idea Group Reference. Bink, S. (2004). Browserless Net use on the rise. Bink.nu. Retrieved February 1, 2006, from http://bink.nu/Article798.bink Chan, S. S., & Fang, X. (2005). Interface design issues for mobile commerce. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 3, pp. 16121616). Hershey, PA: Idea Group Reference. Chen, Y., Xie, X., Ma, W., & Zhang, H. (2005). Adapting Web pages for small-screen devices. IEEE Internet Computing.
Retrieved February 1, 2006, from http://research.microsoft. com/~xingx/tic1.pdf Clarke, I., & Flaherty, T. (2005). Portable portals for mcommerce. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 4, pp. 2293-2296). Hershey, PA: Idea Group Reference. Fleet, G. J. (2003, October 16-18). The devolution of the Web browser: The fracturing of Internet Explorer. Proceedings of the Atlantic Schools of Business Conference, Halifax, Nova Scotia, Canada. Global Reach. (2004). Global Internet statistics by language. Retrieved February 1, 2006, from http://global-reach.biz/ globstats/index.php3 Kuniavsky, M. (2006). User experience and HCI. In J. Jacko & A. Sears (Eds.), The human-computer interaction handbook (2nd ed.). Retrieved February 1, 2006, from http://www. orangecone.com/hci_UX_chapter_0.7a.pdf Kunz, T., & Gaddah, A. (2005). Adaptive mobile applications. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 1, pp. 47-52). Hershey, PA: Idea Group Reference. Lawton, G. (2001). Browsing the mobile Internet. IEEE Computer, 35(12), 18-21. Nugent, J. H. (2005). Critical trends in telecommunications. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 1, pp. 634-639). Hershey, PA: Idea Group Reference. Nayak, R. (2005). Wireless technologies to enable electronic business. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 5, pp. 3101-3105). Hershey, PA: Idea Group Reference. Shortflip.com. (2006). The future of the iPod. Retrieved February 1, 2006, from http://www.shortflip.com/article/TheFuture-of-the-iPod-149.html Webopedia. (2004). Browser. Accessed February 1, 2006, from http://www.webopedia.com/TERM/B/browser.html Wiberg, M. (2005). “Anytime, anywhere” in the context of mobile work. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and technology (Vol. 1, pp. 131-134). Hershey, PA: Idea Group Reference. Wikipedia. (2006a). Web browser. Accessed February 1, 2006, from http://en.wikipedia.org/wiki/Web_browser Wikipedia. (2006b). Microbrowser. Accessed February 1, 2006, from http://en.wikipedia.org/wiki/Microbrowser
Browser-Less Surfing and Mobile Internet Access
Zee News. (2005). Mobile phones outpace landline but with grey calls. Retrieved February 1, 2006, from http://www. zeenews.com/znnew/articles.asp?aid=194056&sid=ZNS
KEY TERMS
Ruby on Rails: An new open-source Web application framework. WAP: Wireless application protocol. WebKit: Application framework for Apple’s Safari Web browser. WML: Wireless Markup Language.
Atom: One of a number of Web formats that supports user subscription to online content. Click Wheel: The physical interface on Apple’s iPod for moving through the directories and selecting items. HDML: Handheld Device Markup Language. i-mode: A popular wireless Internet service initially available only in Japan. Javascript: Scripting programming language. LDAP: Lightweight directory access protocol. Podcast: The distribution of audio or video content over the Web using Atom or RSS. RSS: Rich site summary or really simple syndication.
XML: eXtensible Markup Language.
ENDNOTES 1
2
It is true that some PDAs and cell phones are using miniature QWERTY keyboards for an input device, though the tiny size is only marginally better than the keypad. The click wheel interface allows the user to navigate a vertical array of items or folders by rotating the wheel either clockwise or counterclockwise. Selecting an item in the list or moving deeper into the folder structures can be accomplished with the center button or the four buttons placed 90 degrees apart.
B
Category: PP
Building Web Services in P2P Networks Jihong Guan Tongji University, China Shuigeng Zhou Fudan University, China Jiaogen Zhou Wuhan University, China
INTRODUCTION Nowadays peer-to-peer (P2P) and Web services are two of the hottest research topics in computing. Roughly, they appear as two extremes of distributed computing paradigm. Conceptually, P2P refers to a class of systems and applications that employ distributed resources to perform a critical function in a decentralized way. A P2P distributed system typically consists of a large number of nodes (e.g., PCs connected to the Internet) that can potentially be pooled together to share their resources, information, and services. These nodes, taking the roles of both consumer and provider of data and/or services, may join and depart the P2P network at any time, resulting in a truly dynamic and ad-hoc environment. Apart from improving scalability by avoiding dependency on centralized servers, the distributed nature of such a design can eliminate the need for costly infrastructure by enabling direct communication among clients, along with enabling resource aggregation, thus providing promising opportunities for novel applications to be developed (Ooi, Tan, Lu, & Zhou, 2002). On the other hand, Web services technologies provide a language-neutral and platform-independent programming model that can accelerate application integration inside and outside the enterprise (Gottschalk, Graham, Kreger, & Snell, 2002). It is convenient to construct flexible and loosely coupled business systems by application integration under a Web services framework. Considering Web services are easily applied as wrapping technology around existing applications and information technology assets, new solutions can be deployed quickly and recomposed to address new opportunities. With the acceleration of Web services adoption, the pool of services will grow, fostering development of more dynamic models of just-in-time application and business integration over the Internet. However, current proposals for Web services infrastructures are mainly based on centralized approaches such as UDDI: a central repository is used to store services descriptions, which will be queried to discover or, in a later stage, compose services. Such centralized architecture is prone
to introducing single points of failure and hotspots in the network, and exposing vulnerability to malicious attacks. Furthermore, making full use of Web services capabilities using a centralized system does not scale gracefully to a large number of services and users. This difficulty is severe by the evolving trend to ubiquitous computing in which more and more devices and entities become services, and service networks become extremely dynamic due to constantly arriving and leaving service providers. We explore the techniques of building Web services systems in a P2P environment. By fitting Web services into a P2P environment, we aim to add more flexibility and autonomy to Web services systems, and alleviate to some degree the inherent limitations of these centralized systems. As a case study, we present our project BP-Services. BP-Services is an experimental Web services platform built on BestPeer (http://xena1.ddns.comp.nus.edu.sg/p2p/)a generic P2P infrastructure designed and implemented collaboratively by the National University of Singapore and Fudan University of China (Ng, Ooi, & Tan, 2002).
FITTING WEB SERvICES INTO A P2P FRAMEWORK A Web service can be seen as an interface that describes a collection of operations that are network accessible through standardized XML messaging (Gottschalk et al., 2002). Web services consist of three roles and three operations: the roles are providers, requesters, and registrars of services, and the operations are publish, find, and bind. The service providers are responsible for creating Web services and corresponding service definitions, and then publishing the services with a service registry based on UDDI specification. The service requesters first find the services requested via the UDDI interface, and the UDDI registry provides the requesters with WSDL service descriptions and URLs pointing to these services themselves. With the information obtained, the requesters can then bind directly with the services and invoke them.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Building Web Services in P2P Networks
Figure 1. Network topology of BestPeer
B
LIGLO Servers
LI GLO Ser ver LI GLO Ser verLI GLO Ser ver
Peers
Peer
Peer
Peer Peer
Peer
Peer
Peer
Over the last few years, many P2P systems have been developed and deployed for different purposes and with different technologies, such as Napster (http://www. napster.com/), Gnutella (http://gnutella.wego.com/), and Freenet (http://freenet.sourceforge.com/), to name a few. The architecture of these systems can be categorized into three groups mainly based on their network topologies: centralized P2P, pure P2P, and hybrid P2P systems (Yang & Garcia-Molina, 2001). In a centralized P2P network, there is a central server responsible for maintaining indexes on the metadata for all the peers in the network. Pure P2P is simply P2P systems with fully autonomous peersthat is, all nodes are equal, no functionality is centralized, and the communication between peers is also symmetric. Hybrid P2P is a kind of tradeoff between centralized P2P and pure P2P, which is structured hierarchically with a supernode layer and a normal peers layer. Fitting Web services into P2P framework is to adapt Web services to P2P environment, which results in the socalled P2P Web services, or simply P2P services. Here P2P service is different from the ordinary Web services at least in three aspects. First, typically a peer in P2P services takes all three roles of services provider, consumer, and registrar, whereas in ordinary Web services, a node can typically be a producer and/or a consumer, but not a registrar at the same time. Second, generally speaking, servers in ordinary Web services systems are well-known hosts, with static IP addresses and on the outside of a firewall. However, this is not usually the case in the P2P world. A services node may join or depart the P2P services network at any time. Third, the preferred method of finding Web services in the ordinary architecture is currently through a central repository known as a UDDI operator. Nevertheless, P2P services systems have no central server to hold UDDI registry; each peer node manages its own UDDI registry locally. So, new and efficient mechanisms for services discovery in P2P services environment are required.
Corresponding to the architecture of P2P systems, there may also be three schemes for building P2P services applications: centralized P2P services, pure P2P services, and hybrid P2P services. For centralized P2P services, there is a central server in P2P services systems. However, the central server is not used as a central UDDI registry server; instead it is used for storing metadata of services to facilitate services discovery, which includes business names, services types, URLs, and so forth. In pure P2P services systems, services UDDI registry is distributed on every services node, so there is no need for services publication of the ordinary sense, and UDDI registry maintenance is also simplified because all services information is published and maintained locally. And in hybrid P2P services systems, the supernodes will be used for storing services metadata. It is useful for services discovery to cluster services nodes based on metadata, and then register the nodes in the same cluster under the same supernode.
BP-SERvICES: BESTPEER-BASED WEB SERvICES As mentioned, the BP-Services project aims to develop an experimental P2P-based Web services platform as a test-bed for further P2P and Web services research. BestPeer (Ng et al., 2002) is a generic P2P system with an architecture more pure P2P than hybrid P2P. The BestPeer system consists of two types of nodes: a large number of normal computers (i.e., peers), and a relatively fewer number of Location-Independent Global names Lookup (LIGLO) servers. Every peer in the system runs the BestPeer software, and will be able to communicate and share resources with any other peers. There are two types of data in each peer: private data and public (or sharable) data. For a certain peer, only its public data can be accessed by and shared with other
Building Web Services in P2P Networks
Figure 2. The internals of a BP-Services peer node BPServices Node
Mobile Agent
WSAgent
BPServices Node WSAgent
BPServices Node WSAgent
Cache Manaer
Services Manager User Interface User
Service Discovery Engine
Service Composer
Service Deployer
peers. Figure 1 shows the network topology of BestPeer. In the top layer are LIGLO servers, and in the bottom layer are normal peers.
The Architecture of BP-Services In BP-Services, except for the LIGLO servers adhering to BestPeer, each peer node takes both the roles of a services provider and a services consumer, as well as a services registrar. That is to say, there is no central UDDI registry in BP-Services; all services and their definitions are distributed over the peer nodes. Figure 2 illustrates the internals of a BPServices peer node. There are essentially seven components that are loosely integrated. The first component, also the most important component, is the Services Manager that facilitates services discovery, services composition, and services deploying. Corresponding to its functionalities, the services manager consists of three sub-components: the services discovery engine, the services composer, and the services deployer. The services discovery engine is responsible for the publication and location of services. The services composer provides facilities for defining new composite services from existing services, and editing existing services (local). The services deployer facilitates the binding and invocation of requested services, as well as coordination of composite services.
Services' Keyword Indexes
Local UDDI Registry
Local Services Repository
The second component is the Web Services Agent System, or simply WSAgent. The WSAgent provides the environment for mobile agents to operate on. Each BP-Services node has a master agent that manages the services discovery and services description retrieval. In particular, it will clone and dispatch worker agents to neighboring nodes, receive results, and present to the user. It also monitors the statistics and manages the network reconfiguration policies. The third component is a Cache Manager, which is used for caching the results of services discovery and retrieval. Furthermore, by collaboration among the cache managers, a P2P cache subsystem can be formed under the BP-Services framework so that all peers can share the caching results among themselves. The fourth component is the User Interface, which consists of several interface modules, corresponding to services discovery and retrieval, services composition, and deploying. The other three components are Services Indexes, Local UDDI Registry, and Local Services Repository respectively. The services repository keeps the services provided locally. Local UDDI registry holds the description (or publication) information of local services. And services indexes are simply inverted lists of services keywords extracted from the description information of local services, mainly business names and service types. Extracting and keeping services keywords speeds up services discovery.
Building Web Services in P2P Networks
Neighbor Nodes Finding in BP-Services In BP-Services, given a participant node, its neighbor nodes are defined as those nodes that can provide as many services as possible similar to that in the given node. Here we use the information retrieval method to find neighbor nodes. We can treat each node in BP-Services as a document, whose content is the services description information (UDDI registry) contained in that node. Thus we can cluster the nodes in BP-Services by using documents clustering methods (Baesa-Yates & Ribeiro-Neto, 1999). Roughly speaking, nodes in the same cluster may provide more similar services than those from different clusters. However, traditional documents clustering methods are based on a global data view, which is not realistic because it is not easy, if not impossible, to gather data of all nodes in a dynamic P2P network. In BP-Services, we adopt a simple local clustering strategy. We use a Boolean model to represent a peer services node, for the Boolean model is easier to evaluate than the vector space model (VSM), and it is difficult to set the document vector space without deterministic global data view in a P2P environment. Given a peer services node p, there exists a set of keywords extracted from the services description document of p. We denote the set of keywords Kp, and treat it equal to the node p itself. For two services nodes p and q, suppose their keyword sets are Kp and Kq; the similarity of the two nodes are defined as follows. Here, |•| indicates the cardinality of a set.
sim( K p , K q ) =
| K p ∩ Kq | | K p ∪ Kq |
.
(1)
As in BestPeer, when a services node would like to become a participant of BP-Services, then it first registers with a LIGLO server, and the LIGLO server will issue the node with a global and unique identifier (i.e., BPID, BestPeerID), and meanwhile the LIGLO server will also send the node a list of peer nodes that have already registered in the network (i.e., the initial direct peers of the node). We term the links between these initial direct peers and the new participant node the initial links of the node. After joining the network, the node (say p) can begin to find its neighbors by the following steps: (1) Through the ping-pong messages, it contacts the set of peers within a certain number (say k) of hops away from it. Let denote the set of peers as Peer(p, k)={q1,q2,…qn}, and get these peers’ keywords sets {Ki|i=1∼n}. (2) Calculate the similarity of p and each peer in Peer(p, k)that is, {sim(p, qi)| i=1∼n}. (3) Suppose q is the peer in Peer(p, k) that has largest similarity with p, then take q as p’s neighbor node, and connect p and q by a direct link, which is termed neighbor link of p and q.
Through the process of neighbor finding, the peers that share services tend to be connected together by neighbor links, and consequently form clusters of services peers. Considering the dynamism of the P2P system, the peers should update their neighbors regularly.
Services Discovery in BP-Services Services discovery is the key process of P2P services. Because P2P services’ UDDI registries are distributed on the peer nodes, it is inefficient to search the targeted services by traversing all peers one by one. Note that service discovery in P2P is different from P2P information retrieval. In service discovery, once a service that satisfies the requester’s requirements is found, the discovery process can be stopped. It is not necessary to find a lot of similar services for a certain requester’s specific service requirements. In BP-Services, once a requester submits his or her service requirement, say a service query Q, the following process will be launched: 1. 2.
3.
4.
5. 6.
Extracting keywords from Q, the service search process is equal to carrying out keywords matching in information retrieval. First, search at the local peer. The searching task is done by using the local services indexes as in traditional IR. If there are services matching the query, then go to (3); otherwise, go to (4). Return the matched services’ descriptions to the user, and the user browses the services descriptions to see whether there are services (s)he wants. If there is at least one service (s)he wants, then the process of service discovery is over; otherwise, go to (5). Select randomly an initial link of the local peer, then clone a working agent and dispatch it with the service query to the peer at the other end of the selected initial link. At that remote peer, do the searching as at the local peer. Clone a working agent and dispatch it with the service query to the local peer’s neighbor. At the neighbor peer, do the searching as at the local peer. At the remote peer, once there are services matching the query, then return the matching services’ descriptions to the user, who decides whether the returned results contain the target service. If the target service is found, then the search task is over and the working agent would return the source peer or be destroyed at the remote peer. If no target service is found, the working agent has to continue the search target until the target service is found or the working agent’s TTL is 0.
Note in the above process, when the working agent gets to a peer along a neighbor link, its TTL will not decrease; only walking along initial link, its TTL will decrease.
B
Building Web Services in P2P Networks
RELATED WORK
REFERENCES
Recently, combining P2P and Web services is gaining importance both in industry and academia. From the industry, two ambitious projects were launched, Sun Microsystems’ JXTA (Li, 2001) and Microsoft’s .net, more recently Hailstorm. Both JXTA and Hailstorm are trying to provide a general, language/environment-independent P2P services environment by putting forward a set of protocols for communication among peers. From research institutions, Hoschek (2002) proposed a unified peer-to-peer database framework for scalable service discovery; Schlosser Sinteck, Decker, and Nejdl (2002) put forward a scalable and ontology-based infrastructure for semantic Web services; Sheng, Benatallah, Dumas, and Mak (2002) developed a platform for rapid composition of Web services in peer-to-peer environment; and Abiteboul, Benjelloun, Manolescu, Milo, and Weber (2002) designed a kind of active XML document to integrate peer-to-peer data and Web services. Unlike the projects above, BP-Services is based on the BestPeer platform. We use an information retrieval method for services discovery, which is quite different from other P2P services projects. BP-Services is easy to implement because, except for the ordinary Web services protocols, it does not need any additional and complex protocols.
Abiteboul, S., Benjelloun, O., Manolescu, I., Milo, T., & Weber, R. (2002). Active XML: Peer-to-peer data and Web services integration. Proceedings of the 28th International Conference on Very Large Databases (pp. 1087-1090), Hong Kong, China.
CONCLUSION To overcome the limitations of Web services systems caused by their centralized architecture, we explore the techniques of building Web services applications under a P2P environment. The ongoing project, BP-Services, is presented as a case study to demonstrate our approach. BP-Services is an experimental Web services platform developed on the propriety BestPeer infrastructure. Future work will focus on developing some concrete applications on BP-Services put on a campus network as a test-bed for future research on P2P and Web services. And semantic Web service will also be considered in BP-Services in the future.
ACKNOWLEDGMENTS This work was supported by grants numbered 60573183 and 60373019 from the NSFC, grant No. 20045006071-16 from the Chenguang Program of Wuhan Municipality, grant No. WKL(04)0303 from the Open Researches Fund Program of LIESMARS, and the Shuguang Scholar Program of Shanghai Education Development Foundation.
Baesa-Yates, R., & Ribeiro-Neto, B. (1999). Modern information retrieval (pp. 124-127). Boston: Addison-Wesley/ACM Press. Christensen, E., Curbera, F., & Meredith, G. (2001). Web Services Description Language (WSDL) 1.1. W3C Note 15. Retrieved from http://www.w3.org/TR/wsdl Gottschalk, K., Graham, S., Kreger, H., & Snell, J. (2002). Introduction to Web services architecture. IBM Systems Journal, 41(2), 170-177. Hoschek, W. (2002). A unified peer-to-peer database framework and its application for scalable service discovery. Proceedings of the 3rd International IEEE/ACM Workshop on Gird Computing, Baltimore, MD (pp. 126-144). Li, G. (2001). JXTA: A network programming environment. IEEE Internet Computing, (May-June), 88-95. Ng, W. S., Ooi, B. C., & Tan, K.L. (2002). BestPeer: A self-configurable peer-to-peer system. Proceedings of the 18th International Conference on Data Engineering (pp. 272-272). Ooi, B. C., Tan, K-L., Lu, H., & Zhou, A. (2002). P2P: Harnessing and riding on peers. Proceedings of National Database Conference, Zhengzhou, China, (pp. 1-5). Schlosser, M., Sinteck, M., Decker, S., & Nejdl, W. (2002). A scalable and ontology-based infrastructure for semantic Web services. Proceedings of the 2nd International Workshop on Agents and Peer-to-Peer Computing, Linköping, Sweden, (pp. 104-111). Sheng, Q., Benatallah, B., Dumas, M., & Mak, E. (2002). SELF-SERV: A platform for rapid composition of Web services in a peer-to-peer environment. Proceedings of the 28th International Conference on Very Large Databases, Hong Kong, China, (pp. 1051-1054). SOAP. (2000). Simple Object Access Protocol (SOAP) 1.1. W3C Note 8. Retrieved from http://www.w3.org/TR/soap Yang, B., & Garcia-Molina, H. (2001). Comparing hybrid peer-to-peer systems. Proceedings of the 27th International Conference on Very Large Databases, Roma, Italy, (pp. 561-570).
Building Web Services in P2P Networks
Web Services Conceptual Architecture. (n.d.). Retrieved from http://www.ibm.com/software/solutions/webservices/ documentation.html
KEY TERMS Centralized P2P: In a centralized P2P network, there is a central server responsible for maintaining indexes on the metadata for all the peers in the network. Hybrid P2P: A kind of tradeoff between centralized P2P and pure P2P, which is structured hierarchically with a supernode layer and a normal peers layer. Peer-to-Peer (P2P): A class of systems and applications that employ distributed resources to perform a critical function in a decentralized way. A P2P distributed system typically consists of a large number of nodes that can share resources, information, and services, taking the roles of both consumer and provider, and may join or depart the network at any time, resulting in a truly dynamic and ad-hoc environment. Pure P2P: A P2P system with fully autonomous peersthat is, all nodes are equal, no functionality is centralized, and the communication between peers is also symmetric.
Service Discovery: An operation of finding Web services. After Web services are created and published in Web services registries such as UDDI, the service users or consumers need to search Web services manually or automatically. The implementation of UDDI servers should provide simple search APIs or Web-based GUI to help find Web services. Universal Description, Discovery and Integration (UDDI): The protocol for Web service publishing. It should enable applications to look up Web services information in order to determine whether to use them. Web Service: Can be seen as an interface that describes a collection of operations that are network accessible through standardized XML messaging. Software applications written in various programming languages and running on various platforms can use Web services to exchange data over computer networks due to the interoperability of using open standards. Web Services Description Language (WSDL): An XML language for describing Web services. eXtensible Markup Language (XML): A meta-language written in SGML that allows one to design a markup language, used to allow for the easy interchange of documents on l is misd Wide Web.
B
0 Category: M-Business and M-Commerce
Business and Technology Issues in Wireless Networking David Wright University of Ottawa, Canada
INTRODUCTION A major development in the enabling technologies for mobile computing and commerce is the evolution of wireless communications standards from the IEEE 802 series on local and metropolitan area networks. The rapid market growth and successful applications of 802.11, WiFi, is likely to be followed by similar commercial profitability of the emerging standards, 802.16e, WiMAX, and 802.20, WiMobile, both for network operators and users. This article describes the capabilities of these three standards and provides a comparative evaluation of features that impact their applicability to mobile computing and commerce. In particular, comparisons include the range, data rate in Mbps and ground speed in Km/h plus the availability of quality of service for voice and multimedia applications.
802.11 WiFi WiFi (IEEE, 1999a, 1999b, 1999c, 2003) was originally designed as a wireless equivalent of the wired local area network standard IEEE802.3, Ethernet. In fact there are many differences between the two technologies, but the packet formats are sufficiently similar that WiFi packets can easily be converted to and from Ethernet packets. Access points can therefore be connected using Ethernet and can communicate with end stations using WiFi. WiFi can transport both real-time communications such as voice and video plus non-real time communications such as Web browsing, by providing quality of service, QoS, using 802.11e (IEEE, 2005). There are 2 QoS options. One provides four priority levels allowing real-time traffic to be transmitted ahead of non-real-time traffic, but with no guarantee as to the exact delay experienced by the real-time traffic. The other allows the user to request a specific amount of delay, for example, 10 msecs., which may then be guaranteed by the access point. This is suited to delay sensitive applications such as telephony and audio/video streaming. WiFi has a limited range of up to 100 metres, depending on the number of walls and other obstacles that could absorb or reflect the signal. It therefore requires only low powered transmitters, and hence meets the requirements of operating in unlicensed radio spectrum at 2.4 and 5 GHz in
North America and other unlicensed bands as available in other countries. WiFi is deployed in residences, enterprises and public areas such as airports and restaurants, which contain many obstacles such as furniture and walls, so that a direct line of sight between end-station and access point is not always possible, and certainly cannot be guaranteed when end stations are mobile. For this reason the technology is designed so that the receiver can accept multipath signals that have been reflected and/or diffracted between transmitter and receiver as shown in Figure 1(a). WiFi uses two technologies that operate well in this multipath environment: DSSS, Direct Sequence Spread Spectrum, which is used in 802.11b, and OFDM, Orthogonal Frequency Division Multiplexing, which is used in 802.11a and g (Gast, 2002). A key distinguishing factor between these alternatives, which is important to users, is spectral efficiency, that is, the data rate that can be achieved given the limited amount of wireless spectrum available in the unlicensed bands. DSSS as implemented in 802.11b uses 22 MHz wireless channels and achieves 11 Mbps, that is, a spectral efficiency of 11/22 = 0.5. OFDM achieves a higher spectral efficiency and is therefore making more effective use of the available wireless spectrum. 802.11g has 22 MHz channels and delivers 54 Mbps, for a spectral efficiency of 54/22 = 2.5 and 802.11a delivers 54 Mbps in 20 MHz channels, with a spectral efficiency of 54/20 = 2.7. A recent development in WiFi is 802.11n (IEEE, 2006a), which uses OFDM in combination with MultiInput, MultiOutput, MIMO, antennas as shown in Figure 1(b). MIMO allows the spectral efficiency to be increased further by exploiting the multipath environment to send several streams of data between the multiple antennas at the transmitter and receiver. At the time of writing the details of 802.11n are not finalized, but a 4x4 MIMO system (with 4 transmit and 4 receive antennas) will probably generate about 500 Mbps in a 40 MHz channel, that is, a spectral efficiency of 500/40 = 12.5. 802.11n is suited to streaming high definition video and can also support a large number of users per access point. The data rates in WiFi are shared among all users of a channel, however some users can obtain higher data rates than others. Network operators may choose to police the data rate of individual users and possibly charge more for higher rates, or they may let users compete so that their data rates vary dynamically according to their needs and the priority levels
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Business and Technology Issues in Wireless Networking
Figure 1. (a) Receiver recovers a single signal from multiple incoming signals; (b) MIMO receiver recovers multiple signals using multiple antennas
Multipath Environment Wall Reflection Diffraction Post
(a)
Wall Reflection Diffraction Post
(b)
of their traffic. This provides considerable flexibility allowing many users to spend much of their time with low data rate applications such as VoIP, e-mail and Web browsing, with occasional high data rate bursts for audio/video downloads and data-intensive mesh computing applications. Many deployments of WiFi use multiple access points to achieve greater coverage than the range of a single access point. When the coverage of multiple access points overlaps they should use different radio channels so as not to interfere with each other, as shown in Figure 2. For instance, in the North American 2.4 GHz band there is 79 MHz of spectrum available and the channels of 802.11b and g are 22 MHz wide. It is therefore possible to fit 3 non-overlapping channels into the available 79 MHz, which are known as channels 1, 6 and 11. Other intermediate channels are possible, but overlap with channels 1, 6 and 11. In Figure 2, the top three access points are shown connected by Ethernet implying that they are under the control of a single network operator, such as an airport. As an end-station moves among these access points the connection is handed off from one access point to another using 802.11r (IEEE, 2006b), while maintaining an existing TCP/IP session. Movement can be up to automobile speeds using 802.11p (IEEE, 2006c). Standard technology, 802.21 (IEEE, 2006d), is also available to handoff a TCP/IP session when a mobile end-station moves from an access point of one network operator to that of another, and this requires a business agreement between the two operators. 802.11 networks can therefore span extensive areas by interconnecting multiple access points, and city-wide WiFi networks are available in, for example, Philadelphia in the U.S., Adelaide in Australia, Fredericton in Canada
and Pune in India. The broad coverage possible in this way greatly expands the usefulness of WiFi for mobile computing and electronic commerce. Enterprise users can set up secure virtual private networks from laptops to databases and maintain those connections while moving from desk to conference room to taxi to airport. A VoIP call over WiFi can start in a restaurant, continue in a taxi and after arriving at a residence. The features of WiFi, IEEE 802.11, that are of particular importance for mobile computing and commerce are: • • • •
Broad coverage achieved by handing off calls between access points, using 802.11r and 802.21, in cities where there are sufficient access points. Multimedia capability achieved by QoS, 802.11e. Flexibility in data rates achieved by allowing the total data rate of an access point to be shared in dynamically changing proportions among all users. Low cost achieved by using unlicensed spectrum, low power transmitters and mass produced equipment.
The downside to WiFi, IEEE 802.11, is limited coverage in cities that do not have extensive access point deployment.
802.16E WIMAX 802.16E (IEEE, 2006e) has a greater range than 802.11, typically 2-4 km and operates between base stations and subscriber stations. The initial IEEE standard 802.16 is for fixed applications, which compete with DSL and cable
B
Business and Technology Issues in Wireless Networking
Figure 2. WiFi handoff among access points
Channel 6
Internet Ethernet Switch
Channel 1
VoIP session maintained using 802.11r
Channel 11 Channel 6
VoIP session maintained using 802.21
modems. Mobile applications including handoff capability among base stations, which we deal with here, are provided by 802.16E, and are based on similar but incompatible technology. In 802.16E, WiMAX, mobility is limited to automobile speeds, up to about 100 Km/h so that it has limited use in high speed trains and aircraft. WiMAX uses the terminology “subscriber” stations, implying that customers are paying for a public service. Since the geographic range extends well into public areas, this is certainly one application. Another mobile application is a private campus network in which a central base station serves a business park or university campus. Initial deployment of WiMAX uses licensed spectrum, although low power applications in unlicensed spectrum are also specified in the standard. WiMAX has sophisticated QoS capabilities, which allow customers to reserve capacity on the network including a reserved data rate plus quality of service. The data rate is specified by a minimum reserved traffic rate, MRTR, on which quality of service is guaranteed (Figure 3). The customer is allowed to send at a higher rate, up to a maximum sustainable traffic rate, MSTR, without necessarily receiving QoS, and above that rate, traffic will be policed by the network operator, that is, it may be discarded. The QoS parameters that can be specified by the customer are latency and jitter, plus a priority level, which is used by the base station to distinguish among service flows that have the same latency and jitter requirements. The combination of latency and jitter can be used to distinguish among service flows, and further detail on the performance of WiMAX is given by Ghosh et al. (2005).
Combinations of QoS parameters and data rates make WiMAX highly suited to mobile computing and commerce. Each subscriber can set up multiple service flows, for example, for Web browsing during a multimedia conference, and use data rates that are quite different from those of other customers. The service provider can charge based on a combination of data rate and QoS. WiMAX is based on OFDM, thus achieving a high spectral efficiency. There are a number of options within 802.16E for the channel widths and modulation techniques, resulting in a corresponding range of data rates and spectral efficiencies. It is important to recognize that the spectral efficiency depends on the distance between the base station and the subscriber station (Figure 4). As the signal degrades with distance it is not possible to encode so many bps within each Hz and 802.16E assigns encodings that take this into account. Closer to the base station the data rate is therefore higher. The exact distance depends on the operating environment
Figure 3. WiMAX traffic rate guarantees bps MSTR MRTR
Traffic Policing, e.g., Discard of Excess Traffic No QoS Guarantee QoS Guaranteed
Time
Business and Technology Issues in Wireless Networking
Figure 4. Spectral efficiency and maximum data rates for WiMAX
•
Flexibility in data rates achieved by allowing the total data rate of a base station to be shared in dynamically changing proportions among all users.
The downside to 802.16E is the cost of licensed spectrum.
802.20 WIMOBILE At the time of writing, (1Q06), the specification of 802.20, (IEEE, 2006, f), is under development, so that less detail is available than for 802.11 and 802.16e. The key features of 802.20 are:
Spectral Efficiency = 4.5 Data Rate = 90 Mbps Spectral Efficiency = 3 Data Rate = 60 Mbps
• •
Spectral Efficiency = 1.5 Data Rate = 30 Mbps
• since 802.16E uses multipath signals involving reflections and diffractions. The data rates shown in Figure 4 are the maximum achievable with the highest channel bandwidth allowed according to the standard—20 MHz—and can vary not only with distance but also according to how much forward error correction is used. The features of 802.16E that are of particular importance for mobile computing and commerce are: • •
Good range, enabling city-wide coverage with a reasonable number of base stations. Multimedia capability achieved by QoS, and guaranteed data rates.
•
It operates in licensed spectrum below 3.5 GHz. It is designed from the start for an all-IP environment and interfaces to IP DiffServ QoS service classes, (Grossman, 2002) which provide for prioritization of users’ traffic. It interfaces to “Mobile IP” (Montenegro, 2001) as part of its mobility capability. Mobility includes not just automobile speed, but also high speed trains at up to 250 Km/h. It uses OFDM with MIMO antennas to achieve a very high spectral efficiency, so that large numbers of users can share access to a single base station.
COMPARATIvE EvALUATION Mobile computing and commerce involves communicating from mobile devices for a variety of purposes including: data transfer for processing intensive applications and for Web browsing; voice and multimedia calls between human users;
Table 1. Comparative evaluation of technologies for mobile computing and commerce 802.11, WiFi
802.16e, WiMAX
802.20, WiMobile
Range
100 metres
2-4 Km
2-4 Km
Coverage
Hot spots. Some city-wide deployments.
Designed for city-wide deployment
Designed for national deployment
Data Rate
11, 54, 500 Mbps flexibly shared among all users
Up to 90 Mbps flexibly shared among all users
> 1 Mbps per user
QoS
(a) Prioritization mechanism (b) data rate and QoS guarantees
Data rate and QoS guarantees
Data rate guarantees and QoS prioritization
Mobility Speed
100 Km/h
100 Km/h
250 Km/h
Cost
Very low unit cost access points. End-station interfaces built into phones, laptops, PDAs. Large number of access points required. Unlicensed spectrum.
Medium unit cost access points. End-station interfaces built into phones, laptops, PDAs. Licensed or unlicensed spectrum.
Medium unit cost access points. End-station interfaces built into phones, laptops, PDAs. Licensed spectrum.
B
Business and Technology Issues in Wireless Networking
downloading audio, video and multimedia from a server, (a) streaming for real-time playout to human users and (b) file transfer for subsequent access on the mobile device. Each of these requires appropriate data rate and quality of service. Cost is also an important factor, since subscription may be required to a public network operator or an enterprise may need to build its own wireless network. Employees using mobile computing devices within a building require mobility only at pedestrian speeds. In public areas such as city streets, automobile speeds are required and between cities high speed trains may be used. The type of mobile computing application determines which speed is appropriate. Table 1 provides a comparison among the three technologies described in this paper.
CONCLUSION Mobile computing and commerce users have a wide range of emerging wireless communication technologies available: WiFi, WiMAX and WiMobile. Each of them offers high data rates and spectral efficiencies, and will therefore likely be available at low cost. They are the major enabling telecommunication technologies for mobile computing and are likely to be deployed in public areas and private campuses for in-building and outdoor use. WiFi is already extensively deployed and WiMAX is being deployed in Korea in 2006 and can be expected in many other countries in 2007. The WiMobile standard has not yet been specified (as of the time of writing 1Q06) and commercial equipment can be expected after WiMAX.
REFERENCES Gast, M. (2002). 802.11 wireless networks: The definitive guide. O’Reilly. Ghosh, A., Wolter, D. R., Andrews, J. G., & Chen, R. (2005). Broadband wireless access with WiMax/8O2.16: Current performance benchmarks and future potential. IEEE Communications Magazine, 43(2), 129-136. Grossman, D. (2002). New terminology and clarifications for Diffserv. RFC3260. Internet Engineering Task Force. IEEE. (1999a). 802.11 wireless LAN: Medium access control (MAC) and physical layer (PHY) specifications. New York: IEEE Publications. IEEE. (1999b). 802.11a high-speed physical layer in the 5 GHz band. New York: IEEE Publications. IEEE. (1999c). 802.11b higher-speed physical layer (PHY) extension in the 2.4 GHz band. New York: IEEE Publications.
IEEE. (2003). 802.11g further higher-speed physical layer extension in the 2.4 GHz band. New York: IEEE Publications. IEEE. (2005). 802.11e wireless LAN: Quality of service enhancements. New York: IEEE Publications. IEEE. (2006a). 802.11n wireless LAN: Enhancements for higher throughput (In progress). Retrieved March 2006, from http://standards.ieee.org/board/nes/projects/802-11n.pdf. IEEE. (2006 b). 802.11r wireless LAN: Fast BSS transition (In progress). Retrieved March 2006, http://standards.ieee. org/board/nes/projects/802-11n.pdf. IEEE. (2006c). 802.11p wireless LAN: Wireless access in vehicular environments. (In progress). Retrieved March 2006, from http://standards.ieee.org/board/nes/projects/80211p.pdf. IEEE. (2006d). 802.21 media independent handover services. (In progress). Retrieved March 2006, from http://grouper. ieee.org/groups/802/21/. IEEE. (2006e). 802.16E-2005 air interface for fixed and mobile broadband wireless access systems: Amendment for physical and medium access control layers for combined fixed and mobile operation in licensed bands. New York: IEEE Publications. IEEE. (2006f). 802.20 mobile broadband wireless access systems. (In progress). Retrieved March 2006, from http:// grouper.ieee.org/groups/802/20/. Montenegro, G. (2001) Reverse tunneling for mobile IP. RFC3024. Internet Engineering Task Force.
KEY TERMS Direct Sequence Spread Spectrum (DSS): A transmission technique in which data bits are multiplied by a higher frequency code sequence, so that the data are spread over a wide range of frequencies. If some of these frequencies fade, the data can be recovered from the data on the other frequencies together with a forward error correction code. Mobile IP: An Internet standard that allows a mobile user to move from one point of attachment to the network to another while maintaining an existing TCP/IP session. Incoming packet to the user are forwarded to the new point of attachment. Multipath: A radio environment in which signals between transmitter and receiver take several different spatial paths due to reflections and diffractions. Orthogonal Frequency Division Multiplexing (OFDM): A transmission technique in which data bits are
Business and Technology Issues in Wireless Networking
transmitted on different frequencies. The data transmitted on one frequency can be distinguished from those on other frequencies since each frequency is orthogonal to the others.
WiFi: A commercial implementation of the IEEE 802.11 standard in which the equipment has been certified by the WiFi Alliance, an industry consortium.
Quality of Service (QoS): Features related to a communication, such as delay, variability of delay, bit error rate and packet loss rate. Additional parameters may also be included, for example, peak data rate, average data rate, percentage of time that the service is available, mean time to repair faults and how the customer is compensated if QoS guarantees are not met by a service provider.
WiMAX: A commercial implementation of the IEEE 802.16 standard in which the equipment has been certified by the WiMAX Forum, an industry consortium. WiMobile: Another name for the IEEE 802.20 standard which is in course of development at the time of writing (1Q06).
B
Category: M-Business and M-Commerce
Business Strategies for Mobile Marketing Indranil Bose University of Hong Kong, Hong Kong Chen Xi University of Hong Kong, Hong Kong
INTRODUCTION With the appearance of advanced and mature wireless and mobile technologies, more and more people are embracing mobile “things” as part of their everyday lives. New business opportunities are emerging with the birth of a new type of commerce known as mobile commerce or m-commerce. M-commerce is an extension to electronic commerce (ecommerce) with new capabilities. As a result, marketing activities in m-commerce are different from traditional commerce and e-commerce. This chapter will discuss marketing strategies for m-commerce. First we will give some background knowledge about m-commerce. Then we will discuss the pull, push, and viral models in m-marketing. The third part will be the discussion about the future developments in mobile marketing. The last part will provide a summary of this article.
BACKGROUND Popularity of Mobile Services From the research done by Gartner Dataquest (BusinessWeek, 2005), there will be more than 1.4 billion mobile service subscribers in the Asia-Pacific region by 2009. Research analysts of Gartner Dataquest also estimated that China will have over 500,000 subscribers, and more than 39% of the people will use mobile phones at that time. In India, the penetration rate of mobile phones is expected to increase from 7% in 2005 to 28% in 2008. The Yankee Group has also reported a growing trend of mobile service revenues from 2003 to 2009. Although the revenue generated by traditional text-based messaging service will not change much, revenue from multimedia messaging services will rise to a great extent. Other applications of mobile services, such as m-commerce-based services and mobile enterprise services, will continue to flourish. One thing that is very important in driving Asia-Pacific mobile service revenue is mobile entertainment services. Revenue from mobile entertainment services will make up almost half of the total revenues from all kinds of mobile data services from now on. Not only in the region of Asia-Pacific, but mobile services will increase
in popularity in other parts of the world as well. In the United States, it is expected that the market for m-commerce will reach US$25 billion in 2006.
The Development of Mobile Technologies Two terms are frequently used when people talk about mobile information transmission techniques: the second-generation (2G) and the third-generation (3G) wireless systems. These two terms actually refer to two generations of mobile telecommunication systems. Three basic 2G technologies are time division multiple access (TDMA), global system for mobile (GSM), and code division multiple access (CDMA). Among these three, GSM is the most widely accepted technology. There is also the two-and-a-half generation (2.5G) technology of mobile telecommunication, such as general packet radio service (GPRS). 2.5G is considered to be a transitional generation of technology between 2G and 3G. They have not replaced 2G systems. They are mostly used to provide additional value-added services to 2G systems. The future of mobile telecommunication network is believed to be 3G. Some standards in 3G include W-CDMA, TD-SCDMA, CDMA 2000 EV-DO, and CDMA EV-DV. The advancement in mobile telecommunication technology will bring in higher speed of data transmission.. The speed of GSM was only 9.6 kilobits per second (kbps), while the speed of GPRS can reach from 56 to114 kbps. It is believed that the speed of 3G will be as fast as 2 Megabits per second (mbps). The acceptance of 3G in this world began in Japan. NTT DoCoMo introduced its 3G services in 2001. Korea soon followed the example of Japan. In 2003, the Hutchison Group launched 3G commercially in Italy and the UK, and branded its services as ‘3’. ‘3’ was later introduced in Hong Kong, China in 2004. Mainland China is also planning to implement 3G systems. Some prototypes or experimental networks have been set up in the Guangdong province. It is expected that 3G networks will be put into commercial use in 2007 using the TD-SCDMA standard that has been indigenously developed in China. Mobile information transmission can also be done using other technical solutions such as wireless local area network (WLAN) and Bluetooth. The interested reader may refer to Holma and Toskala (2002) for a fuller
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Business Strategies for Mobile Marketing
description of 3G systems, and to Halonen, Romero, and Melero (2003) for details of 2G and 2.5G systems. The most popular mobile devices currently in use include mobile phones, wireless-enabled personal digital assistants (PDAs), and wireless-enabled laptops (Tarasewich, Nickerson, & Warkentin, 2002). Smartphones are also gaining favor from customers. Mobile phones are the most pervasive mobile devices. Basically, mobile phones can make phone calls, and can send and receive short text messages. More advanced mobile phones have color screens so that they can send or receive multimedia messages, or have integrated GPRS modules so that they can connect to the Internet for data transmission. PDAs are pocket-size or palm-size devices which do limited personal data processing such as recording of telephone numbers, appointments, and notes on the go. Wireless-enabled PDAs have integrated Wi-Fi (wireless fidelity)—which is the connection standard for W-LAN or Bluetooth—which helps them access the Internet. Some PDAs can be extended with GPRS or GSM modules so that they can work as a mobile phone. PDAs nowadays usually have larger screens than that of mobile phones and with higher resolution. They are often equipped with powerful CPUs and large storage components so that they can handle multimedia tasks easily. Smartphones are the combination of mobile phones and PDAs. Smartphones have more complete phoning function than PDAs, while PDAs have more powerful data processing abilities. However, the boundary between smartphones and PDAs are actually becoming more and more fuzzy.
The Need for Mobile Marketing The rapid penetration rate of mobile devices, the huge amounts of investment from industries, and the advancement of mobile technologies, all make it feasible to do marketing via mobile devices. Mobile commerce refers to a category of business applications that derive their profit from business opportunities created by mobile technologies. Mobile marketing, as a branch of m-commerce (Choon, Hyung, & Kim, 2004; Varshney & Vetter, 2002), refers to any marketing activities conducted via mobile technologies. Usually mcommerce is regarded as a subset of e-commerce (Coursaris & Hassanein, 2002; Kwon & Sadeh, 2004). That is true, but due to the characteristics of mobile technologies, mobile marketing is different from other e-commerce activities. The first difference is caused by mobile technologies’ ability to reach people anywhere and anytime; therefore mobile marketing can take the advantage of contextual information (Zhang, 2003). Dey and Abowd (2001) defined context as “any information that characterizes a situation related to the interaction between users, applications, and the surrounding environment.” Time, location, and network conditions are three of the key elements of context. The second difference is caused by the characteristics of mobile devices. Mobile
devices have limited display abilities. The screens are usually small, and some of the devices cannot display color pictures or animations. On the other hand, mobile devices have various kinds of screen shapes, sizes, and resolutions. Thus, delivering appropriate content to specific devices is very important. Mobile devices also have limited input abilities, and this makes it difficult for customers to respond. Mobile marketing shares something in common with e-commerce activities. An important aspect of e-commerce is to deliver personalized products/services to customers. Mobile marketing inherits this feature. Mobile marketing also inherits some of the problems from e-commerce, especially the problem of spamming. Personalization in mobile marketing is to conduct marketing campaigns which can meet the customer’s needs by providing authorized, timely, location-sensitive, and device-adaptive advertising and promotion information (Scharl, Dickinger, & Murphy, 2005).
MOBILE MARKETING Benefits of Mobile Marketing There are two main approaches to advertise and promote products in industry—mass marketing and direct marketing. The former uses mass media to broadcast product-related information to customers without discrimination, whereas the latter is quite different in this regard. Mobile marketing takes a direct marketing approach. Using mobile marketing, marketers can reach customers directly and immediately. Similarly, customers can also respond to marketers rapidly. This benefit makes the interaction between marketers and customers easy and frequent. Compared to direct marketing using mail or catalogs, mobile marketing is comparatively cost effective and quick. Compared to telephone direct marketing, mobile marketing can be less interruptive. Compared to e-mail direct marketing, mobile marketing can reach people anytime and anywhere, and does not require customers to sit in front of a computer. Therefore, to some extent, mobile marketing can be a replacement for other types of marketing channels such as mail, telephone, or e-mail. Advertisement or promotion information sent via the Internet can be sent via a mobile device. Mobile marketing can enhance marketing by adding new abilities like time-sensitive and location-sensitive information. On the other hand, mobile commerce can generate new customers’ data, like mobile telecommunication usage data and mobile Internet surfing data. Mobile marketing is the first choice for conducting marketing activities for m-commerce applications. However, due to limited size of screens of mobile devices, only brief information can be provided in mobile marketing solicitations, while e-mail or mail marketing can provide very detailed information. On the other hand telephone marketing requires the good communication skill of telesales. Once telemarketers have
B
Business Strategies for Mobile Marketing
acquired this skill, the interaction between marketers and customers is quicker and more effective. It is not clear if mobile marketing is as effective or as popular as mass marketing agents like television and newspaper, but it can be said that it is indeed a powerful medium that is likely to gain in popularity in the future.
Models for Mobile Marketing Mobile marketing usually follows one of the three kinds of models—push, pull (Haig, 2002; Zhang, 2003), and viral (Ahonen, 2002; Ahonen, Kasper, & Melkko, 2004; Haig, 2002), as illustrated in Figure 1.
Push Model The push model sends marketing information to customers without the request of the customer. If the push model is used, besides knowing the targeted customers’ interests, understanding the context of customers at the time the marketing activities are to be carried out is very critical. The timing in sending mobile information should also be appropriate. The content delivered to customers should also be displayable on their mobile devices. Permissions from customers are necessary before any solicitations can be sent. In the push model, marketers like to make it easier for customers to respond because of the poor input ability of mobile devices. G2000, a Hong Kong clothing chain, launched a mobile marketing campaign in November 2004. Mobile coupons in the format of SMS were sent to the mobile phones of selected customers. Customers could then use these coupons stored in their mobile phones when purchasing items in designated G2000 stores in order to get discounts. The campaign was considered to be a success because a number of customers responded to this program and used mobile coupons at G2000 stores.
Figure 1. Push, pull and viral models for mobile marketing Mobile marketers
Push
Push Pull
Mobile device
Mobile device
Viral Viral
98
Mobile device
Pull Model In the pull model, the marketer waits for the customer to send a request for a solicitation. The marketer prepares the marketing information in a format that is displayable in all possible mobile devices and scalable for various connection speeds. The significance of the pull model is that the information from customers is very useful for understanding customers’ preferences, such as the preferred marketing time and interests. Mobile marketing following a pull model can be conducted in many ways. One possible approach is to let the customers select and download coupons to their mobile devices. Mobile service providers can build a Web site using a mobile Internet protocol such as wireless application protocol (WAP), and place various text-based coupons on this Web site. Customers can use their GPRS-enabled phone to browse the Web site and download coupons they like in the format of SMS. Each coupon will have a unique identity number. When the coupon is redeemed, related information, such as the phone number of the customer who downloaded the coupon and the time it was redeemed, is recorded. This kind of information is later used to analyze the behaviors of customers and build a profile for the customer. China Mobile, a mobile telecommunication operator in China, had established such a Web site for customers in Xiamen (a city located in southeast China) in 2005. The customers of China Mobile could download coupons displayed on this Web site onto their mobile phones using text messaging.
Viral Model The phrase viral marketing was created by Steve Jurvetson in 1997 to describe the burgeoning use of Hotmail (Jurvetson & Draper, 1997). The principle of the viral model is based on the fact that customers forward information about products/services to other customers. The viral model enlarges the effect of other marketing activities while it costs the marketers very little in monetary terms. The viral model enables customer-to-customer communication. Like the pull model, the format of information that is delivered by viral model should be displayable in different devices and scalable for different connection speeds. Actually mobile marketing has the ability to be viral inherently because it is quite easy for people to forward mobile advertising or promotion information to their friends. However, viral marketing information has to be interesting and attractive enough to make the customers willing to forward it to other people. For example, “reply to this message in order to win $5000” may be a very attractive viral marketing message. Usually, viral marketing begins with push marketing activities to customers. According to Linner (2003), when the movie “2 Fast 2 Furious” was running in movie theatres, marketers tried to create a viral promotion using a mobile marketing strategy. Fans were asked to send SMS to enter a certain film-related
Business Strategies for Mobile Marketing
competition. Besides inviting fans on every major phone network through advertisements on television, newspapers, and also through posters, a special code was designed and a low fee was offered to customers in order to encourage them to forward promotion information to their friends. Exciting gifts were offered as prizes in this competition (such as a replica of the vehicle, the EvO VII, that was used in the movie) to spur the enthusiasm of customers. These three models of direct marketing can be complimentary to each other. Push-based mobile marketing can be used to stimulate pull-based marketing activities. For example, book marketers can send a short introduction to customers via SMS with a remark at the end saying “for more details, please reply to XXXXX.” Once a customer responds to this by replying using SMS, more promotion or advertising information on this book can be sent to him. All three models—push, pull, and viral—can even be integrated together in a mobile marketing campaign. An example of an integrated mobile marketing approach was adopted by Fox Txt Club for the movie “phone booth” (Linner, 2003). At the beginning of the marketing campaign, members of the Fox Txt Club were sent invitations via SMS to a preview. The aim of this was to pull customers to the campaign. A competition that invited people to send SMS about questions pertaining to various details in the film was set up. The forwarding of SMS about the movie and the competition among club members and their friends, together with other media such as entertainment and event listings magazines and city-center posters, made the marketing campaign viral. The details of those who responded were recorded by Fox Txt Club, and this helped in building the database of customers for future release promotions. This could be used for push marketing for another movie in the future.
Strategies for Mobile Marketing The most fundamental task for marketing activities following the push model is to send advertising or promotion information about products or services that the targeted customers once bought. This is the most direct and easy way to decide what is to be offered to customers in a solicitation. However, just marketing products already existing in customers’ transaction records is not enough for marketers. It is necessary for marketers to explore the needs of customers. Two of the most commonly used marketing strategies are cross-selling and up-selling. Cross-selling is the practice of suggesting similar products or services to a customer who is considering buying something, such as showing a list of ring tones on a mobile Internet Web page that are similar to the one a customer has downloaded. Up-selling is the practice of suggesting higher priced, better versions of products or services to a customer who is considering a purchase, such as a mobile phone plan with higher fees and additional features. Two approaches can be used to find opportunities
for up-selling or cross-selling. One is to find products or services that are similar to the ones a customer has bought. The other is to find people who have characteristics that are similar to a targeted customer. Products or services those people have bought and the targeted customer has not can be recommended to the target customers. Pull-based marketing is relatively passive compared to push-based marketing. Usually in pull marketing, customers are responsible for searching for useful advertising or promotion information. The marketers’ responsibility is to help customers find what they want more efficiently. Therefore, knowing what customers may request is very important in pull marketing. Instead of sending related information to customers like push marketing, marketers doing pull marketing can make information about products or services available on their mobile Internet Web site or ordinary Internet Web site. In viral marketing, marketers stand in a more passive position than even in pull marketing. However, for both pull and push marketing, some push activities should be carried out to start the marketing. Whatever model one may use when carrying out mobile marketing activities, one issue must always be kept in mind and that is the necessity of obtaining explicit permission from customers (Bayne, 2002). Mobile technology makes connections so direct that it can interfere with customers’ privacy very easily. Therefore, sending advertising or promotion information to people will cause trouble if permissions are not sought before solicitations or customers’ wishes about not receiving a solicitation are not respected.
Understanding Customers in Mobile Marketing All of the three models require good understanding of customers’ needs. Marketing information that is not well designed will be regarded as spam by customers. Once a customer identifies some information from a company as spam, he or she will pay very little attention to or simply discard any information from that company. If a customer cannot find useful information on the Web site a company provides, it may be ok for the first time, a pity for the second time, but for the third time it will mean business lost forever. If information sent to customers is not interesting, customers may not want to forward them to their friends. All these situations may lead to failure of a marketing campaign. To avoid these situations, marketers need to understand customers well enough in order to send personalized marketing information. Customer profiling is a necessary approach to understand customers better. Customer profiling aims to find factors that can characterize customers. These factors are found by comparing customers to each other in order to discover similarities and differences among customers. Customer profiling encompasses two tasks—customer clustering and customer behavior pattern recognition. Customer clustering
B
Business Strategies for Mobile Marketing
aims to classify customers into different groups. Customers within the same group are said to be more similar to each other than to customers in different groups. Marketers cluster customers using various data. Traditionally, customers are clustered according to their geographic locations, demographic characteristics, and the industries they are working for. They can also be clustered based on information about their purchasing history, such as what they bought, when they bought, and how much they spent. With the appearance of mobile services and m-commerce, usage data of new customer data services can also be used for clustering. For example, messaging services that customers subscribed to, GPRS surfing and download records, the type of mobile devices the customers use, and monthly mobile phone usage including use of IDD and roaming can yield many interesting information about the customers. Aside from these hard facts, marketers may also want to infer some soft knowledge about customers’ behaviors as well. To recognize customer behavior the marketers must discover relationships between hard facts. For example, customers that download ring tones of game music may download games-related screensavers later on. Since mobile technologies can enable context-sensitive marketing activities, marketers should gather knowledge about customers’ location preferences and time preferences. For example, when does a customer usually go shopping and which place does he/she visit on the shopping trips? Marketers can find this kind of soft knowledge from various mobile network usage data. Again, collecting information on location and time requires permission from customers. Based on customer profiling, more sophisticated personalized advertising or promotion information can be sent to customers.
FUTURE TRENDS Mobile technologies will advance further in the future. New technologies will enable new kinds of marketing activities. For example, the implementation of fourth-generation (4G) wireless systems will make the bandwidth much larger than that in current networks. On the other hand, the mobile device will have larger screens with higher resolutions. These two factors together will make interactive audio and even interactive video marketing possible. Generally speaking, the limitation of current mobile technologies will be weakened or removed in the future. As a result, more emphasis may be put on time- and location-related marketing, as well as on better understanding customers’ interests. The principle is not only to know what customers want, but also to know when and where they may have a certain kind of need. Data mining techniques can be used in the future to find customer behavior patterns with time and location factors. Data mining techniques have been used widely in direct marketing for targeting customers (Ling & Li, 1998). There are also data 00
mining techniques for clustering customers such as selforganizing-map (SOM—Kohonen, 1995) and techniques for discovering customer behavior such as association rules mining (Agrawal & Srikant, 1994). In the future, the availability of huge amounts of data about customers will compel marketers to adopt strong data mining tools to delve deep into customers’ nature.
CONCLUSION Equipped with advanced mobile technologies, more sophisticated marketing activities can be conducted now and in the future. In this article, we have discussed the benefits of mobile marketing, the role of mobile marketing in m-commerce, and the models used in mobile marketing. Although mobile marketing is powerful, it cannot replace other methods of marketing and should only be used as a powerful complement to traditional marketing. Mobile marketing should be integrated into the whole marketing strategy of a firm so that it can work seamlessly with other marketing approaches.
REFERENCES Agrawal, R., & Srikant, R. (1994). Fast algorithms for mining association rules. In Proceedings of the 20th International Conference on Very large Databases (pp. 487-499), Santiago, Chile. Ahonen, T. T. (2002). M-profits: Making money from 3G services. West Sussex, UK: John Wiley & Sons. Ahonen, T. T., Kasper, T., & Melkko, S. (2004). 3G marketing: Communities and strategic partnerships. West Sussex, UK: John Wiley & Sons. Bayne, K. M. (2002). Marketing without wires: Targeting promotions and advertising to mobile device users. New York: John Wiley & Sons. BusinessWeek. (2005). Special advertising section: 3G the mobile opportunity. BusinessWeek (Asian ed.), (November 21), 92-96. Choon, S. L., Hyung, S. S., & Kim, D. S. (2004). A classification of mobile business models and its applications. Industrial Management & Data Systems, 104(1), 78-87. Coursaris, C., & Hassanein, K. (2002). Understanding mcommerce. Quarterly Journal of Electronic Commerce, 3(3), 247-271. Dey, A. K., & Abowd, G. D. (2001). A conceptual framework and a toolkit for supporting the rapid prototyping of context-aware applications. Human-Computer Interaction, 16(2-4), 97-166.
Business Strategies for Mobile Marketing
Haig, M. (2002). Mobile marketing: The message revolution. London: Kogan Page. Halonen, T., Romero, J., & Melero, J. (2003). GSM, GPRS and EDGE performance: Evolution towards 3G/UMTS. West Sussex, UK: John Wiley & Sons. Holma, H., & Toskala, A. (2002). WCDMA for UMTS (2nd ed.). West Sussex, UK: John Wiley & Sons. Jurvetson, S., & Draper, T. (1997). Viral marketing. Retrieved from http://www.dfj.com/cgi-bin/artman/publish/ steve_tim_may97.shtml
Code Division Multiple Access (CDMA): A kind of 2G technology that allows users to share a channel by encoding data with channel-specified code and by making use of the constructive interference properties of the transmission medium. Enhanced Data rates for GSM Evolution (EDGE): A kind of 2.5G technology. A new modulation scheme is implemented in EDGE to enable transmission speed of up to 384 kbps within the existing GSM network.
Kohonen, T. (1995). Self-organizing maps. Berlin: SpringerVerlag.
General Packet Radio Service (GPRS): Belongs to the family of 2.5G. GPRS is the first implementation of packet switching technology within GSM. The speed of GPRS can reach up to 115 Kbps.
Kwon, O.B., & Sadeh, N. (2004). Applying case-based reasoning and multi-agent intelligent system to contextaware comparative shopping. Decision Support Systems, 37(2), 199-213.
Global System for Mobile (GSM) Communications: One of the 2G wireless mobile network technologies and the most widely used today. It can now operate in the 900 MHz, 1,800 MHz, and 1,900 MHz bands.
Ling, C. X., & Li, C.-H. (1998). Data mining for direct marketing: Problems and solutions. Proceedings of the 4th International Conference on Knowledge Discovery and Data Mining (pp. 73-79), New York.
3G: The third generation of mobile telecommunication technologies. 3G refers to the next generation of mobile networks which operate at frequencies as high as 2.1 GHz, or even higher. The transmission speeds of 3G mobile wireless networks are believed to be able to reach up to 2 Mbps.
Linner, J. (2003). Hitting the mark with text messaging. Retrieved from http://wireless.sys-con.com/read/41316.htm Scharl, A., Dickinger, A., & Murphy, J. (2005). Diffusion and success factors of mobile marketing. Electronic Commerce Research and Applications, 4, 159-173. Tarasewich, P., Nickerson, R.C., & Warkentin, M. (2002). Issues in mobile e-commerce. Communications of the Association for Information Systems, 8, 41-84. Varshney, U., & Vetter, R. (2002). Mobile commerce: Framework, applications and networking support. Mobile Networks and Applications, 7, 185-198. Zhang, D. (2003). Delivery of personalized and adaptive content to mobile devices: A framework and enabling technology. Communications of the Association for Information Systems, 12, 183-202.
KEY TERMS Bluetooth: Used mostly to connect personal devices wirelessly like PDAs, mobile phones, laptops, PCs, printers, and digital cameras.
Time Division Multiple Access (TDMA): Divides each network channel into different time slots in order to allow several users to share the channel. Time Division Synchronous Code Division Multiple Access (TD-SCDMA): A 3G mobile telecommunications standard developed in China. 2G: The second generation of mobile telecommunication technologies. It refers to mobile wireless networks and services that use digital technology. 2G wireless networks support data services. 2.5G: The second-and-a-half generation of mobile telecommunication technologies. 2.5G wireless system is built on top of a 2G network. 2.5G networks have the ability to conduct packet switching in addition to circuit switching. 2.5G supports higher transmission speeds compared to 2G systems. W-CDMA: Developed by NTT DoCoMo as the air interface for its 3G network called FOMA. It is now accepted as a part of the IMT-2000 family of 3G standards. Wireless Local Area Network (WLAN): Connects users wirelessly instead of using cables. WLAN is not a kind of mobile telecommunication technology. The coverage of WLAN may vary from a single meeting room to an entire building of a company.
0
B
0 Category: Mobile Software Engineering
Cache Invalidation in a Mobile Environment Say Ying Lim Monash University, Australia
INTRODUCTION
BACKGROUND
The rapid development, as well as recent advances in wireless network technologies, has led to the development of the concept of mobile computing. A mobile computing environment enables mobile users to query databases from their mobile devices over the wireless communication channels (Cai & Tan, 1999). The potential market for mobile computing applications is projected to increase over time by the currently increasingly mobile world, which enables a user to satisfy their needs by having the ability to access information anywhere, anytime. However, the typical nature of a mobile environment includes low bandwidth and low reliability of wireless channels, which causes frequent disconnection to the mobile users. Often, mobile devices are associated with low memory storage and low power computation and with a limited power supply (Myers & Beigl, 2003). Thus, for mobile computing to be widely deployed, it is important to cope with the current limitation of power conservation and low bandwidth of the wireless channel. These two issues create a great challenge for fellow researchers in the area of mobile computing. By introducing data caching into the mobile environment, it is believed to be a very useful and effective method in conserving bandwidth and power consumptions. This is because, when the data item is cached, the mobile user can avoid requests for the same data if the data are valid. And this would lead to reduced transmissions, which implies better utilization of the nature of the wireless channel of limited bandwidth. The cached data are able to support disconnected or intermitted connected operations as well. In addition, this also leads to cost reduction if the billing is per KB data transfer (Lai, Tari, & Bertok, 2003). Caching has emerged as a fundamental technique especially in distributed systems, as it not only helps reduce communication costs but also offloads shared database servers. Generally, caching in a mobile environment is complicated by the fact that the caches need to be kept consistent at all time. In this article, we describe the use of caching that allows coping with the characteristics of the mobile environment. We concentrate particularly on cache invalidation strategy, which is basically a type of caching strategy that is used to ensure that the data items that are cached in the mobile client are consistent in comparison to the ones that are stored on the server.
Caching at the mobile client helps in relieving the low bandwidth constraints imposed in the mobile environment (Kara & Edwards, 2003). Without the ability to cache data, there will be increased communication in the remote servers for data and this eventually leads to increased cost and, with the nature of an environment that is vulnerable to frequent disconnection, may also lead to higher costs (Leong & Si, 1997). However, the frequent disconnection and the mobility of clients complicate the issue of keeping the cache consistent with those that are stored in the servers (Chand, Joshi, & Misra, 2004). Thus, when caching is used, ensuring data consistency is an important issue that needs considerable attention at all times (Lao, Tari, & Bertok, 2003). This is because the data that has been cached may have been outdated and no longer valid in comparison to the data from the corresponding servers or broadcast channel. Figure 1 shows an illustration of a typical mobile environment that consists of mobile clients and servers, which are also know as mobile host (MH) and mobile support system (MSS) respectively. The mobile clients and servers communicate via a wireless channel within a certain coverage, known as cell (Chand, Joshi, & Misra, 2003; Cai & Tan, 1999). There are two approaches for sending a query in a mobile environment, which are: (a) The mobile clients are free to request data directly from the server via the wireless channel and the server will process and pass the desired data items back and (b) the mobile clients can tune into the broadcast channel to obtain the desired data items and download it to his/her mobile device. This can be illustrated in Figure 1a and Figure 1b respectively. The assumption is that updates are only able to occur at the server side and mobile clients can only have a read only feature.
CACHE INvALIDATION Due to the important issue in the mobile environment, which is the ability to maintain data consistency, cache invalidation strategy is of utmost significance to ensure that the data items cached in the mobile client are consistent with those that are stored on the server. In order to ensure that data that are about to be used is consistent, a client must validate its cache prior to using any data from it.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Cache Invalidation in a Mobile Environment
Figure 1. Mobile environment architecture
C Wireless Cell
Wireless Cell
C2
Processing
Processing
C3
Moving
WIRED NETWORK
Moving
Received Requested Items
Sending Query for Requesting Item C1
C6
Processing
C4
C5 Wireless Cell
There are several distinctive and significant benefits that cache invalidation brings to a mobile computing environment. If cache data are not validated to check for consistency, it will become useless and out-of-date. However, if one can utilize the cache data then the benefits it may bring include energy savings—that is, by reducing the amount of data transfer—and in return result in cost savings.
Using Cache Invalidation in a Mobile Environment This can be done by using the broadcasting concept in communicating cache validation information to mobile clients. The server broadcasts the cache information, which is known as cache invalidation report (IR), periodically on the air to help clients validate their cache to ensure they are still consistent and can be used. It appears that the broadcast mechanism is more appropriate for the mobile environment due to its characteristic of salability, which allows it to broadcast data to an arbitrary number of clients who can listen to the broadcast channel anytime (Lai, Tari, & Bertok, 2003). By using the broadcasting approach, whereby the server periodically broadcasts the IR to indicate the change data items, it eliminates the need to query directly to the server for a validation cache copies. The mobile clients would be able to listen to the broadcast channel on the IR and use them to validate their local cache respectively (Cao, 2002). Although cache invalidation strategy is important in a mobile environment, it will be vulnerable to disconnection
and the mobility of the clients. One of the main reasons that cause mobile clients frequent disconnection is the limited battery power, and that is why mobile clients often disconnect to conserve battery power. It may appear to be very expensive at times to validate the cache for clients that experience frequent disconnection, especially with narrow wireless links. Other drawbacks would include long query latency, which is associated with the need of the mobile client to listen to the channel for the next IR first before he is able to conclude whether the cache is valid or not before answering a query. Another major drawback is the unnecessary data items in the IR that the server keeps. This refers to data items that are not cached by any mobile clients. This is thereby wasting a significant amount of wireless bandwidth. Example 1: A mobile client in a shopping complex denoted as C1 in Figure 2 wanted to know which store to visit by obtaining a store directory. The client has previously visited this store and already has a copy of the result in his cache. In order to answer a query, the client will listen to the IR that are broadcasted and use it for validation against its local cache to see if it is valid or not. If there is a valid cached copy that can be used in answering the query, which is getting the store directories, then the result will be returned immediately. Otherwise, if the store directories have changed and now contain new shops, then the invalid caches have to be refreshed via sending a query to the server (Elmagarmid et al., 2003). The server would keep track of the recently updated data and broadcast the up-to-date IR every now and 0
Cache Invalidation in a Mobile Environment
Figure 2. Using cache invalidation in a mobile environment
…………. Broadcast
D1
(3) Validate against received IR with cache to determine if the data is valid or not
(5) If cache data not valid, tune into the broadcast channel to refresh and obtain valid data
(2) Retrieve IR for validation
(1) Listened to the IR for validation
D3
(6) New valid data is obtained
D2
(4) If cache data is valid, result returned imediately from local cache
then for the clients to tune in. This can be done either by sending a request directly to the server (pull-based system) or tune into the broadcast channel (push-based system). Figure 2 illustrates an example of a push-based system. In summary, effective cache invalidation strategies must be developed to ensure consistency between cached data in the mobile environment and the original data that are stored on the server (Hu & Lee, 1998).
Designing Cache Invalidation Strategies It is important to produce an effective cache invalidation strategy to maintain a high level of consistency between cached data in the mobile devices with those that are stored on the server. In general, there are three possible basic ways in designing the invalidation strategies that described as follows (Hu & Lee, 1998). Assuming the server is stateful, whereby it knows which data are cached and by which particular mobile clients. Whenever there are changes in the data item in the server, the server would send a message to those clients, which has cached that particular item that has been updated or changed. In this way, the server would be required to locate the mobile clients. However, there is a major limitation in this method, that is, particularly in cases of disconnection. This is because mobile clients that are disconnected cannot be contacted by the server and thus its cache would have turned into invalid upon reconnection. Another aspect is if the mobile client moves to a new location, it will have to notify the server of the relocation. And all these issues, such as disconnection 0
and mobility, have to be taken into account because it incurs costs from sending the messages to and from the server via the uplink and downlink messages. The second possible way is to have the mobile client query the server directly in order to verify the validity of the cache data prior to using it. This appears to be straightforward and easy, but one has to bear in mind that this method would generate a lot of uplink traffic in the network. In contrast to stateful method, another way that can be taken into account in designing the invalidation is using a stateless method. This method is in direct opposite from the first possible way, which is the stateful method. In this method, the server is not aware of the state of the client’s cache and the client location and disconnected status. The server would not care about all these but just periodically broadcasts an IR containing the data items that have been updated or changed in comparison to its previous state. Thus the server just keeps track of which item is recently updated and broadcasts them in an IR. Then only the client determines whether its cache is valid or not by validating it against the IRs that are broadcasted on the wireless channel. Another challenging issue that involves determining an efficient invalidation strategy is to optimize the organization of the IR. Commonly, a large-sized report provides more information and appears to be more effective. But publishing a large report also brings drawbacks, such as implying a long latency for mobile clients to listen to the report due to the low bandwidth wireless channel. There have been several methods proposed in addressing the report optimization issue in other works, such as using the dual report scheme and bit sequence scheme (Tan, Cai, & Ooi, 2001; Elmagarmid et al., 2003).
Cache Invalidation in a Mobile Environment
Figure 3. Architecture of location dependent query processing
P1 P
C
2
P7 P5
P3
P6
User move from Location A towards Location B
Query 1
P4
Location-Dependent Cache Invalidation Due to the fact that mobile users in a typical mobile environment move around frequently by changing location has opened up a new challenge of answering queries that is dependent on the current geographical coordinates of the users (Barbara, 1999; Waluyo, Srinivasan, & Taniar, 2005). This is known as location dependent queries (Kottkamp & Zukunft, 1998). In this location dependent query, the server would produce answers to a query based on the location of the mobile client issuing the query. Thus, a different location may sometimes yield a different result even though the query is taken from a similar source. Figure 3 depicts an illustration of a location dependent query processing. This shows that when the mobile client is in Location A, the query would return a set of results and when the mobile client moves towards a new Location B, another set of results will be returned. However there are cases of results overlapping between nearby locations. An example of a location dependent query can be: “Find the nearest restaurants from where I am standing now.” This is an example of static object whereby restaurants are not moving. An example of a dynamic object would be: “What is the nearest taxi that will pass by me” (Lee et al., 2002). With the frequent movement of mobile users, very often the mobile clients would query the same server to obtain results, or with the frequent movement of mobile users from location to location, very often the mobile clients would suffer from scarce bandwidth and frequent disconnection, especially when suddenly moved towards a secluded area (Jayaputera & Taniar, 2005). Hence, is essential to have data caching that can cope with cases of frequent disconnections. And often data may have become invalid after a certain point of time, especially in the area of location dependent. Example 2: A mobile user who is in Location A, cached the results of the nearby vegetarian restaurants in Location A. As he moves to Location B, he would like another list of
P8
Query 2
nearby vegetarian restaurants. The user is sending a query to the same source, but the results returned are different due to location dependent data. And because there are data previously cached, this data—which is the result obtained when he is in Location A—would become invalid since he has now moves to Location B. Hence, location dependent cache invalidation serves the purpose of maintaining the validity of the cached data when the mobile client moves from one location to another location (Zheng, Xu, & Lee, 2002). The emergence of this location dependent cache invalidation is due to mobile client’s movement and thus the data value for a data item is actually dependent on the geographical location. Hence, traditional caching that does not consider geographical location is inefficient for location dependent data. There are both advantages and disadvantages of location dependent cache invalidation. The major benefit that the attached invalidation information provided is that it provides a way for the client to be able to check for validity of cached data in respect to a certain location. These are necessary, especially in cases of when the mobile client wishes to issue the same query later when he/she moves to a new location. Another situation for the importance in checking the validity of the cached data is that because mobile clients keep on moving even right after they submit a query, they would have arrived to a new location when the results are returned. This may occur if there is a long delay in accessing data. Thus, if this two situations occur, then it is significant to validate the cached data because it may have become invalid (Zheng, Xu, & Lee, 2002).
FUTURE TRENDS There have been several researches done in the area of exploring cache invalidation in a mobile environment. The usage of cache invalidation has obviously provoked extensive 0
Cache Invalidation in a Mobile Environment
complicated issues. There are still many limitations of the nature of the mobile environment as well as mobility of the users that generate a lot of attention from research in finding a good cache strategy that can cope well with frequent disconnection and low power consumption. In the future, it is critical to build an analytical model to get a better understanding of how cache invalidation works and how well it can cope in the mobile environment. Developing caching strategies that support cache invalidation for a multiple channel environment is also desirable, whereby a mixture of broadcast and point-to-point channels are being used. Including a dynamic clustering is also beneficial in order to allow the server to group data items together as their update changes. Besides these, further investigation on other cache replacement policies, as well as granularities issues, is also beneficial. Due to the non-stop moving clients, further research on adapting cache invalidation into location dependent data is favorable. Another possible issue that could open up for future work may involve minimizing the waiting time for the mobile client in acquiring the IR, since the mobile client has to obtain an IR prior to their cache being validated. Thus, it is essential to be able to reduce waiting time. Another aspect is due to wireless channels that are often error prone due to their instability, bandwidth, and so on. Thereby, having techniques to handle errors in a mobile environment is definitely helpful. Last but not least, having further study on integrating several different strategies to obtain a more optimal solution in coping with mobile environment is advantageous.
CONCLUSION Although there is a significant increase in the popularity of mobile computing, there are still several limitations that are inherent, be it the mobile device itself or the environment itself. These include limited battery power, storage, communication cost, and bandwidth problems. All these have become present challenges for researchers to address. In this article, we have described the pros and cons of adopting cache invalidation in a mobile environment. We include adapting cache invalidation strategy in both location and non-location dependent queries. Discussion regarding the issue in designing cache invalidation is also provided in a preliminary stage. This article serves as a valuable starting point for those who wish to gain some introductory knowledge about the usefulness of cache invalidation.
REFERENCES Barbara, D., & Imielinski, T. (1994). Sleepers and workaholics: Caching strategies in mobile environments. MOBIDATA: An Interactive Journal of Mobile Computing, 1(1). 0
Cai, J., & Tan, K. L. (1999). Energy efficient selective cache invalidation. Wireless Networks, 5(6), 489-502. Chan, B. Y., Si, A., & Leong, H. V. (1998). Cache management for mobile databases: Design and evaluation. In Proceedings of the International Conference on Data Engineering (ICDE) (pp. 54-63). Chand, N., Joshi, R., & Misra, M. (1996). Energy efficient cache invalidation in a disconnected mobile environment. In Proceedings of the Twelfth International Conference on Data Engineering (pp.336-343). Cao, G. (2003). A scalable low-latency cache invalidation strategy for mobile environment. IEEE Transaction on Knowledge and Data Engineering (TKDE), 15(2), 12511265. Cao, G. (2002). On improving the performance of cache invalidation in mobile environment. Mobile Networks and Applications, 7(4), 291-303. Deshpande, P. M., & Ramasamy, K. (1998). Caching multidimensional queries using chunks. In Proceedings of the ACM SIGMOD Conference on Management of Data (pp. 259-270). Dong Jung, Y. H., You, J., Lee, W., & Kim, K. (2002). Broadcasting and caching policies for location dependent queries in urban areas. In Proceedings of the 2nd International Workshop on Mobile Commerce (pp. 54-60). Elmagarmid, A., Jing, J., Helal, A., & Lee, C. (2003). Scalable cache invalidation algorithms for mobile data access. IEEE Transaction on Knowledge and Data Engineering (TKDE), 15(6), 1498-1511. Hu, Q., & Lee, D. (1998). Cache algorithms based on adaptive invalidation reports for mobile environment. Cluster Computing, pp. 39-48. Hurson A.R., & Jiao, Y. (2005). Data broadcasting in mobile environment. In D. Katsaros, A. Nanopoulos, & Y. Manolopaulos (Eds.), Wireless information highways (Chapter 4). Hershey, PA: IRM Press. Imielinski, T., & Badrinath, B. (1994). Mobile wireless computing: Challenges in data management. Communications of the ACM, 37(10), 18-28. Jayaputera, J., & Taniar, D. (2005). Data retrieval for location-dependent queries in a multi-cell wireless environment. Mobile Information Systems, 1(2), 91-108. Kara, H., & Edwards, C. (2003). A caching architecture for content delivery to mobile devices. In Proceedings of the 29th EUROMICRO Conference: New Waves in System Architecture (EUROMICRO’03).
Cache Invalidation in a Mobile Environment
Kottkamp, H.-E., & Zukunft, O. (1998). Location-aware query processing in mobile database systems. In Proceedings of ACM Symposium on Applied Computing (pp. 416-423). Lai, K.Y., Tari, Z., & Bertok, P. (2003). Cost efficient broadcast based cache invalidation for mobile environment. In Proceedings of the 2003 ACM symposium on Applied Computing (pp. 871-877). Lee, G., Lo, S-C., & Chen, A. L. P. (2002). Data allocation on wireless broadcast channels for efficient query processing. IEEE Transactions on Computers, 51(10), 1237-1252. Leong, H. V., & Si, A. (1997). Database caching over the air-storage. The Computer Journal, 40(7), 401-415. Lee, D-L., Zhu, M., & Hu, H. (2005). When location-based services meet databases. Mobile Information Systems, 1(2), 81-90. Lee, D. K., Xu, J., Zheng, B., & Lee, W-C. (2002). Data management in location-dependent information services. IEEE Pervasive Computing, 2(3), 65-72. Prabhajara, K., Hua, K. A., & Oh, J.H. (2000). Multi-level, multi-channel air cache designs for broadcasting in a mobile environment. In Proceedings of the 16th International Conference on Data Engineering (pp. 167-186). Park, K., Song, M., & Hwang, C. S. (2004). An efficient data dissemination schemes for location dependent information services. In Proceedings of the First International Conference on Distributed Computing and Internet Technology (ICDCIT 2004) (Vol. 3347, pp.96-105). Springer-Verlag. Tan, K. L., Cai, J., & Ooi, B. C. (2001). An evaluation of cache invalidation strategies in wireless environment. IEEE Transactions on Parallel and Distributed Systems, 12(8), 789-807. Waluyo, A. B., Srinivasan, B., & Taniar, D. (2005). Research on location-dependent queries in mobile databases. International Journal on Computer Systems: Science and Engineering, 20(3), 77-93. Waluyo, A. B., Srinivasan, B., & Taniar, D. (2005). Research in mobile database query optimization and processing. Mobile Information Systems, 1(4). Xu, J., Hu, Q., Tang, X., & Lee, D. L. (2004). Performance analysis of location dependent cache invalidation scheme for mobile environments. IEEE Transaction on Knowledge and Data Engineering (TKDE), 15(2), 125-139. Xu, J., Hu, Q., Lee, D. L., & Lee, W.-C. (2000). SAIU: An efficient cache replacement policy for wireless on-demand broadcasts. In Proceedings of the 9th International Conference on Information and Knowledge Management (pp. 46-53).
Xu, J., Hu, Q., Lee, W.-C., & Lee, D. L. (2004). Performance evaluation of an optimal cache replacement policy for wireless data dissemination. IEEE Transaction on Knowledge and Data Engineering (TKDE), 16(1), 125-139. Yajima, E., Hara, T., Tsukamoto, M., & Nishio, S. (2001). Scheduling and caching strategies for correlated data in push-based information systems. ACM SIGAPP Applied Computing Review, 9(1), 22-28. Zheng, B., Xu, J., & Lee, D.L. (2002, October). Cache invalidation and replacement strategies for location-dependent data in mobile environments. IEEE Transactions on Computers, 51(10), 1141-1153.
KEY TERMS Caching: Techniques of temporarily storing frequently accessed data designed to reduce network transfers and therefore increase speed of download Cache Invalidation Strategy: A type of caching strategy that is used to ensure that the data items that are cached in the mobile client are consistent in comparison to the ones that are stored on the server. Caching Management Strategy: A strategy that relates to how client manipulates the data that has been cached in an efficient and effective way by maintaining the data items in a client’s local storage. Invalidation Report (IR): An informative report in which the changed data items are indicated; it is used for mobile clients to validate against their cache data to check if it is still valid or not. Location-Dependent Cache Invalidation: maintaining the validity of the cached data when the mobile client changes locations. Mobile Environment: Refers to a set of database servers, which may or may not be collaborative with one another, that disseminate data via wireless channels to multiple mobile users. Pull-Based Environment: Also known as an on demand system, which relates to techniques that enable the server to process request that are sent from mobile users. Push-Based Environment: Also known as a broadcast system where the server would broadcast a set of data to the air for a population of mobile users to tune in for their required data.
0
C
0 Category: Service Computing
Communicating Recommendations in a Service-Oriented Environment Omar Khadeer Hussain Curtin University of Technology, Australia Elizabeth Chang Curtin University of Technology, Australia Farookh Khadeer Hussain Curtin University of Technology, Australia Tharam S. Dillon University of Technology, Sydney, Australia
INTRODUCTION The Australian and New Zealand Standard on Risk Management, AS/NZS 4360:2004 (Cooper, 2004), states that risk identification is the heart of risk management. Hence risk should be identified according to the context of the transaction in order to analyze and manage it better. Risk analysis is the science of evaluating risks resulting from past, current, anticipated, or future activities. The use of these evaluations includes providing information for determining regulatory actions to limit risk, and for educating the public concerning particular risk issues. Risk analysis is an interdisciplinary science that relies on laboratory studies, collection, and exposure of data and computer modeling. Chan, Lee, Dillon, and Chang (2002) state that the advent of the Internet and its development has simplified the way transactions are carried out. It currently provides the user with numerous facilities which facilitate transaction process. This process evolved into what became known as e-commerce transactions. There are two types of architectures through which e-commerce transactions can be conducted. They are: (a) client-server business architecture, and (b) peer-to-peer business architecture. In almost all cases, the amount of risk involved in a transaction is important to be understood or analyzed before a transaction is begun. This also applies to the transactions in the field of e-commerce and peer-to-peer business. In this article we will emphasize transactions carried out in the peer-to-peer business architecture style, as our aim is to analyze risk in such transactions carried out in a serviceoriented environment. Peer-to-peer (P2P) architecture is so called because each node has equivalent responsibilities (Leuf, 2002). This is a type of network in which each workstation or peer has equivalent capabilities and responsibilities. This differs from client/server architecture, in which some computers or
central servers are dedicated to serving others. As mentioned by Oram (2001), the main difference between these two architectures is that in peer-to-peer architecture, the control is transferred back to the clients from the servers, and it is the responsibility of the clients to complete the transaction. Some of the characteristics of peer-to-peer or decentralized transactions are: 1. 2. 3.
There is no server in this type of transaction between peers. Peers interact with each other directly, rather than through a server, as compared to a centralized transaction where the authenticity can be checked. Peers can forge or create multiple identities in a decentralized transaction, and there is no way of checking the identity claimed by the peer to be genuine or not.
The above properties clearly show that a decentralized transaction carries more risks and hence merits more detailed investigation. Similarly, in a service-oriented peer-to-peer financial transaction, there is the possibility of the trusted agent engaging in an untrustworthy manner and in other negative behavior at the buyer’s expense, which would result in the loss of the buyer’s resources. This possibility of failure and the degree of possible loss in the buyer’s resource is termed as risk. Hence, risk analysis is an important factor in deciding whether to proceed in an interaction or not, as it helps to determine the likelihood of loss in the resources involved in the transaction. Risk analysis by the trusting agent before initiating an interaction with a trusted agent can be done by: • •
determining the possibility of failure of the interaction, and determining the possible consequences of failure of the interaction.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Communicating Recommendations in a Service-Oriented Environment
Figure 1. The riskiness scale and its associated levels Riskiness Levels
Magnitude of Risk
Riskiness Value
Star Rating
Unknown Risk
-
-1
Not Displayed
Totally Risky
91-100% of Risk
0
Not Displayed
Extremely Risky
71-90% of Risk
1
Largely Risky
70% of Risk
2
Risky
26-50% of Risk
3
Largely Unrisky
11-25% of Risk
4
Unrisky
0-10% of Risk
5
The trusting agent can determine the possibility of failure in interacting with a probable trusted agent either by: a. b.
considering its previous interaction history with the trusted agent, if any, in the context of its future interaction, or soliciting recommendations for the trusted agent in the particular context of its future interaction, if it does not have any previous interaction history with it.
When the trusting agent solicits for recommendations about a trusted agent for a particular context, then it should consider replies from agents who have previous interaction history with the trusted agent in that particular context. The agents replying back with the recommendations are called the recommending agents. But it is possible that each recommending agent might give its recommendation in its own way, and as a result of that, it will be difficult for the trusting agent to interpret and understand what each element of the recommendations mean. Hence, a standard format for communicating recommendations is needed so that it is easier for the trusting agent to understand and assimilate them. Further, the trusting agent has to determine whether the recommendation communicated by the recommending agent is trustworthy or not before considering it. In this article we propose a methodology by which the trusting agent classifies the recommendation according to its trustworthiness. We also define a standard format for communicating recommendations, so that it is easier for the trusting agent to interpret and understand them.
BACKGROUND Security is the process of providing sheltered communication between two communicating agents (Singh & Liu, 2003;
From From From From From
C
to to to to to
Chan et al., 2002). We define risk in a peer-to-peer serviceoriented environment transaction as the likelihood that the transaction might not proceed as expected by the trusting agent in a given context and at a particular time once it begins resulting in the loss of money and the resources involved in it. The study of risk cannot be compared with the study of security, because securing a transaction does not mean that there will be no risk in personal damages and financial losses. Risk is a combination of: a. b.
the uncertainty of the outcome; and the cost of the outcome when it occurs, usually the loss incurred.
Analyzing risk is important in e-commerce transactions, because there is a whole body of literature based on rational economics that argues that the decision to buy is based on the risk-adjusted cost-benefit analysis (Greenland, 2004). Thus it commands a central role in any discussion of e-commerce that is related to a transaction. Risk plays a central role in deciding whether to proceed with a transaction or not. It can broadly be classified as an attribute of decision making that reflects the variance of its possible outcomes. Peer-to-peer architecture-type transactions are being described as the next generation of the Internet (Orlowska, 2004). Architectures have been proposed by researchers (Qu & Nejdl, 2004; Schmidt & Parashar, 2004; Schuler, Weber, Schuldt, & Schek, 2004) for integrating Web services with peer-to-peer communicating agents like Gnutella. However, as discussed earlier, peer-to-peer-type transactions suffer from some disadvantages, and risk associated in the transactions is one of them. Hence, this disadvantage has to be overcome so that they can be used effectively with whatever service they are being integrated with. Through the above discussion, it is evident that risk analysis is necessary when a transaction is being conducted in a 0
Communicating Recommendations in a Service-Oriented Environment
peer-to-peer architecture environment. As mentioned before, risk analysis by the trusting agent can be done by determining the possibility of failure and the possible consequences of failure in interacting with a probable trusted agent. In order for the trusting agent to determine and quantify the possibility of failure of an interaction, we define the term riskiness. Riskiness is defined as the numerical value that is assigned by the trusting agent to the trusted agent after the interaction, which shows the level of possibility of failure of an interaction on the riskiness scale. The numerical value corresponds to a level on the riskiness scale, which gives an indication to other agents about the level of possibility of failure in interacting with a particular trusted agent. The riskiness scale as shown in Figure 1 depicts different levels of possibility of failure that could be present in an interaction. The riskiness value to the trusted agent is assigned by the trusting agent after assessing the level of un-commitment in its actual behavior with respect to the promised commitment. The promised commitment is the expected behavior by which the trusted agent was supposed to behave in the interaction. The expected behavior is defined by the trusting agent according to its criteria, before starting its interaction with the trusted agent. The actual behavior is the actual commitment that the trusted agent showed or behaved in the interaction. Criteria are defined as the set of factors or bases that the trusting agent wants in the interaction and later against which it determines the un-committed behavior of the trusted agent in the interaction. If the trusting agent has interacted previously with the trusted agent in the same context as its future interaction, then it can determine the possibility of failure in interacting with it by analyzing the riskiness value that it assigned to the trusted agent in their previous interaction. If a trusting agent has not interacted previously with a trusted agent in a particular context, then it can determine the possibility of failure in their future interaction, by soliciting for its recommendation from other agents who have dealt with the same trusted agent previously in the same context as that of the trusting agent’s future interaction. As mentioned earlier the agents giving recommendations are called recommending agents. But it would be difficult for the trusting agent to assimilate the data that it gets from the recommending agents and draw a conclusion if each agent gives its recommendation in its own format. It would rather be easier for the trusting agent if the recommendations came in a standard set or format that enables the trusting agent to ascertain the meaning of each element in the recommendations. But even in the same context, each recommending agent might have different criteria in its interaction with the trusted agent. Consequently the riskiness value that it recommends for the trusted agent depends on its assessment of un-commitment in the trusted agent’s actual behavior with respect to its expected behavior in those criteria. It would be baseless for 0
the trusting agent to consider recommendations for a trusted agent in criteria of assessment which are not similar to those in its future interaction with that particular trusted agent. Additionally it is highly unlikely that the recommendations provided by the recommending agents would be completely reliable or trustworthy. Some agents might be communicating un-trustworthy recommendations. The trusting agent has to consider all these scenarios before it assimilates the recommendations from the recommending agents to assess the risk in dealing with a trusted agent. In order to propose a solution to these issues, in the next sections we will define a methodology by which the trusting agent can classify the recommendations according to its trustworthiness. We also define a standard format for communicating recommendations so that the trusting agent can ascertain the meaning of each element of the recommendations before assimilating them, and consider only those whose criteria are of interest to it in its future interaction.
CLASSIFYING THE RECOMMENDATIONS AS TRUSTWORTHY OR UN-TRUSTWORTHY As stated earlier, it is possible that the recommendation communicated by a recommending agent might not be trustworthy. The recommending agent might be communicating recommendations that the trusting agent finds to be incorrect or misleading after its interaction with the trusted agent. So the trusting agent has to determine whether the recommendation is trustworthy or not before assimilating it. To achieve that, we propose that each recommending agent is assigned a riskiness value while giving recommendations called riskiness of the recommending agent (RRP). The riskiness value of the recommending agent is determined by the difference between: • •
the riskiness value that the trusting agent found out for the trusted agent after interacting with it, and the riskiness value that the recommending agent recommend for the trusted agent to the trusting agent when solicited for.
When the trusting agent broadcasts a query soliciting for recommendations about a trusted agent in a particular context, it will consider replies from those agents who have interacted with that particular trusted agent previously in that same context. Hence, whatever riskiness value the recommending agents recommend to the trusting agent will be greater than -1, as -1 on the riskiness scale represents the riskiness value as Unknown Risk, which cannot be assigned to any agent after an interaction. After an interaction a value only within the range of (0, 5) on the riskiness scale can be
Communicating Recommendations in a Service-Oriented Environment
assigned. So, the maximum range for the riskiness value of the recommending agent (RRP) is between (-5, 5), since this is the maximum possible range of difference between the riskiness value that the trusting agent might determine for the trusted agent after its interaction with it and the riskiness value recommended by the recommending agent for the trusted agent to the trusting agent. We adopt the approach mentioned by Chang, Dillon, and Hussain (2006) which states that a recommending agent is said to be communicating trustworthy recommendations if its riskiness value while giving recommendations (RRP) is in the range of (-1, 1). A value within this range will state that there is a difference of one level in the riskiness value that the trusting agent found out after the interaction and what the recommending agent suggested for the trusted agent. If the riskiness value of the recommending agent is beyond those levels, then it hints that the recommending agent is giving recommendations that the trusting agent finds to vary a lot after the interaction, and there is at least a difference of two levels on the riskiness scale between what the trusting agent found and what the recommending agent recommended. An agent whose Riskiness value while giving recommendation (RRP) is beyond the level of (-1, 1) is said to be an Un-trustworthy recommending agent. Chang et al. (2006) mention that the trusting agent should only consider recommendations from agents who are either Trustworthy or Unknown in giving recommendations and leave the recommendation from agents who are Un-trustworthy in giving them. Hence the recommendation from agents with riskiness values beyond the levels of (-1, 1) will not be considered. If the recommending agent gives more than one recommendation in an interaction, then its riskiness value while giving recommendation can be determined by taking the average of the difference of each recommendation. Hence riskiness of the recommending agent (RRP) =
1 N
N
∑ i =1
(Ti – Ri)
where Ti is the riskiness value found out by the trusting agent after the interaction, Ri is the riskiness value recommended by the recommending agent for the trusted agent, and N is the number of recommendations given by a particular agent.
DEFINING A STANDARD FORMAT FOR COMMUNICATING RECOMMENDATIONS Whenever a trusting agent interacts with a trusted agent, a risk relationship is formed between them. The risk relationship consists of a number of factors. These include the trusting agent:
1.
2. 3. 4.
considering its previous experience with the trusted agent in the context of its future interaction with it, or soliciting recommendations for the trusted agent in the context of its future interaction if they have not interacted before; determining the riskiness value of the trusted agent according to its previous interactions or recommendations; predicting the future riskiness value of the trusted agent, within the time period of its interaction with the trusted agent; or taking into consideration the cost of the interaction and assigning a riskiness value to the trusted agent after completing the interaction.
A risk relationship exists between a trusting agent and a trusted agent only if they interact with each other. Between a trusting agent and a trusted agent, there might exist one or more risk relationships depending on the number of times they interact with each other. For each interaction a new risk relationship is formed. Hence, the trusting agent and the trusted agent are in a ternary association (Eriksson & Penker, 2000) with the risk relationship as shown in Figure 2. But the risk relationship exists only if the trusting agent interacts with the trusted agent, and hence it is realized by a transaction between them. The risk relationship in turn is dependent on a number of factors. Figure 2 shows the risk relationship and the factors on which it is dependent. As mentioned earlier, the trusting agent will consider recommendations from agents who have previous interaction history with the particular trusted agent in question in context similar to that of its future interaction with the trusted agent. A recommending agent when solicited for recommendation by a trusting agent for a particular trusted agent in a particular context will give its recommendation depending on its previous interaction with the particular trusted agent in the particular context. In other terms, it gives the risk relationship that it had formed with the trusted agent in that particular interaction as its recommendation. We propose that when any trusting agent solicits for recommendations for a trusted agent, then the recommending agents should give their replies in a standard format so that it is easier for the trusting agent to interpret their recommendation. The standard format is represented by a risk set. The risk set is formed from the risk relationship that the recommending agent had from the last time it interacted with the particular trusted agent. Alternately, a risk set exists between any two agents only if there is a risk relationship between them, and hence it is dependent on the risk relationship as shown in Figure 2. Once the risk relationship between any two agents has been established, then a risk set can be defined. The risk set contains the same elements as that of the risk relationship but in an ordered way. The order of appearance of the
C
Communicating Recommendations in a Service-Oriented Environment
Figure 2. Risk relationship that exists between any two agents
Trusting Agent
Trusted Agent
1
1
* Transaction
Medium of the Transaction
Risk Relationship
Outcome of the Previous Transaction or Recommendations from the Recommending peers
Context of the Transaction
Risk Set
Time of the Transaction
Cost of the Transaction
Criteria in the Context
elements of the risk set is: {TP1, TP2, Context, CR, R’, (Criteria, Commitment level), R, Cost, Start time, End time, RRP}, where: • • • •
• •
TP1 denotes the trusting agent in the interaction. This is also the recommending agent while communicating recommendations. TP2 denotes the trusted agent in the interaction. Context represents the context of the interaction. CR represents the ‘current riskiness’ value of the trusted agent before its interaction with the recommending agent. This is achieved either by the previous interaction history between the recommending agent and the trusted agent in the same context or by soliciting recommendations for the trusted agent by the recommending agent before its interaction, R’ shows the predicted riskiness value of the trusted agent as determined by the recommending agent within the time slot of its interaction. (Criteria, Commitment Level) shows the factors or bases that the recommending agent used in its interac-
•
tion with the trusted agent to assign it a riskiness value. These criteria are necessary to mention while giving recommendations, so that a trusting agent who asks for recommendation knows the factors on which this particular trusted agent has assigned the recommended riskiness value and only considers those recommendations which are of interest to it according to its criteria. Commitment level specifies whether the particular criterion was fulfilled by the trusted agent or not. A value of either 0 or 1 assigned here is based on its commitment. A value of 0 signifies that the criterion was not fulfilled by the trusted agent according to the expected behavior, whereas a value of 1 signifies that the criterion was fulfilled according to the expected behavior. R is the riskiness value assigned by the recommending agent to the trusted agent after its interaction. As discussed earlier, the riskiness value is determined after the interaction by assessing the level of un-commitment in the trusted agent’s actual behavior with respect to the expected behavior.
Communicating Recommendations in a Service-Oriented Environment
• • • •
Cost represents the cost of the interaction. Start Time is the time at which the recommending agent started the interaction with the trusted agent. End Time is the time at which the interaction of the recommending agent ended with the trusted agent. RRP is the riskiness value of the recommending agent while giving recommendations. This value determines whether the recommendation is trustworthy or not.
Recommendation from agent ‘D’ is: {Agent ‘D’, Logistic Company ‘LC’, Transport, 4, 4, ((C1, 1) (C2, 1), (C3, 1)), 5, UNKNOWN, 07/04/2006, 10/04/2006, 0} The properties to be followed while forming or representing the risk set are:
To highlight the advantages of communicating the recommendations in a standard format by risk set and the usefulness of its elements, let us consider a scenario in which Bob wants to interact with a logistic company ‘LC’. The context of its interaction with the logistic company is to ‘transport its goods’ on April 15, 2006. Let us represent the context as ‘Transport’. The goods are worth $1,500. The criteria put up by Bob in its interaction with the logistic company ‘LC’ are:
1.
1. 2.
5.
3.
Packing the goods properly. Pickup of the goods on time by the logistic company. Delivering the goods to the correct address on time as promised.
For explanation sake the criteria in the interaction are represented by C1, C2, and C3 respectively. The trusting agent Bob does not have any previous dealings with the logistic company ‘LC’, and in order to analyze the risk before proceeding in a business transaction with it, Bob solicits for recommendations from other agents who have previously dealt with logistic company ‘LC’ in a context similar to that in this interaction. The agents who had interacted previously in the same context give their recommendations to Bob in the form of a risk set, which relates to their previous interactions with the trusted agent ‘LC’. Let us suppose that Bob receives recommendations from agents ‘A’, ‘B’, ‘C’, and ‘D’ in the form of risk set. The recommendation from agent ‘A’ is: {Agent ‘A’, Logistic Company ‘LC’, Transport, 5, 5, ((C3, 1) (C1, 0), (C2, 1)), 5, $5000, 15/07/2005, 22/07/2005, 0.8} Similarly, recommendation from agent ‘B’ is: {Agent ‘B’, Logistic Company ‘LC’, Transport, 3, 3, ((C5, 1) (C6, 0)), 3, $1000, 1/02/2006, 22/02/2006, -1} Recommendation from agent ‘C’ is: {Agent ‘C’, Logistic Company ‘LC’, Transport, 2, 3, ((C1, 0) (C2, 0), (C3, 1)), 2, UNKNOWN, 01/04/2006, 03/04/2006, -2.5}
2. 3.
4.
6.
The elements should be represented in the same order as defined above. Each element of the risk set is mandatorally to be defined except the element ‘cost’. Each criteria and its commitment level should be represented inside a single ‘(‘, ‘)’ bracket, separated by a comma, so as to differentiate it from the other criteria and the elements of the risk set. If the cost is not represented, then it should be written as UNKNOWN. If the riskiness value of the recommending agent (RRP) is not known, then it should be represented as UNKNOWN. The elements of the risk set should be separated by a comma ‘,’. From the above recommendations it can be seen that:
•
•
•
•
The recommendation from the recommending agent ‘A’ is trustworthy and exactly according to the criteria of the trusting agent’s future interaction with the trusted agent. But there is a huge gap in time between the recommending agent’s interaction with the trusted agent and the future interaction of the trusting agent with the trusted agent. The recommendation from recommending agent ‘B’ is trustworthy, but the criteria of its recommendation does not match with those of the trusting agent’s future interaction with the trusted agent and so it is baseless for it to consider this recommendation. The criteria of recommending agent ‘C’ is similar to those of the trusting agent, but the riskiness value of the recommending agent ‘C’ (RRP) is not within the range of (-1,1). So it can be concluded that this is an un-trustworthy recommendation and it will not be considered by the trusting agent. The recommendation from recommending agent ‘D’ is trustworthy and in the criterions that the trusting agent wants in its interaction.
Hence, as can be seen, the trusting agent, by making use of the risk set, can interpret the meaning of each element of the recommendation that would help it to understand the recommendation better and assimilate it easily.
C
Communicating Recommendations in a Service-Oriented Environment
The advantages of communicating the recommendations in the form of a risk set are: 1. 2.
3.
4.
The recommendations come in a standard format and it is easier for the trusting agent to understand them. Even if the context of two interactions is the same, the criteria might differ considerably, and the riskiness value assigned to each interaction is in accordance with its corresponding criteria. Therefore, while giving recommendations, the recommending agent must specify the criteria apart from the context of the interaction. By doing so, the trusting agent who is soliciting for recommendations might know the exact criteria in which the trusted agent was assigned the riskiness value recommended by the recommending agent and consider only those recommendations that are of interest to it. The risk set communicates the criteria along with the recommendations and also specifies the commitment level of the trusted agent in those criteria. The risk set specifies the riskiness value of the trusted agent as determined by recommending agent before starting an interaction with it (CR), the predicted future riskiness value of the trusted agent within the time space of its interaction (R’), and the actual riskiness value of the trusted agent determined by the recommending agent (R) after its interaction with it depending on the level of un-commitment in the actual behavior of the trusted agent with respect to the expected behavior. The risk set specifies the time of the interaction between the recommending agent and the trusted agent. As defined in the literature, risk is dynamic and it keeps on changing. It is not possible for an agent to have the same impression of another agent that it had at a given point of time. Hence, the trusting agent should give more weight to those recommendations which are near to the time slot of its interaction as compared to the far recent ones while assimilating them. This is achieved by using the proposed risk set, which specifies the context along with the accessing criteria and the riskiness values of the trusted agent according to the time assigned, in an ordered way.
CONCLUSION In this article we discussed the need to analyze risk that could be associated in a peer-to-peer financial transaction. Further we discussed how a trusting agent can assess the possible risk beforehand that could be present in interacting with a particular trusted agent. We proposed a methodology of classifying the recommendations according to its trustworthiness. Further we discussed the risk relationship that exists between a trusting agent and a trusted agent in the
post-interaction phase, and ascertain the factors on which the risk relationship is dependent. From that relationship we defined the risk set, which is an ordered way of representing the details of the transaction between the agents. This risk set is utilized by the recommending agents while communicating recommendations to the trusting agents, so that it can be interpreted and understood easily.
REFERENCES Chan, H., Lee, R., Dillon, T. S., & Chang, E. (2002). E-commerce: Fundamentals and applications (1st ed.). New York: John Wiley & Sons. Chang, E., Dillon, T., & Hussain F. K. (2006). Trust and reputation in service oriented environments (1st ed.). New York: John Wiley & Sons. Cooper, D. F. (2004). The Australian and New Zealand standard on risk management, AS/NZS 4360:2004, tutorial notes: Broadleaf Capital International Pty Ltd. Retrieved from http://www.broadleaf.com.au/tutorials/Tut_Standard. pdf Eriksson, H., & Penker, M. (2000). Business modeling with UML: Business patterns at work. New York: John Wiley & Sons. Greenland, S. (2004). Bounding analysis as an inadequately specified methodology. Risk Analysis, 24(5), 1085-1092. Leuf, B. (2002). Peer to peer collaboration & sharing on the Internet. Pearson Education. Oram, A. (2004). Peer-to-peer: Harnessing the power of disruptive technologies. Retrieved February 16, 2004, from http://www.oreilly.com/catalog/peertopeer/chapter/ch01. html Orlowska, M. E. (2004). The next generation messaging technology—makes Web services effective. Proceedings of the 6th Asia Pacific Web Conference (pp. 13-19). Berlin/ Heidelberg: Springer-Verlag. Qu, C., & Nejdl, W. (2004). Interacting the Edutella/JXTA peer-to-peer network with Web services. Proceedings of the International Symposium on Applications and the Internet (SAINT’04), Tokyo, (pp. 67-73). Schmidt, C., & Parashar, M. (2004). A peer-to-peer approach to Web service discovery. World Wide Web Journal, 7(2), 211-229. Schuler, C., Weber, R., Schuldt, H., & Schek, H. (2004). Scalable peer-to-peer process management—the OSIRIS approach. Proceedings of the 2nd International Conference on Web Services, San Diego, (pp. 26-34).
Communicating Recommendations in a Service-Oriented Environment
Singh, A., & Liu, L. (2003). TrustMe: Anonymous management of trust relationships in decentralized P2P systems. Proceedings of the 3rd IEEE International Conference on P2P Computing (pp. 142-149), Linköping, Sweden.
Riskiness Value: A value that is assigned to the trusted agent by the trusting agent after its interaction with it. This value specifies a level of risk on the riskiness scale that the trusted agent deserves according to the level of un-committed behavior in its interaction with the trusting agent.
KEY TERMS
RRP: Stands for riskiness value of the recommending agent. This value is used to determine if the recommending agent is communicating trustworthy recommendations or not.
Recommending Agent: An agent who gives its recommendation about a trusted agent to a trusting agent, when solicited for. Risk Set: A standard format for giving recommendations by the recommending agents. Riskiness Scale: A scale that represents different levels of risk that could be possible in an interaction.
Trusted Agent: An agent with whom the trusting agent deals with and reposes its faith in. Trusting Agent: An agent who controls the resources and interacts with another agent after reposing its faith in it.
C
Category: Mobile Software Engineering
Content Personalization for Mobile Interfaces Spiridoula Koukia University of Patras, Greece Maria Rigou University of Patras, Greece and Research Academic Computer Technology Institute, Greece Spiros Sirmakessis Technological Institution of Messolongi and Research Academic Computer Technology Institute, Greece
INTRODUCTION The contribution of context information to content management is of great importance. The increase of storage capacity in mobile devices gives users the possibility to maintain large amounts of content to their phones. As a result, this amount of content is increasing at a high rate. Users are able to store a huge variety of content such as contacts, text messages, ring tones, logos, calendar events, and textual notes. Furthermore, the development of novel applications has created new types of content, which include images, videos, MMS (multi-media messaging), e-mail, music, play lists, audio clips, bookmarks, news and weather, chat, niche information services, travel and entertainment information, driving instructions, banking, and shopping (Schilit & Theimer, 1994; Schilit, Adams, & Want, 1994; Brown, 1996; Brown, Bovey, & Chen, 1997). The fact that users should be able to store the content on their mobile phone and find the content they need without much effort results in the requirement of managing the content by organizing and annotating it. The purpose of information management is to aid users by offering a safe and easy way of retrieving the relevant content automatically, to minimize their effort and maximize their benefit (Sorvari et al., 2004). The increasing amount of stored content in mobile devices and the limitations of physical mobile phone user interfaces introduce a usability challenge in content management. The physical mobile phone user interface will not change considerably. The physical display sizes will not increase since in the mobile devices the display already covers a large part of the surface area. Text input speed will not change much, as keyboard-based text input methods have been the most efficient way to reduce slowness. While information is necessary for many applications, the human brain is limited in terms of how much information it can process at one time. The problem of information management is more complex in mobile environments (Campbell & Tarasewich, 2004). One way to reduce information overload and enhance content management is through the use of context metadata.
Context metadata is information that describes the context in which a content item was created or received and can be used to aid users in searching, retrieving, and organizing the relevant content automatically. Context is any information that can be used to characterize the situation of an entity. An entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and the applications themselves (Dey, 2001). Some types of context are the physical context, such as time, location, and date; the social context, such as social group, friends, work, and home; and the mental context, which includes users’ activities and feelings (Ryan, Pascoe, & Morse, 1997; Dey, Abowd, & Wood, 1998; Lucas, 2001). By organizing and annotating the content, we develop a new way of managing it, while content management features are created to face efficiently the usability challenge. Context metadata helps the user find the content he needs by enabling single and multi-criteria searches (e.g., find photos taken in Paris last year), example-based searches (e.g., find all the video clips recorded in the same location as the selected video clip), and automatic content organization for efficient browsing (e.g., location-based content view, where the content is arranged hierarchically based on the content capture location and information about the hierarchical relationships of different locations).
DATE, TIME, LOCATION, AND PROXIMITY While context can be characterized by a large number of different types of attributes, the contribution of context attributes to content management is of great importance. We focus on a small number of attributes, which are considered the most important in supporting content management and also have the most practical implementations in real products, such as date, time, location, and proximity (nearby Bluetooth devices). Bluetooth is a short-range wireless technology used
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Content Personalization for Mobile Interfaces
to create personal area networks among user mobile devices and with other nearby devices. The first two attributes, date and time, are the most common in use in a wide range of applications. They are used to organize both digital and analog content, and offer an easy way of searching and retrieving the relevant content automatically. For example, many cameras automatically add the date and time to photographs. Furthermore, the location where content is created is another useful attribute for searching the content (e.g., home, workplace, summer cottage). Mobile devices give users the possibility to create content in many different locations. Users can associate the location with the equivalent content in order to add an attribute to it that will enable them to find it easier. Finally, proximity also plays an important role in content management, as nearby Bluetooth devices can provide information both in social and physical context. While each Bluetooth device can be uniquely identified, information can be provided on nearby people by identifying their mobile phones. An example for physical context is the case of a Bluetooth-based hands-free car kit that can be used to identify that the user is in a car.
USABILITY ISSUES AND PROBLEMS The expansion of the dimension of context information in order to include location, as well as proximity context, can be of benefit to users while they are able to store, access, and share with others their own location-based information such as videos and photos, and feel the sense of community growing among them (Kasinen, 2003; Cheverist, Smith, Mitchell, Friday, & Davies, 2001). But when it comes to proximity to be included in context information, the problem of privacy emerges. It appears that users are willing to accept a loss of privacy when they take into account the benefits of receiving useful information, but they would like to control the release of private information (Ljungstrand, 2001; Ackerman, Darrel, & Weitzner, 2001). While context metadata is attached to content, when users share content, they have to decide if they share all the metadata with the content or they filter out all or some part of them. The cost for memory and transmission of metadata, as it is textual information, is not an important factor to influence this decision. When the user receives location and proximity information attached to content, he or she may also find out where and with whom the creator of the content was when the content was created. As a result, both the location of the content creator and the location of nearby people are shared along with the content information. If this information is private, the sharing of it could be considered as a privacy violation. This violation may be ‘multiplied’ if the first recipient forwards the content and the metadata to other users.
However, users seem to be willing to share context metadata attached to content, as it would be convenient if context metadata were automatically available with the content (so that users do not have to add this information manually). Furthermore, it would be very helpful for the recipient if the received content was annotated with context metadata so that the recipient does not have to annotate it manually and be able to manage the content more easily. For example, in the case of image and video content, the filtering of context metadata such as location and people could be useless, since these same items appearing in the image or video can be identified visually from the image content itself. But what is meaningful information to the end user? It seems that users want meaningful information, but they are not willing to put too much effort in creating it, unless this information is expected to be very useful. In the case of location, it would be difficult for users to type the name of the place and other attributes manually, since it would require their time and effort. Thus it would be important if meaningful context metadata, which include the required information, are automatically generated. Proximity information also needs to be meaningful. In this way, meaningfulness is important when attaching information on nearby devices in the form of metadata. If the globally unique Bluetooth device address and the real name of the owner of the device could be connected, this functionality would give meaningful information to the user. It is hard to determine which information is useful, while what is useful information in one situation might be totally useless in another. For example, when looking at photo albums, what is thought to be useful information varies a lot. When one is looking at family pictures taken recently, it is needless to write down the names of the people, since they were well known and discernable. But it is different looking at family pictures taken many years ago: the same people may not be that easily recognizable. It appears that useful information depends on a user’s location, what the information is used for, and in which time span. In order to create meaningful information, users need to put much effort into getting the data, organizing it, and annotating it with context metadata. Ways to minimize their effort and maximize their benefit should be developed.
CONCLUSION The increasing amount of stored content in mobile devices and the limitations of physical mobile phone user interfaces introduce a usability challenge in content management. The efficient management of large amounts of data requires developing new ways of managing content. Stored data are used by applications which should express information in a sensible way, and offer users a simple and intuitive way of
C
Content Personalization for Mobile Interfaces
organizing, searching, and grouping this information. Inadequate design of user interface results in poor usability and makes an otherwise good application useless. Therefore, it is necessary to design and built context-aware applications. Issues of usefulness and meaningfulness in utilizing context metadata need to be further investigated. Usefulness depends on the type of metadata. As far as location and proximity are concerned, it appears that the more time has passed since the recording of the data, the more accurate the information needs to be. Furthermore, in the case of location information, the closer to one’s home or familiar places the data refers to, the more detailed the information needs to be. A main usability challenge is the creation of meaningful context metadata automatically, without users having to add this information manually. There exist many ways for automatic recording of information about a user’s context, but the generated information is not always meaningful. Another field that requires further research is privacy. It seems that users are willing to accept a loss of privacy, provided that the information they receive is useful and they have control over the release of private information. Content management provides users with a safe, easy-to-use, and automated way of organizing and managing their mobile content, as well as retrieving useful information efficiently.
Kim, H., Kim, J., Lee, Y., Chae, M., & Choi, Y. (2002). An empirical study of the use contexts and usability problems in mobile Internet. Proceedings of the 35th Annual International Conference on System Sciences (pp. 1767-1776). Ljungstrand, P. (2001). Context-awareness and mobile phones. Personal and Ubiquitous Computing, 5, 58-61. Lucas, P. (2001). Mobile devices and mobile data—issues of identity and reference. Human-Computer Interaction, 16(2), 323-336. Ryan, N., Pascoe, J., & Morse, D. (1997). Enhanced reality fieldwork: The context-aware archaeological assistant. In V. Gaffney, M. v. Leusen, & S. Exxon (Eds.), Computer applications in archaeology. Schilit, B., & Theimer, M. (1994). Disseminating active map information to mobile hosts. IEEE Network, 8(5), 22-32. Schilit, B., Adams, N., & Want, R. (1994). Context-aware computing applications. IEEE Proceedings of the 1st International Workshop on Mobile Computing Systems and Applications, Santa Cruz, CA, (pp. 85-90).
REFERENCES
Sorvari, A., Jalkanen, J., Jokela, R., Black, A., Kolil, K., Moberg, M., & Keinonen, T. (2004). Usability issues in utilizing context metadata in content management of mobile devices. Proceedings of the 3rd Nordic Conference on HumanComputer Interaction, Tampere, Finland, (pp. 357-363).
Ackerman, M., Darrel, T., & Weitzner, D. J. (2001). Privacy in context. Human Computer Interaction, 16, 167-176.
KEY TERMS
Brown, P. J. (1996). The stick-e document: A framework for creating context-aware applications. IFIP Proceedings of Electronic Publishing ’96, Laxenburg, Austria, (pp. 259-272). Brown, P. J., Bovey, J. D., & Chen, X. (1997). Context-aware applications: From the laboratory to the marketplace. IEEE Personal Communications, 4(5), 58-64. Campbell, C., & Tarasewich, P. (2004). What can you say with only three pixels? Proceedings of the 6th International Symposium on Mobile Human-Computer Interaction, Glasgow, Scotland, (pp. 1-12). Cheverist, K., Smith, G., Mitchell, K., Friday, A., & Davies, N. (2001). The role of shared context in supporting cooperation between city visitors. Computers &Graphics, 25, 555-562. Dey, A. K., Abowd, G. D., & Wood, A. (1998). CyberDesk: A framework for providing self–integrating context-aware services. Knowledge Based Systems, 11(1), 3-13. Dey, A. K. (2001). Understanding and using context. Personal & Ubiquitous Computing, 5(1), 4-7. Kaasinen, E. (2003). User needs for location-aware mobile services. Personal Ubiquitous Computing, 7, 70-79.
Bluetooth: A short-range wireless technology used to create personal area networks among user devices and with other nearby devices. Content Management: Ways of organizing and annotating content in order to retrieve and search it more efficiently. Context: Any information that can be used to characterize the situation of an entity. Context Metadata: Information that describes the context in which a content item was created or received. Entity: A person, place, or object that is considered relevant to the interaction between a user and an application, including the user and the applications themselves. Location: The place where content is created by the user. Usability: The effectiveness, efficiency, and satisfaction with which users can achieve tasks in the environment of mobile devices.
Category: Mobile Software Engineering
Content Transformation Techniques Ioannis Antonellis Research Academic Computer Technology Institute, Greece University of Patras, Greece Christos Bouras Research Academic Computer Technology Institute, Greece University of Patras, Greece Vassilis Poulopoulos Research Academic Computer Technology Institute, Greece University of Patras, Greece
INTRODUCTION The expansion of the Web is enormous and, more and more, people everyday access its content trying to make their life easier and their informational level complete. One can realize that lately the advances in computers are such that many appliances exist in order to offer to its users the chance to access any type of information. The use of microcomputers, such as PDAs, laptops, palmtops, mobile phones and generally mobile devices, has lead to a situation where a way had to be found in order to offer to the users the same information as if they had a normal screen device. Almost all the mobile devices offer “Web-ready” functionality, but it seems that few of the Web sites are considering offering to the mobile users the opportunity to access their pages from the mobile devices. On the one hand, the widespread use of mobile devices introduces a new big market and many chances for research and development. On the other hand, the use of small screen devices introduces a basic constraint both to the constructors of the devices and to the users: the small screen limitation. This is making difficult for the users to establish a mental model of the data, often leading to user disorientation and frustration (Albers & Kim, 2000). Many other restrictions have to be taken under consideration when using small devices, especially the low resolution, the amount of the memory, and the speed of the processor. Additionally, when using such devices the users are often in places with distractions of noise, interruptions and movement of the handheld device (Jameson et al., 1998). Many companies exist in order to offer to the users of small screen devices the opportunity to access Web pages by doing syntactic translation (AvantoGo, DPWeb, Palmscape, and Eudora). Syntactic translation recodes the Web content in a rote manner, usually tag-for-tag or following some predefined templates or rules. This method seems to be successful especially for the devices that have graphical
display. But, in order to achieve this, the Web pages are scaled down and small devices like mobile phone (very small screen and low resolution) are problematic. This happens because either the graphics are too small or the letters and links cannot be explored. Another major problem of the use of small screen devices is that users often migrate from device to device during a day and they demand to be able to work in the same way whether they work on their personal computer or their mobile phone. This is the main issue that is going to be analyzed in this article: the way of migrating data from device to device without damaging the integrity of the data and without distracting the user. Migration is the process of taking data originally designed for display on a large screen and transforming it to be viewed on the small screen (Jameson et al., 1998). The main techniques that exist and are used for data migration are direct migration, data modification, data suppression and data overview. The first one, direct migration, is a very simple. The data are sent directly to the small screen device and the user navigates to the data by scrolling horizontally and vertically on the page. The second method is more complicated and data is shortened and minimized in order to be viewable in a small screen device. Data suppression technique removes parts of the data and presents parts of them and the latest technique is based on the focus and context model (Spence, 2001). All the aforementioned techniques are useful and any of them can be used efficiently for different types of data. This is a difficult part for the construction of the small screen devices. The constructors of the devices cannot include all the implementations of the techniques or, even if they do, the user has to be asked which one to choose or try the different implementations while viewing a source of data. The differences between the aforementioned techniques are focused on the quality of the information shown to the user and the range of information that is shown. This means that
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
C
Content Transformation Techniques
in some techniques the quality of the information shown is high but the amount of information shown is quite poor. One can think that the quality of information is more important while another can think that the amount of information is more important. This is a question that cannot be answered simply. What we can safely note is that the answer depends on the type of data that we want to present to the users. The rest of this article is structured as follows. In the next section we present the efforts of some companies that offer to the users of small screen devices the opportunity to access Web pages by doing syntactic translation. The first method of transforming information, its use, its advantages and disadvantages are presented to the third section. The fourth section presents the data modification technique and how it is implemented, and the fifth section the data suppression technique. The next section covers the issues concerning data overview technique and the last section presents a summarization and general overview of the techniques.
DIRECT MIGRATION TECHNIQUE The most simple and most often used technique is the direct migration technique. It is used mostly for Web pages and its scope is to send to the users exactly the same data regardless of the device in use. The users are free to interact with the data and they are actually responsible for making themselves comfortable with the amount of data that they are presented. We cannot say that it is a user-centric technique but it is very easy to be implemented, very fast and does not require much effort either for machine or human. The main problem, which is actually a failure of the technique, is that it produces data that needs horizontal scrolling in order to be accessed and that way the user is much distracted. Some additional techniques are used together with the data migration technique in order to reduce or remove the horizontal scrolling problem. The additional technique is mainly the wrapping technique, which removes the horizontal scrolling by putting the extra data under the main page that is shown to the small screen. The problem is not solved but it becomes minor, because it does not lessen the amount of data but transforms the horizontal scrolling to vertical. Another additional technique requires duplicate creation of the data. It is used very often for Web sites and the method is creating two kinds of pages for the same data: one for large screen devices and one for small screen. Surely, this technique has major problems. One is that someone has to create two totally different pages for the same content. The other and more crucial problem is the size of the World Wide Web and the fact that almost nobody has made any effort to create two types of Web pages makes the technique difficult to be applied. Research has shown that the users react better when they are confronting vertical scrolling rather than horizontal 0
(Nielsen, 1999). However even vertical scrolling—generally any kind of scrolling—affects negatively the completion of any task (Albers & Kim, 2000; Dyson & Haselgrove, 2001; Jones et al., 1999). The above implies that this technique can be suitable only for situations where the user just wants to access and read some kind of information and the interaction level between the user and the data remains low. Summarizing, we can say that this technique is very suitable for short text, sequential text, lists and menus that can be displayed within the width constraints of small screens (the impact of migration). It is not recommended to be used when the data include big tables and images (big, high resolution) because these types of data add horizontal scrolling that cannot be transformed.
DATA MODIFICATION In this section we will analyze the second method for data migration, which is the data modification technique. Its main idea approaches the direct migration technique, but the data modification technique has countered the problem of big images and tables. When the data are to be presented to a small device, the size of the images, tables and lists is reduced and some parts of the text are summarized. In this way the users can save in download time and device memory (Mani, 2001). The text summarization is the difficult part of the technique and it introduces a whole new theme for discussion. Many approaches have been proposed (Buyukkokten et al., 2000; Fukushima, 2001, Mani, 2001; Amitay & Paris, 2000). Some of them require a human expert to create the summaries while some others are based on machines. The data that is presented to the users is a reduced form of the actual data. The user has the option to scroll vertically through the data that he comes up with. He can also select a part of the reduced data in order to “open” in another page of his small screen device the real text, which is hidden behind. This procedure can be algorithmic. When data are presented in this way to the user, then the procedure is to read the summarized, reduced data, select a specific topic that suits the user’s needs, read the whole data that is hidden behind the summarized and then go back. The procedure then starts from the beginning. We can say that this technique is very similar to the aforementioned direct migration technique but it goes one step further. It is used mainly for Web browsing where the data are already reduced and offer the user a style of navigation. The summarization that is included, whether it is for images (lower size, resolution) or text (summary), is very helpful for the end-user as it lessens the scrolling either vertical or horizontal. Actually this method does not have horizontal scrolling at all except for some specific, very rare conditions (very large images or tables).
Content Transformation Techniques
Summarizing, we can say that this technique is very useful when users are determined of the information and can easily understand what they are looking for, from a summary of text or simple keywords. It cannot be useful for very specialized texts with difficult and mannered terminology. In general, the summaries have to be very specific and represent accurately the meaning of the text. The main problem of all the summarization techniques is that they do not succeed very often and cannot replace numerical data like financial information, weather information and dates. If one can think that some users want their small screen devices for accessing their bank accounts, watching the weather in a place that they visit or finding the financial exchange then this technique cannot be recommended.
DATA SUPPRESSION As we are able to figure out from the name of this technique, what it actually does is to remove parts of the data that “seem” to be unimportant. What is presented to the user is the basic frame of the data. Displaying only skeleton information can simplify navigation and may reduce disorientation (Spence, 2001). The data is not removed randomly, but there exist several techniques that help in this direction. Some methods for suppressing data is to select only some of the keywords (that are produce from text summarization), present only a specific number of words from each sentence or Z-thru mapping that imposes selective display (Spence, 2001). This approach is very similar to the previous but it seems to be more compact. Very few data is presented to the user and most of the time there is no scrolling at all. The absence of scrolling has advantages and disadvantages. When there is no scrolling the user is not distracted from completing his task, but no scrolling means that the data is extremely reduced in order to fit the screen and it may be difficult for a simple user to locate the information he/she wants. Navigating through the data in this technique is like a file system. The user has a list of words (like a folder) for each amount of data and by selecting an element of the list the information is expanded (files, subfolders) and shown to the user. Every time the user is able to return to the starting frame of data and start exploring from the beginning. Like the previous, this technique has applications where the users know the exact information that they are looking for and they can figure it out from just a heading or a set of keywords. Searching through this type of data is almost impossible because the little amount of data that is presented is often not representative of the data that it comes from. However it is very useful for browsing through news portal when just a title or part of the title is enough for the user to understand the meaning of the whole article. It is used for
structured data, which include information hierarchically structured. Sequential data with little or no structure could be less compatible to manipulate into categories for suppression (impact of migration).
DATA OvERvIEW The last technique that is used for data migration is data overview technique. In reverse to the aforementioned techniques, which reduce parts of the data, this technique creates an overview of the whole data and presents it to the users. The whole data is minimized and the whole information is presented to the user minimized in order to fit the small screen of the device. It is based on the “focus and expand” method. When the user points a specific set of data that is contiguous then it is expanded and shown bigger to the screen in order to fit the screen and be readable. The approach makes it easier for the users to access at once a very large amount of data without losing or not seeing any part of it and, in this way, the disorientation is lessened (Spence, 2001; Storey et al., 1999). Some methods that are used concerning the data overview technique are: • • • •
Focus and context (Spence, 2001; Buyukkokten et al., 2000; Bjork, 2000) Fisheye Techniques (Spence, 2001; Storey et al., 1999) Zoom and pan (Good et al., 2002; Spence, 2001) Content lens (Dieberger et al., 2002)
In general, the technique seems to be problematic as the user is presented with a large amount of data in a small screen. The data is shrunk in order to fit the screen and may be difficult for the user even to see it and figure out what he is looking for. Movements while using a small screen device could create further distortion, or could make it difficult to discern what has been distorted (MacKay & Watters, 2003). The nature of this approach produces both positive and negative points for the end-users. The point that the user is presented the whole information can be both positive and negative depending on the amount of data. However, it is very useful for the users to have full observation of the information they are looking for. The navigation is easy and is based at presenting in large the parts that are focused from the user, but the user can focus only on a part of information and he is not able to combine parts of data. In general this method seems to be the best when the information that is accessed by the user includes large images, big tables, maps, graphs and in general everything that a “focus” method cannot distort but help.
C
Content Transformation Techniques
OvERvIEW OF THE TECHNIQUES In the previous sections we have discussed and analyzed the most common methods for data migration from large screen displays to small screen. As we can obviously see, each method has its advantages and disadvantages making difficult the selection of only one of them in order to cope with every type of data. Direct migration cannot preserve scrolling and it is the fastest and easiest way to present data that are for reading. Its simplicity is its power but we can admit that is not user friendly. Data modification technique solves many problems of the previous technique but still scrolling is an issue. At least paging of the data is preserved and the user can see a large part of the information in only one screen. The matter that rises from this technique is the summarization of the information, which may be distracting or not useful depending on the type of information. It could be seen as a good method for Web browsing. Data suppression goes one step further than the previous technique by removing parts of data and summarizing the rest. It is named as the best method for browsing news portals where just the keywords of a news title can represent successfully the whole article. It is very weak for textual data and for searching, as it provides in a hierarchic manner only some keywords and often distracts a user that does not know exactly what he is looking for. Data overview has a different angle of view than the three previous methods. It is based on the idea “focus and expand” and the philosophy is to present to the user all the information. When the data include large images, big tables and graphs, data overview is the best method for migrating data because it does not lessen or break into many pages all this information, which is by nature connected. On the contrary, when the user wants to read a text or browse in a big portal then this technique seems to be weak, as it provides to the user all the information in one screen and the data are often unreadable. Summarizing, all the techniques offer to the users the opportunity to access any kind of information through their small screen devices like they would do to big screen ones. It is not fair to select one of them as the best one because each one is created for coping with different types of information and data. A device that could combine the implementation of all the aforementioned techniques could be a solution, but the complexity of modern life would prevent us to permute to the users the effort of data migration and thus make modern life more complex.
REFERENCES Albers, M. J., & Kim, L. (2000). User Web browsing characteristics using Palm handhelds for information retrieval.
In Proceedings Of IPCC/SIGDOC Technology & Teamwork (pp. 125-135). September, 2000, Cambridge, MA: IEEE. Amitay, E., & Paris, C. (2000, November). Automatically summarizing Web sites: Is there a way around it? In Proceedings of the 9th Internet Conference on Information and Knowledge Management (pp. 173-179). McLean, VA. AvantGo, Inc. (n.d.). AvantGo. Retrieved from http://www. avantgo.com Bjork, S. (2000, May). Hierarchical flip zooming: Enabling parallel exploration of hierarchical visualization. In Proceedings of the Working Conference on Advanced Visual Interfaces (pp. 232-237). Palermo, Italy. Buyukkokten, O., Garcia-Molina, H., & Paepcke, A. (2001). Seeing the whole in parts: Text summarization for Web browsing on handheld devices. Retrieved from http://wwwconf. ecs.soton.ac.uk/archive/00000067/01/index.html Dieberger, A., & Russell, D. M. (2002, January). Exploratory navigation in large multimedia documents using context lenses. In Proceedings of 35th Hawaii International Conference on System Sciences (pp. 1462-1468). Big Island, Hawaii. Digital Paths LLC. DPWeb. Http://www.digitalpaths.com/ prodserv/dpwebdx.htm Dyson, M., & Haselgrove, M. (2001). The influence of reading, speed and line length and effectiveness of reading from screen. International Journal Human Computer Studies, 54(4), 585-612. Fukusima, T., & Okumura, M. (2001, June). Text summarization challenge: Text summarization evaluation in Japan. In Proceedings North American Association for Computational Linguistics (pp. 51-59). ittsburgh, Philadelphia, Association of Computational Linguistics. Good, L., Bederson, B., Stefik, M., & Baudisch, P. (2002). Automatic text reduction for changing size constraints. In Proceedings of Conference on Human Factors in Computer Systems, Extended Abstracts (pp. 798-799). April 2001, Minneapolis, MN. ILINX, Inc. (n.d.). Palmscape. Retrieved from http://www. ilinx.co.jp/en/products/ps.html Jameson A., Schafer, R., Weis, T., Berthold A., & Weyrath, T. (1998). Making systems sensitive to the user’s time and working memory constraints. In Proceedings of 4th international Conference on Intelligent User Interfaces (pp. 79-86). December 1998, Los Angeles, CA: ACM Press. Jones, M., Marsden, G., Mohd-Nasir, N., Boone, K., & Buchanan, G. (1999). Improving Web interaction on small displays. In Proceedings of the 8th International WWW
Content Transformation Techniques
Conference. May 1999, Toronto, Canada. Retrieved from http://www8.org/w8-papers/1b-multimedia/improving/improving.html MacKay, B., & Watters, C. (2003, Winter). The impact of migration of data to small screens on navigation. IT&Society, 1(3), 90-101. Mani, I. (2001, October). Text summarization and question answering: Recent developments in text summarization. In Proceedings of the 10th International Conference on Information and Knowledge Management (pp. 529-531). Nielsen, J. (1999, December). Changes in usability since 1994. QUALCOMM, Inc. (n.d.). Eudora Internet Suite. Retrieved from www.eudora.com/internetsuite/eudoraweb.html
KEY TERMS Content Transformation: The procedure that leads to changes to content in order to make it interoperable. Migration: Migration is the process of taking data originally designed for display on a large screen and transforming it to be viewed on the small screen. Small Screen Devices: Devices with small screen size where it is difficult to access large-sized blocks of information. Syntactic Translation (of WWW Data): The recoding of the Web content in a rote manner, usually tag-for-tag or following some predefined templates or rules.
Spence, Robert. (2001). Information visualization. New York: ACM Press. Storey, M. D., Fraachia, F., Davic, M., & Hausi, A. (1999, June). Customizing a fisheye view algorithm to preserve the mental map. Journal of Visual Languages and Computing, 254-267.
C
Category: Location and Context Awareness
Context-Adaptive Mobile Systems Christian Kaspar University of Goettingen, Germany Thomas Diekmann University of Goettingen, Germany Svenja Hagenhoff University of Goettingen, Germany
INTRODUCTION
AUTOMATED CONTEXT-AWARENESS
Even though a major part of the industrialized world works with computers on a daily basis and operating computers became much easier since the introduction of graphic interfaces, many users do not experience their computers as work relief, but rather as an increased burden in their everyday lives. One of the most important reasons for this attitude is the unnatural mode of communication between user and computer: the natural interpersonal communication takes information from the communication situation (e.g., the location of the interacting communicators, their personal preferences, or their relationship with each other) implicitly into account. On the other side, despite the development of new interfacessuch as voice and character recognition, which are much closer to interpersonal communication than keyboard terminalscommunication between user and computer is still complex and characterized by little intuition. This is where the objectives of the context-adaptive systems come into play: it is the aim of context-adaptive systems to implicitly collect information about the situation of a system request (context) in order to enable more efficient communication between user and computer. Currently, the concept of context-awareness and context-adaptation has attracted particular attention in the area of mobile communication. This is largely due to the fact that the obligatory requirement of the devices’ portability leads to certain constraints of mobile devices. Small-sized screens, low data processing capacity, and inconvenient ways of navigation and data entry are some examples for these constraints. To overcome these limitations is particularly relevant in the area of multimedia Internet content and therefore requires the communication to be as efficient as possible. One possible option to reduce the resulting problem of presentation and selection of content on mobile devices is to automatically offer the user only those contents relevant for the concrete situation of the service request. Such services require that the computer can sense the particular situation of the service request and autonomously respond with appropriate actions.
In order for real situations to be sensed automatically by computing devices, the situations of these system requests have to be computed as abstract, automatically understood events, so-called contexts. A context is any information that is used to characterize relevant situations of people, locations, or objects that are important for the interaction between application and user (Dey, 2001). A common classification of context information traces back to Schilit, Adams, and Want (1994). They distinguish between the technical context of participating and available computing resources, the social context of users that are involved in the system interaction, and the physical context of the location of the system interaction. Computing context describes available network connections and network bandwidth. Additionally, computing context includes accessible peripherals such as printers, screens, or additional terminals. For example, if a multimedia application knows the user’s available network bandwidth, it is able to adapt a video stream to its capacity and ensures streaming without jerks and with the highest possible resolution. Furthermore, if the multimedia application is aware of a high-resolution display close to the user, it can suggest this device as an alternative screen for displaying the video stream. To be able to identify each other, mobile computing devices must have radio or infrared sensors. A computing device equipped with radio or infrared sensors spans a distinct logical space (a so-called “smart space”) within its sensor coverage. If a device enters another device’s sensor space, the device will identify itself and send its network address or appropriate commands for application requests. The social context contains information about the users involved in the interaction. The user’s information, such as identity, age, gender, and preferences, can be gathered either explicitly from surveys or implicitly by observing the user’s behavior. Surveying each user’s personal characteristics and preferences is the most common form of gathering user information. Most often, surveying user information is directly linked with service registration. Because the provider
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Context-Adaptive Mobile Systems
Figure 1. Components of an agent system Agent
Sensor
Event
Environment
Filter rules
Trigger Actions
has little means to control (usually voluntary) submitted information, information from user surveys is often of poor quality. Additionally, profiles that were gathered from a onetime survey remain static over time. Therefore, apart from voluntary authentication information on a specific Web site, the user can be additionally identified on the basis of his or her behavior. Every Web server has a protocol component that logs every server activity and stores these logs chronologically into different application-oriented protocol files. Analyzing these server protocols, it can be determined what requests for which resources have been completed during a specific unique Web site visit. To link recorded requests with an individual user, the IP address of the user’s device or identification data stored as cookie on the user’s device can be used. Information about the physical context can be collected from a multitude of data sources such as contact-, thermo-, humidity-, acceleration-, torsion-, or photo-sensors, cameras, and microphones. Sensors that are equipped with processors not only collect data but also pre-process this data. Additionally, they can identify specific patterns such as fingerprints. The user’s interaction location is of particular importance for the perception of the physical interaction context. “Location-based-services” are services that take the location of the user into consideration. These services promise to have great chances on the market (Lehner, 2003). The geographical location of a user that is required for services of this kind can either be determined by terminal-locating or by external network-locating. Terminal-locating is carried out by an especially designed device that autonomously executes location measurements. The global positioning system (GPS) operated by the U.S. military is the best-known technology for self-locating. Receiving positioning signals that are beamed down by GPS satellites, a GPS device can accurately triangulate its position for up to 10 meters. Techniques that can position a location or object from a photo are more sophisticated than the GPS system. Yet they strongly resemble human orienta-
tion. Using photo cameras, these methods can calculate the angle and distance to a specific object (such as a building) with the help of a stored three-dimensional model. Network-locating fixes a device’s position using network information. The best-known method for network-locating is the cell identity technique (or cell of origin technique). This technique locates a mobile device within a cellular radio network using the network’s cell-ID. Other networklocating techniques fix a particular position based on time differences of signals arriving at different base stations, the angle of arrival, or attenuation of signals from different base stations. While Schilit et al. (1994) differentiated between three forms of context, Dey (2001) adds the primary and secondary context to these categories (Conlan, Power, & Barrett, 2003). The location, the type of device, the behavior of the user, and the time of inquiry represent primary request contexts. On the other hand, secondary request contexts are composed of a combination of primary context data. By taking the location and other people in the vicinity into consideration to form the secondary social context, the social situation of the user can be determined; for example, the user might be in his office with his colleagues, or out with friends. Furthermore, Chen and Kotz (2000) differentiate between the active and the passive context. The active context determines a change in behavior for the present application (e.g., by defining sections of interest for an adaptive online newspaper). A passive context shows the change in context conditions for the system inquiry only as extra information for the user. For example, a recommendation system can suggest a specific purchase in an online shop, or the user can see his location when using a navigation system.
CONTEXT-RELATED SYSTEM ADAPTION A system is considered context-adaptive if it uses context information to offer its users relevant information or, rather, services (Dey & Abowd, 2000). Generally speaking, a context-adaptive system is characterized by a certain degree of autonomy when fulfilling its tasks. Therefore, adaptive systems are also referred to as agent systems (Russel & Norvig, 2003). Agents are software systems that, with the help of sensors, identify their environment as applicationrelated events and by using pre-defined rules that activate respective events (see Figure 1). The actions that have been triggered by the agent can refer to information or they can contain control commands for other systems. The agent can either perform the action directly and autonomously (active context-awareness), or these actions are only a suggestion to the user (passive context-awareness). A context-adaptive application consists of software objects that are automatically requested when the system senses
C
Context-Adaptive Mobile Systems
a certain event in the systems environment. If the application identifies such a key event, it executes the respective object, or rather, the relevant methods that are part of this object. Usually, the adaptive application needs to combine raw data from its sensory perception of its environment to compute such a key event (e.g., GPS-coordinates, entries in a Web-server’s log file, or identified, closely related network resources) since an individual piece of raw data is only able to hint at the relevant situation. An adaptive traffic information service, for instance, needs to compute information about the current date and time (system clock), the user’s current position (e.g., GPS), the user’s destination, and his or her preferred routes and means of transportation (user profile). In this context, we can distinguish different adaptation methods, depending on the type and extent of combined raw data collected while computing such a key event. Raw sensor data that closely related to the system (such as time, temperature, or the device’s data processing capacity)whose values are, in accordance with the closed-world-assumption, known ex anteare usually directly linked to the respective application-specific event. Therefore, dynamic key-value-pairs are composed, whereby the application-event contains the key and the sensor system contributes the key value (Chen & Kotz, 2000). With regard to raw sensor data that cannot be directly linked to application events, such as user preferences or information about the location, however, the process is different. In this case, values from various sensor systems need to be aggregated. For instance, a standardized locating technique that is able to locate a user in closed rooms and outside does not yet exist. In order to locate users accurately inside and outside of buildings, data from different locating systems need to be combined. Since different locating systems utilize different measures (e.g., distance, angular separation, or geometrical position), these values need to be translated into standardized measures first and then have to be transferred into a joint data model (Hightower, Brumitt, & Borriello, 2002). This process is commonly known as “sensor fusion” (Chen, Li, & Kotz, 2004). Adaptive systems that process contexts whose quality rating is subject to a high level of changeability (e.g., a user’s movement in a room or outside of a building) need a high number of key values to be computed. Generally speaking, it does not make much sense to compose one key-value-pair for each of these values. Instead, artificial intelligence techniques, such as artificial neural networks that have been developed to process knowledge, can be utilized to compute these values. An artificial neural network (ANN) is a system that consists of a multitude of identical, networked computing elements, so-called neurons. ANNs are thus able to simultaneously process extensive and changeable raw data in an efficient manner (Van Laerhoven, Aidoo, & Lowette, 2001).
EXAMPLES FOR CONTEXT-ADAPTIvE MOBILE SERvICES Scholars focusing on mobile systems were among the first to publish research on adaptive systems. Olivetti and XEROX, two of the leading producers of copying machines at that time, were pioneers in this new research area of context-adaptive computing. In the early 1990s, their research facilities introduced the first prototypical adaptive systems. These early applications include automatic call-forwarding systems that are based on where in the office building the person who receives the call is located (Want, Hopper, Falcao, & Gibbons, 1992; Wood, Richardson, Bennett, Harter, & Hopper, 1997), browser software that can be adapted to specific locations (Voelker & Bershad, 1994), as well as location-aware shopping assistants (Asthana, Cravatts, & Krzyzanowski, 1994). These groundbreaking applications are not so much derived from data communication that is supported by mobile radio technology, but from communication that is connected to ubiquitous or, rather, pervasive computing. They primarily use infrared or radio transponders (so-called active badges). Special room sensors are able to detect users, who carry these transponders with them at all times. In later years, scholars developed various kinds of adaptive systems: these new developments include tour guides (Bederson, 1995; Long, Kooper, Abowd, & Atkeson, 1996; Davies, Cheverst, Mitchell, & Friday, 1999), software assistants for conference participants (Dey, Futakawa, Salber, & Abowd, 1999), and field researchers (Pascoe, 1998). In the late 1990s, researchers proposed adaptive service concepts for commercial, content-related mobile radio services. One of the first studies offered a solution to the problem of content recipients’ partially bound attention: this study developed an adaptive screen for GSM terminals which is able to change the screen’s font and brightness in accordance with the room conditions and the user’s activity. As a reaction to the deregulation of the cell-based user locating services in GSM-networks for commercial purposes in 2001, scholars proposed a number of other options for location-specific mobile services (Lehner, 2004). These suggestions include gas station search services that take locating and vehicle data into consideration, and adaptive multimedia applications for cars that are able to switch between different reception options (e.g., GSM or DVB), depending on the specific reception quality (Herden, Rautenstrauch, Zwanziger, & Planck, 2004). In addition to these cell-based location services, researchers have also discussed adaptive services based on GPS (Diekmann & Gehrke, 2003). Furthermore, some authors suggest individualizing concepts that are attuned to the special features of mobile terminals. These concepts would be able to adapt contents to individual preferences. In this context, it has to be distinguished between those individualization concepts that carry out the adaptation on
Context-Adaptive Mobile Systems
the aggregation level (or rather, on a mobile portalSmyth & Cotter, 2003; Kaspar & Hagenhoff, 2004) and those that aim at the individual service (Anderson, Domingos, & Weld, 2001). For the most part, scholars have discussed individualization techniques that are based on explicit information from users and on limited potentialities. Currently, however, research is also trying to find ways to distribute contents to a multitude of different types of mobile devices, a development that has become necessary because of the growing diversity of mobile devices that are equipped with varying hardware and software. One way to achieve this is to use markup transformations that are based on schematic libraries for different devices and standards such as WML, XHTML, und HTML. In order to identify a mobile device, various standards that allow users and providers to exchange the respective configuration during each data communication process have been developed. The best-known standards include the “Composite Capabilities/Preference Profiles” (CC/PP), which was developed by W3C in 2004 or its earlier implementationthe “User Agent Profile” (U-AProf)by the WAP Forum in 2003. After identifying the respective device configuration, it is possible to adapt the syntax, for example on the basis of the style sheet transformation language, XSLT.
Asthana, A., Cravatts, M., & Krzyzanowski, P. (1994). An indoor wireless system for personalized shopping assistance. Proceedings of the IEEE Workshop on Mobile Computing Systems and Applications, Santa Cruz, CA, (pp. 69-74).
CONCLUSION
Dey, A., & Abowd, G. (2000, June). The context toolkit: Aiding the development of context-aware applications. Proceedings of the Workshop on Software Engineering for Wearable and Pervasive Computing, Limerick, Ireland, (pp. 434-441).
Most of the current examples of context-adaptive systems represent isolated solutions that are based on a closed-world assumption. At this point, these systems have little commercial value. This lack of commercial relevance is basically rooted in two main problems that have to be overcome in the future. On the one hand, the development of adaptive systems is very cost intensive. In addition, currently existing solutions are usually based on proprietary data models, which keep them from interacting with different systems and do not allow them to add additional contexts. On the other hand, an adaptive system is dependent on a comparatively large set of personal information that has to be gathered and processed automatically. This causes user concerns about possible abuses of personal data and intrusions of privacy. So far, neither legal measures nor technical control instruments have been able to eliminate users’ apprehensions of permanent surveillance by a “big brother.”
REFERENCES Anderson, C., Domingos, P., & Weld, D. (2001). Personalizing Web sites for mobile users. Retrieved May 31, 2005, from http://www.cs.washington.edu/ai/proteus/www10.pdf
Bederson, B. (1995). Audio augmented reality: A prototype automated tour guide. Proceedings of the Conference on Human Factors and Computing Systems, Denver, (pp. 210-211). Chen, G., & Kotz, D. (2000). A survey of context-aware mobile computing research. Dartmouth Computer Science Technical Report TR2000-381. Retrieved October 31, 2005, from http://www.cs.dartmouth.edu/~dfk/papers/chen: survey-tr.pdf Chen, G., Li, M., & Kotz, D. (2004, August). Design and implementation of a large-scale context fusion network. Proceedings of the 1st Annual International Conference on Mobile and Ubiquitous Systems: Networking and Services, Boston. Davies, N., Cheverst, K., Mitchell, K., & Friday, A. (1999). Caches in the air: Disseminating tourist information in the GUIDE system. Proceedings of the 2nd IEEE Workshop on Mobile Computing Systems and Applications, New Orleans.
Dey, A. (2001). Understanding and using context. Personal and Ubiquitous Computing Journal, 5(1), 4-7. Dey, A., Futakawa, M., Salber, D., & Abowd, G. (1999). The conference assistant: Combining context-awareness with wearable computing. Proceedings of the 3rd International Symposium on Wearable Computers (ISWC ’99), San Francisco, (pp. 21-28). Diekmann, T., & Gehrke, N. (2003). Ein framework zur nutzung situationsabhängiger dienste. In K. Dittrich, W. König, A. Oberweis, K. Rannenberg, & W. Wahlster (Eds.), Lecture notes in informatics, informatik 2003. Innovative anwendungen (Vol. 1, pp. 217-221). Bonn: Gesellschaft fuer Informatik. Herden, S., Rautenstrauch, C., Zwanziger, A., & Planck, M. (2004). Personal information guide. In K. Pousttchi & K. Turowski (Eds.), Mobile Economy: Proceedings of the 4th Workshop on Mobile Commerce, Augsburg, (pp. 86-102).
C
Context-Adaptive Mobile Systems
Hightower, J., Brumitt, B., & Borriello, G. (2002, June). The location stack: A layered model for ubiquitous computing. Proceedings of the 4th IEEE Workshop on Mobile Computing Systems & Applications, Callicoon, New York, (pp. 22-28). Kaspar, C., & Hagenhoff, S. (2004). Individualization of a mobile news servicea simple approach. In S. Jönsson (Ed.), Proceedings of the 7th SAM/IFSAM World Congress, Gothenburg. Lehner, F. (2003). Mobile und drahtlose informationssysteme. Berlin: Springer-Verlag. Lehner, F. (2004). Lokalisierungstechniken und location based services. WISU, 2, 211-219. Long, S., Kooper, R., Abowd, G., & Atkeson, C. (1996). Rapid prototyping of mobile context-aware applications: The Cyberguide case study. Proceedings of the 2nd Annual International Conference on Mobile Computing and Networking (pp. 97-107), White Plains, NY.
Wood, K., Richardson, T., Bennett, F., Harter, A., & Hopper, A. (1997). Global tele-porting with Java: Toward ubiquitous personalized computing. Computer, 30(2), 53-59.
KEY TERMS Agent: A software system that is able to perceive its surroundings as events that are relevant for the application and that, in accordance with previously defined filter rules, causes actions in accordance with its perceptions. Context: Any piece of information that can be used to characterize the situation of a person, a place, or an object in a way that is significant for the interaction between user and application. Context-Adaptive System: An application that changes its behavior in accordance with information about the respective situation.
Pascoe, J. (1998). Adding generic contextual capabilities to wearable computers. Proceedings of the 2nd International Symposium on Wearable Computers, Pittsburgh, PA.
Location-Based Service: A service that takes the location of the user into consideration. A user’s location can therefore be detected by using network-locating or terminal-locating.
Russel, S., & Norvig, P. (2003). Artificial intelligence, a modern approach. NJ: Pearson Education.
Network-Locating: Detects an end device’s position on the basis of network information.
Schilit, B., Adams, N., & Want, R. (1994, December). Proceedings of the IEEE Workshop on Mobile Computing Systems and Applications, pp. 85-90.
Sensor Fusion: Refers to the translation of values from different locating systems using different measures (e.g., distance, angular separation, or geometrical position) into standardized measures and the transfer of these standardized measures into a joint data model.
Smyth, B., & Cotter, P. (2003). Intelligent navigation for mobile Internet portals. Proceedings of the 18th International Joint Conference on Artificial Intelligence (IJCAI-03), Acapulco. Van Laerhoven, K., Aidoo, K., & Lowette, S. (2001). Realtime analysis of data from many sensors with neural networks. Proceedings of the 5th IEEE International Symposium on Wearable Computers 2001, (p. 115). Want, R., Hopper, A., Falcao, V., & Gibbons, J. (1992). The active badge location system. ACM Transactions on Information Systems, 10(1), 91-102.
Smart Space: The logical space that is covered by a computing device equipped with radio or infrared sensors within its sensor coverage. Terminal-Locating: A procedure that enables a specially designed end device to locate its own position.
Category: Location and Context Awareness
Context-Aware Mobile Geographic Information Systems Slobodanka Djordjevic-Kajan University of Nis, Serbia Dragan Stojanović University of Nis, Serbia Bratislav Predić University of Nis, Serbia
INTRODUCTION A new breed of computing devices is taking more and more ground in the highly dynamic market of computer hardware. We refer to smart phones and PocketPCs, which redefine typical usage procedures we are all familiar with in traditional, desktop information systems. Dimensions of this class of computing devices allow users to keep them at hand virtually at all times. This omnipresence allows development of applications that will truly bring to life the motto: “availability always and everywhere.” Hardware and software characteristics of the aforementioned devices require a somewhat modified approach when developing software for them. Not only technical characteristics should be considered in this process, but also a general set of functionalities such an application should provide. Equally important is the fact that the typical user will be on the move, and his attention will be divided between the application and events occurring in his environment. Fundamentally new and important input to mobile applications is constantly changing the user environment. The term that is used most frequently and describes the user environment is a context, and applications that are able to independently interpret a user’s context and autonomously adapt to it are named context-aware applications. Recent developments in wireless telecommunications, ubiquitous computing, and mobile computing devices allowed extension of geographic information system (GIS) concepts into the field. Contemporary mobile devices have traveled a long way from simple mobile phones or digital calendars and phonebooks to powerful handheld computers capable of performing a majority of tasks, until recently reserved only for desktop computers. Advancements in wireless telecommunications, packet data transfer in cellular networks, and wireless LAN standards are only some of technological advancements GIS is profiting from. This mobile and ubiquitous computing environment is perfect incubation grounds for a new breed of GIS applications, mobile GIS. Advances in mobile positioning have given a
rise to a new class of mobile GIS applications called location-based services (LBS). Such services deliver geographic information and geo-processing services to the mobile/stationary users, taking into account their current location and references, or locations of the stationary/mobile objects of their interests. But the location of the user and the time of day of the application’s usage are not the only information that shapes the features and functionalities of a mobile GIS application (Hinze & Voisard, 2003). Like other mobile and ubiquitous applications, mobile GIS completely relies on context in which the application is running and used. The full potential of mobile GIS applications is demonstrated when used in the geographic environment they represent (Raento, Oulasvirta, Petit, & Toivonen, 2005). Thus, development of mobile GIS applications requires thorough analysis of requirements and limitations specific to the mobile environment and devices. Practices applied to traditional GISs are usually not directly applicable to mobile GIS applications. Limitations shaping future mobile applications, including mobile GISs, are ranging from hardware limitations of client devices to physical and logical environment of the running application. Considering the fact that mobile applications are used in open space and in various situations, the ability of the application to autonomously adapt itself to a user’s location and generally a user’s context significantly increases the application’s usability. Regardless of the type of LBS and mobile GIS application, the part of the system that is handling context is fairly independent and can be separately developed and reused. The proper management of contextual data and reasoning about it to shape the characteristics and functionalities of mobile GIS applications leads to a full context-aware mobile GIS. The second section presents concepts of mobile GISs and context awareness, and the principles of how context can be incorporated into traditional GIS features adapted to mobile devices. Data structures and algorithms supporting context awareness are also given. The third section presents GinisMobile, a mobile GIS and LBS application framework
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
C
Context-Aware Mobile Geographic Information Systems
developed at Computer Graphics and GIS Lab, University of Nis, which demonstrates the concepts proposed in this article. The last section concludes the article, and outlines future research and development directions.
CONTEXT AWARENESS IN MOBILE GIS Even though the concept of mobile GIS is in its infancy, technologies that were prerequisite for development of this niche of GIS applications are today widely available and well known to GIS developers. It is reasonable to expect that there are prototypes available demonstrating all the advantages mobile GIS offers to field fork personnel. ESRI, as one of the leading companies in the GIS field in its palette of products, offers a mobile GIS solution targeting the PocketPC platform. It is called ArcPad (http://www.esri. com/software/arcgis/ bout/arcpad.html). It is a general type of mobile GIS solution with open architecture allowing easy customization and tailoring according to a specific customer’s needs. It therefore offers a set of basic GIS functionalities and tools that are used to extend application with functionalities needed for specific usage scenarios. ESRI bases its ArcPad on four basic technologies: mobile computing device (PocketPC), basic set of spatial analysis and manipulation tools, global positioning system (GPS), and wireless network communication interface. Basic GIS functionality understandably supported by ArcPad is geographic maps visualization in the form of raster images. In order to avoid the need for maps conversion into some highly specialized proprietary raster map format, ArcPad supports usage of all of today’s widely used raster image formats, like JPEG, JPEG 2000, and BMP, as well as MrSID, which is common in GIS applications. Thematically different maps in the form of raster images can be grouped into layers. Apart from raster type, layers can also contain vector data. Also, standard vector type data formats are supported, most importantly the shapefile format. That is the most common vector data format in use in GIS today and is also well supported by other ESRI GIS software like ArcInfo, ArcEditor, ArcView, ArcIMS, and others. Other optimizations which enable sufficient speed in handling spatial data include spatial indexing schemes. Spatial indexing significantly increases speed of spatial objects visualization and search, especially on portable devices with limited processing power. Indexes are prepared on other desktop-type ESRI applications, and afterwards are transferred to a mobile device and used by ArcPad. In order to support usage of ArcPad throughout the world, a majority of map projections are included. ArcPad is conceived as an integral part of the ESRI GIS platform consisting of other products, so there is the possibility of ArcPad functioning as a client for ArcIMS or Geography Network (http://www.geographynetwork.com/). Data is transferred to ArcPad using TCP/IP protocol and 0
any sort of packet-based wireless networking technology (wireless LAN, GSM, GPRS, EDGE, 3G, etc.). Possibly the strongest advantage of ArcPad is its extensibility and adaptability. Forms used for thematic data input and manipulation are created and customizes independently using ArcPad Studio and Application builder development tools. Application toolbars can be adapted to specific user needs. More importantly, specific interfaces can be developed and added to ArcPad, enabling it to acquire data from different database types and sensors (GPS location devices, laser rangefinders, magnetic orientation sensors, etc.). One academic project that encompasses the development of mobile GIS is “Integrated Mobile GIS and Wireless Image Servers for Environmental Modeling and Management,” developed at San Diego State University (2002). The project includes an integrated GIS platform where, in the field, data collection must be performed using a mobile GIS client platform. Effectiveness of the developed system is tested in three different services: campus security, national park preservation service, and sports events. The development group’s decision was not to develop a mobile GIS solution from scratch, but to upgrade and customize ArcPad. Similarly to other mobile GIS solutions, this project is based on modified client/server architecture. Fieldwork personnel are using a PocketPC device with a customized ArcPad version installed. Customization includes components developed specifically for testing on campus. PocketPC is connected with an external GPS device, and therefore it has constant access to user location information. Considering wireless communications, campus grounds are covered with a wireless LAN, and all client PocketPCs are equipped with WLAN adapters. The server side of this system includes a typical set of servers and tools from ESRI including ArcIMS and ArcGIS. When this system is employed by the campus security service, field units use mobile GIS components to locate a reported incident location more easily and swiftly. Mobile GIS is also used to report new incidents to central. Following report-in, information about a new event taking place is momentarily available to all units. Therefore, reaction time is shortened and all patrolling units within campus are synchronized more easily. Demonstration use case shows the field unit receiving a warning about a fire reported at the specified site. The closest field unit is being notified. Using the campus WLAN, the central ArcIMS server is contacted and a map of that part of the campus is acquired, as well as blueprints of buildings endangered by fire. The central server also contains thematic data about the estimated number of people in these buildings, evacuation plans, and similar information. Simultaneously, units on site can update fire reports with more detailed information and therefore shorten response time of other units enroute. The ArcPad application customized for this use and being used in this scenario is shown in Figure 1.
Context-Aware Mobile Geographic Information Systems
Figure 1. Mobile GIS implemented in San Diego State University campus security
Besides the location of the user, contemporary mobile GIS applications, such as the one previously described, lack the support for context awareness. Such support must be developed and integrated into the basic framework or platform on top of which the mobile GIS application is developed. But first we must define the context and basic principles of context awareness. In interpersonal communication, a significant amount of information is transmitted without explicit communication of such information. If we take verbal communication as an example, nonverbal signs will significantly influence the completeness of verbally communicated data. We are referring to facial expressions, body posture, voice tone, and nearby objects and persons included in the past history of communications. All this is helping the process of interpretation of verbally transmitted data. In a typical human-machine communication, there is very little context information available in a form that can be interpreted by machine. Therefore our first step should be to define the context. No matter how obvious this may seem, the definition of the context influences significantly all the decisions in the further process of context-aware application development. Dey and Abowd (2000) give a relatively abstract definition of context influenced by their work on “context toolkit” architecture: We define context as any information that can be used to characterize the situation of an entity, where an entity can be a person, place, or physical or computational object. Schilit and Theimer (1994) give a very concrete context definition which is therefore rather local in its application: Context refers to location, identity of spatially nearby individuals and objects and changes that are relevant to aforementioned individuals and objects. Summarizing numerous definitions of context, we can notice three aspects of context that are standing out:
•
•
•
C
Technical Characteristics of the Environment: Hereby we are mainly referring to technical characteristics of the client device, processing power, available memory capacity, display characteristics, as well as characteristics of network connections available to the device (bandwidth, latency, price, etc.). Logical Characteristics of the User’s Environment: This group contains geographic location, identity of individuals and objects nearby, and general social situation. Physical Characteristics of the User’s Environment: This group contains levels of noise, light, and movement parameters (speed, direction, etc.).
In the process of context modeling and management, the system can use information that is both automatically collected or manually entered by the user. Although the first approach is attractive and seems to be the only true manner of handling contextual data, we believe that manual input should not be excluded. Also, some characteristics of the context (e.g., user preferences, history, and predictions of actions) are much more easily acquired by manual input at the current level of advancements in context management algorithms. The important step in development of a context-aware LBS and mobile GIS is to define the set of functionalities the application should provide to the user, implicitly or explicitly. Numerous types of contextual information produce adequately numerous potential functionalities. We can group them as follows: •
Display of Information and Services: In order to reduce user workload, the system adjusts the set of offered information and functions according to detected and deduced environment of the user. For a typical mobile GIS, a section of the map surrounding the current user’s location is displayed. According to the user’s speed and heading, the central point of the map view is chosen and the speed vector displayed. Also, font and color scheme are adjusted to the situ
Context-Aware Mobile Geographic Information Systems
Figure 2. XML scheme describing profile
•
•
ation the user is in (e.g., the user is steering a vehicle at night). Automated Execution of Commands: An example would be a navigation GIS application that detects the user has missed the intersection and automatically initiates rerouting to find the new shortest path to destination. Storage of Contextual Information: Potential use of stored contextual information would be to enable application to autonomously extract user preferences from previous actions using data mining techniques.
The user context that is of interest in LBS and mobile GIS applications is classified into specific classes. Each class of contextual information is assigned a context variable. Often, in other papers published by researchers in this field, authors have noticed a hierarchical structure of context information, so some sort of graph structure is used for context representation (Meissen, Pfennigschmidt, Voisard, & Wahnfried, 2004). Since one class contains contextual information of various levels of generality, the most appropriate data structure for representing contextual information is directed acyclic graph. This data structure is the closest match to human cognition of structure and connections existing within a context data class. Another advantage of the hierarchical context model is the possibility to narrow
down the choice of possible actions induced by a detected context. In this manner, a set of rules used by the rule-based expert system is kept to a minimum of candidate rules (Biegel & Cahill, 2004). The rule-based expert system is used to perform generalization of raw contextual data acquired from sensors and contained in leaf nodes. In this manner a “vertical” structure within each context class is built. As an example of a rule-based expert system that is widely used in literature, we have opted for the C Language Integrated Production System (CLIPS, 2006). The main advantage of CLIPS in our case is the existence of jCLIPS, a library that enables Java programs to use the CLIPS engine, embedding it in a Java code (jCLIPS, 2005). The typical context data flow path in a context-aware application is as follows: raw data is collected by connected sensors, and the software interface associated with each of the sensors converts the data into facts and stores the facts into the expert system. After each modification of a rule set, CLIPS executes a generalization process and generates the new facts at higher levels of generality. Also, the possibility of performing action is tested. The action is represented by forming an XML file containing configuration parameters for the client device. This XML file generally describes a profile that a context-aware application will use as a response to a context change. The XML scheme of such a profile is shown in Figure 2.
Context-Aware Mobile Geographic Information Systems
Figure 3. The XML file representing the user profile according to the user’s context
C
1000 700 850 0 Gas_Stations Fast_Food_Restaurants 200 320 4092 12 Courier Normal Yellow 24 Arial Bold Red Blue
The particular profile transferred to the client is represented as an XML file described in Figure 3.
GinisMobile: CONTEXT-AWARE MOBILE GIS FRAMEWORK To support efficient development of mobile GIS solutions, GinisMobile as a component mobile GIS framework is developed in the Laboratory for Computer Graphics and Geographic Information Systems (CG&GIS) Lab at the University of Nis. This framework represents the result of continuous development and advancement of GIS frameworks intended for rapid development of desktop and Web GIS applications (Ginis and GinisWeb). Well-tested GIS concepts of Ginis and GinisWeb have been supplemented with application components from the mobility domain (Predic & Stojanovic, 2004).
Mobile GIS architecture is somewhat similar to architectures used for Web or WAP GIS solutions (Fangxiong & Zhiyong, 2004). An extended client/server approach is used as the basis. The most frequent modification of this architecture, which is used in commercial solutions, is the three-tier model. It consists of the presentation layer, and the GIS logic layer including mobile database components and external GIS services layer. The third layer encompasses GIS services like Web map service (WMS) and Web feature service (WFS) (OGC, 2003). The main advantage of this approach is clear separation of functionalities into independent modules, easily upgradeable and substitutable. The grouping of layers in the case of mobile GIS is shown in Figure 4. The presentation layer is tasked with information visualization (spatial data in the form of maps and attributes in alphanumerical tabular form). This layer also receives users’ requests and queries, interprets them, and activates corresponding functions in the application logic layer. Using adapted HTTP protocol, the client is supplied with static
Context-Aware Mobile Geographic Information Systems
Figure 4. Layered architecture of mobile GIS User
Stationary database and third party GIS services layer
Presentation layer
GIS functionality
GIS services
Application logic HTTP
Mobile database HTTP
Mobile database and application logic layer
Presentation layer
Interaction
GIS functionality
Stationary database
Other GIS services (WMS, WFS...)
sections of maps in the form of raster images and vector type spatial data which is XML encoded. XML encoding also contains attribute data. All data the user interacts with and can change are classified and transferred in the vector form. All other data, regardless of their form when stored on the server side, are rasterized and transferred to the client in the form of raster images. Since the data’s role is only auxiliary, this is the most appropriate form of visualization. Database component must contain the functionality of partial replication of the data subset that is of interest to the individual user and data synchronization with local data storage located on the mobile device (Huang & Garcia-Molina, 2004). In our usage scenario (mobile GIS), the client possesses significant processing power and memory capacity which can therefore be used for performance improvement of this typical layered architecture. That is the reason the client-side component in the case of mobile GIS is usually named ‘rich client’ (Predic, Milosavljevic, & Rancic, 2005). The mobile database and application logic layer is physically located at the client device according to its significant processing capacities. Application logic, which is also physically located at the client device, contains a portion of GIS functionalities, basic functionalities which can be performed on the client side solely without data transfer to/from the server side. The stationary database and thirdparty GIS services layer (e.g., Web map server) is physically located on the server side. The stationary database contains the complete set of data available. The mobile GIS client is supplied only with a subset of available data, a subset that is of direct interest to the individual client (fieldwork team). Determining the scope and volume of this subset is the task of the GIS functionalities component located in this layer. Other GIS services belonging to this layer (WMS, WFS) are also controlled by GIS functionalities of this layer. WMS is
tasked with supplying raster map segments which are used at the client to form a continuous geo-referenced map. WFS provides data about geographic features in the vector format encoded in Geography Markup Language (GML, 2004). To support development of context-aware LBS and mobile GIS applications, we have developed and integrated context-aware support and components into GinisMobile, an LBS and mobile GIS application framework (Predic et al., 2005). GinisMobile is a mobile extension of GinisWeb, a Web GIS application framework (Predic & Stojanovic, 2004). As such it includes support for management and presentation of raster and vector spatial data, as well as dynamic data about mobile objects (Stojanovic, Djordjevic-Kajan, & Predic, 2005). The first obstacle encountered when developing LBS and mobile GIS applications is a highly constrained mobile client platform. Therefore, these applications are already aware of the hardware characteristics of the device it is running on and able to automatically adapt to it, enabling full utilization of the device’s capabilities. The type of sensors relevant to context-aware LBS and mobile GIS applications are widely available, and either are already integrated in modern devices (GPS sensors, Bluetooth radios, level of light sensors, etc.) or are available as add-on devices connected via PAN (Bluetooth). Each of these sensors implements its internal data format. Therefore, each sensor has a software interface attached to it. Its task is to convert data from the format used internally by the sensor into a format appropriate to the application. Contextual data concerning technical characteristics of the device are accessible directly by the application and therefore do not require a separate software interface. A compiled set of contextual data is encoded according to a defined XML scheme and transferred to the server for analysis and storage. The proposed architecture of GinisMobile, a context-aware LBS and mobile GIS framework, is illustrated in Figure 5. This model requires minimal changes to starting a mobile GIS architecture and minimizes processing requirements on the client side. On the server side, context information is handled separately from the user commands. It is inserted into the rules and facts database, and is analyzed by a rule-based expert system. Every change in the rules and facts database can lead to either insertion of new context information at the higher logical level into the database or entering a new state. Reaching a new state results in picking out a profile that most adequately fits new change in the user’s context. Handling of other spatial data is the same as in Predic and Stojanovic (2005) and will not be further discussed in this article. Profile, packed with static spatial layers (rasterized to a single map layer) and a dynamic map component (e.g., moving objects), is transferred to the client. On the client side, the XML profile is parsed and used to customize the user interface. Rasterized layers are stored on internal cache and displayed. Static objects are presented on the background map according to display settings, profile,
Context-Aware Mobile Geographic Information Systems
Figure 5. Mobile GIS architecture with context-aware support Visuelization
C
Usage hostory of raster map segments
Vector layers
Mobile objects history database
Display style
Segmented raster map
HTTP POST response
User interface
Commands buffer
XML profile
HTTP POST request
Rule based expert system
Sensor interfaces
Context interpretator and integrator
Creation of raster maps (layers)
GIS functionalities
Coordinates conversion Profiles
RDBMS Geospatioal database
Context information database
Creation of GIS commands
Sensors
Server side Client side
and with appropriate symbols. Finally, moving objects are superimposed on the map display. Since raster segments are static in nature and change rarely, we keep a local cache of frequently used segments. This approach speeds up the visualization process significantly. The adopted least recently used (LRU) algorithm is used to keep memory requirements minimal. To test the context awareness support and context-aware components built into the GinisMobile framework, we have developed a mobile GIS application for vehicle navigation and fleet tracking on top of the GinisMobile. The application setting assumes that the sensors connected to the user’s device are able to determine speed and direction, time of a day, and levels of noise and light. The higher level of contextual information is deduced based on basic contextual data. Based on deduced facts and rules within the knowledge-based engine, the server recognizes that the user drives a vehicle during the evening/night hours. According to this information, an appropriate profile is constructed which describes the user interface with night colors. The navigation data are displayed on the screen with appropriate font size according to the speed of the user’s vehicle. Also, appropriate zoom
level is chosen with the user’s location displaced from the view center. In this manner the user is enabled to see more of the map in front of him. Finally, the view contains vector speed as a reference. We assume the existence of a GPS receiver attached to the mobile device since this is a very common type of sensor today. It provides data on a user’s geographic location as well as motion data (speed and direction). Another “sensor” relies on time of day to detect light conditions (day/night). More specifically, a level-of-light sensor that is present on mobile devices available on the market could be used for this purpose for additional accuracy. The role of context detection and interpretation subsystem is to decrease the workload needed to operate a typical vehicle navigation system and therefore increase safety. In the demonstration application the following scenario is employed: a mobile user drives a vehicle in the urban environment during the day and at night. Screenshots are taken at different levels of adaptation that an LBS application has performed autonomously in response to changing user context. Figure 6(a) shows a user traveling at 20km/h
Figure 6. Screenshots taken from the sample mobile GIS application
(a)
(b)
(c) 135
Context-Aware Mobile Geographic Information Systems
Figure 6. Continued
(d)
along a city street. It is worth noticing that the user’s location (indicated by the cross symbol) is decentralized and a velocity vector is drawn on the map view. The map view also includes speed and heading. As the user increases his speed, the font size for displayed motion data (speed and heading) is increased, the amount of map view decentralization is increased, and the velocity vector is updated accordingly. This is illustrated in Figure 6(b). As the speed further increases above a certain threshold (Figure 6(c)), the map view zoom scale is changed (decreased). This, along with additional decentralization of a map view, allows the user to see more of the map laying in front of him, in the direction of the velocity vector. The effects of further increase of speed and change in direction of velocity vector are shown in Figures 6(d) and 6(e). When the application detects night conditions, it switches to using a set of colors customized to night conditions. The map view customized to night conditions is shown in Figure 6(f). This choice of colors minimizes the distraction effect for the user (driver).
CONCLUSION Considering the general trend in development of information technologies and the computer industry in general, we can notice a constant migration into mobile and ubiquitous computing (Hinze & Voisard, 2003). With further development of wireless communication technologies, data stored in heterogeneous distributed databases will be available at any specific instance in time and at any location. The most important beneficiaries of this newly introduced concept of mobile GIS will be professional users whose job descriptions include a lot of field work with spatial data. Considering the current state of the art, we believe that mobile GIS applications are perfect testing grounds for the context-aware concept. Being used in unconstrained free space, context awareness considerably enhances usability of the mobile GIS applications. Hereby, context-aware applications are a super set of location-based services. As this article has
(e)
(f)
stressed, location information is only one class, although very frequently used, of context information. This information will be used to automate many procedures and decisions, and relieve the user of the repeatable and tedious tasks of frequent reconfiguration.
REFERENCES Biegel, G., & Cahill, V. (2004, March 14-17). A framework for developing mobile, context-aware applications. Proceedings of the 2nd IEEE Conference on Pervasive Computing and Communications (Percom 2004), Orlando, FL. CLIPS. (2006). Version 6.24: A tool for building expert systems. Retrieved from http://www.ghg.net/clips/CLIPS. html Dey, A. K., & Abowd, G. D. (2000, April 1-6). Towards a better understanding of context and context-awareness. Proceedings of the Workshop on the What, Who, Where, When and How of Context-Awareness (affiliated with CHI 2000), The Hague, The Netherlands. Fangxiong, W., & Zhiyong, J. (2004, July 12-23). Research on a distributed architecture of mobile GIS based on WAP. Proceedings of the ISPRS Congress, Istanbul, Turkey. GML. (2004). Geography Markup Language (version 3.1.1)—encoding specification. Retrieved from http://portal. opengeospatial.org/files/?artifact_id=4700 Hinze, A., & Voisard, A. (2003, July 24-27). Location- and time-based information delivery. Proceedings of the 8th International Symposium on Spatial and Temporal Databases, Santorini Island, Greece. Huang, Y., & Garcia-Molina, H. (2004). Publish/subscribe in a mobile environment. Wireless Networks, 10(6), 643-652. JCLIPS. (2005). Retrieved from http://www.cs.vu. nl/~mrmenken/jclips/#developer
Context-Aware Mobile Geographic Information Systems
Meissen, U., Pfennigschmidt, S., Voisard, A., & Wahnfried, T. (2004). Context and situation awareness in information logistics. Proceedings of the Workshop on Pervasive Information Management (held in conjunction with EDBT 2004).
Proceedings of the 5th International Workshop on Web and Wireless Geographical Information Systems (LNCS 3833, pp. 168-182), Lausanne, Switzerland. Berlin: Springer-Verlag.
Open GIS® Reference Model. (2003). Version 0.1.2, Open GIS Consortium, Reference Number: OGC 03-040. Retrieved from http://www.opengis.org/specs/?page=orm
KEY TERMS
Predic, B., & Stojanovic, D. (2004, March 8-12). XML integrating location based services with Web based GIS. Proceedings of YU INFO 2004, Kopaonik, Serbia, and Montnegro. Predic, B., & Stojanovic, D. (2005, May 26-28). Framework for handling mobile objects in location based services. Proceedings of the 8th Conference on Geographic Information Science (AGILE 2005) (pp. 419-427), Estoril, Lisbon, Portugal. Predic, B., Milosavljevic, A., & Rancic, D. (2005, June 5-10). RICH J2ME GIS client for mobile objects tracking. Proceedings of the XLIX ETRAN Conference, Budva. Raento, M., Oulasvirta, A., Petit, R., & Toivonen, H. (2005). Context phone: A prototyping platform for context-aware mobile applications. IEEE Pervasive Computing—Mobile and Ubiquitous Systems, (April-June), 51-59. San Diego State University. (2002). Integrated mobile GIS and wireless Internet image servers for environmental monitoring and management. Retrieved from http://map. sdsu.edu/mobilegis/photo_mtrp.htm Schilit, B., & Theimer, M. (1994). Disseminating active map information to mobile hosts. IEEE Network, 8(5), 22-32. Stojanovic, D., Djordjevic-Kajan, S., & Predic, B. (2005, December 15-16). Incremental evaluation of continuous range queries over objects moving on known network paths.
Context-Aware Application: Application that posses the ability to autonomously and independently detect and interpret a user’s environment parameters, and adapts its performance and functionalities according to detected context. Geographic Information System (GIS): Information system that stores, analyzes, and presents data about geographic entities. Geography Markup Language (GML): The XML grammar defined by the Open Geospatial Consortium (OGC) to express geographical features. Global Positioning System (GPS): Satellite-based system that is using radio triangulation techniques to determine geographic position time and speed of user. Location-Based Service (LBS): Service that delivers geographic information and geo-processing services to a mobile/stationary user, taking into account the user’s current location and references, or locations of the stationary/mobile objects of the user’s interests. Web Feature Service (WFS): Web-accessible service that provides data about geographic entities encoded in GML using HTTP protocol. Web Map Service (WMS): Web-accessible service that provides geo-referenced raster map data using HTTP protocol.
C
Category: Location and Context Awareness
Context-Aware Systems Chin Chin Wong British Telecommunications (Asian Research Center), Malaysia Simon Hoh British Telecommunications (Asian Research Center), Malaysia
INTRODUCTION Fixed mobile convergence is presently one of the crucial strategic issues in the telecommunications industry. It is about connecting the mobile phone network with the fixedline infrastructure. With the convergence between the mobile and fixed-line networks, telecommunications operators can offer services to users irrespective of their location, access technology, or terminal. The development of hybrid mobile devices is bringing significant impact on the next generation of mobile services that can be rolled out by mobile operators. One of the visions for the future of telecommunication is for conventional services such as voice call to be integrated with data services like e-mail, Web, and instant messaging. As all these new technologies evolve, more and more efforts will be made to integrate new devices and services. New markets for services and devices will be created in this converged environment. Services become personalized when they are tailored to the context and adapted to changing situation. A context-aware network system is designed to allow for customization and application creation, while at the same time ensuring that application operation is compatible not just with the preferences of the individual user, but with the expressed preferences of the enterprise or those which own the networks. In a converged world, an extended personalization concept is required. The aspects covered include user preferences, location, time, network, and terminal; these must be integrated and the relationships between these aspects must be taken into consideration to design business models. Next-generation handsets are capable of a combination of services available on a personal digital assistant (PDA), mobile phone, radio, television, and even remote control. This kind of information and communications technology and mobile services together form one of the most promising business fields in the near future. The voice average revenue per user (ARPU) is declining, the competition is getting fiercer, and voice over Internet protocol (VoIP) is entering the market with aggressive pricing strategies. Fixed mobile convergence should help in this context by providing converged services to both consumer and small-business users. For telecommunication companies it is now crucial to attempt to identify concrete
applications and services for commercial offerings based on fixed mobile convergence which go beyond the current hype. Market scenarios and business models for such fixed mobile convergence solutions will be required and are therefore valuable for future strategy decisions. This article examines market aspects, user requirements, and usage scenarios to come up with a roadmap and suggestions on how to deal with this matter.
CURRENT AND FUTURE TRENDS In the past, user movement has often implied interruption of service. With the advent of pocket-size computers and wireless communication, services can be accessed without interruption while the entity using the services is moving (Floch, Hallsteinsen, Lie, & Myrhaug, 2001). There is a strong need for seamless access. Convergence has been taking place for years now. A study performed by the European Commission (1997) defines convergence as allowing both traditional and new communication services, whether voice data, sound, or pictures to be provided over many different networks. An excellent example of convergence in the telecommunications industry is the IP multimedia subsystem (IMS). Similar to other emerging industries, fixed mobile convergence is characterized by a continuously changing and complex environment, which creates uncertainties at technology, demand, and strategy levels (Porter, 1980). Porter (1980) asserts that it is possible to generalize about processes that drive industry evolution, even though their speed and direction vary. According to Ollila, Kronzell, Bakos, and Weisner (2003), these processes are of different types and are related to: • • • • •
market behavior; industry innovation; cost changes; uncertainty reduction; and external forces, such as government policy and structural change in adjacent industries.
Each evolutionary process recognizes strategic key issues for the companies within the industry, and their effects
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Context-Aware Systems
are usually illustrated as either positive or negative from an industry development viewpoint. For example, uncertainty reduction is an evolutionary process that leads to an increased diffusion of successful strategies among companies and the entry of new types of companies into the industry. Both of these effects are believed to contribute to industry development with regards to the fixed mobile convergence value Web. The technological uncertainties are usually caused by fast technological development and the battles for establishing standards, which are common in the beginning stages of the lifecycle of a specific industry as a result of a technological innovation (Camponovo, 2002). Concerning demand, regardless of the generalized consensus about the huge potential of fixed mobile convergence, there are many uncertainties about what services will be developed, whether the users are ready to pay for them, and the level and timeframe of their adoption (Camponovo, 2002). While the wireless industry is often cited as an example of a rapidly changing sector, the period from 2001-2005 could (in some respects) be regarded as relatively stable (Brydon, Heath, & Pow, 2006). Mobile operators have made the vast majority of their service revenue from simple voice telephony and text messaging, while their value chain has remained largely undisturbed (Brydon et al., 2006). However, new services, alternative technologies, and an evolving competitive landscape mean that the possibility of substantial industry change over the course of the next five to 10 years cannot be discounted (Brydon et al., 2006). The telecommunication industry has experienced several waves of changes from the introduction of wired telephony to wireless telephony, and it is currently heading towards fixed-mobile convergence. Users become more demanding: a “user-centric” and not “network-centric” approach is needed.
According to Hellwig (2006), many fixed operators lose their market dominance and merge units (fixed and mobile). New technologies and new actors (e.g., VoIP, Wi-Fi operators) coming into the picture are driving the adoption of fixed mobile convergence. The formation of new roles in the communication industryincluding brokers, aggregators, alliances, and cooperationhave further pushed the stakeholders to take aggressive strategies to gain competitive advantage. However, since new roles have been introduced, it is unclear how the market acceptance in the near future will be. Existing business models might not be applicable in the new business environment. The lack of terminal devices at the moment also hinders the diffusion. In markets where there are high levels of fixed-mobile substitution and where broadband penetration and wireless local area network (WLAN) diffusion in the home are accelerating, it is most likely that consumers will be drawn to fixed mobile convergence, provided the cost savings and added convenience of carrying one device are apparent to the consumer (McQuire, 2005). According to the Yankee Group (2005), almost one-third of users make more calls within the home using their mobile phone than their landline. The trend is stronger among younger respondents. Figure 1 shows the number of fixed mobile convergence households in Spain, the UK, and France from 2006 to 2010. The increasing need for a personal communication device that can connect to any type of networka mobile network, IP network, or even public switched telephone network (PSTN)and that supports all voice and text-based communication services drives the development of context-aware systems. The primary objective of the system is to facilitate acquisition, translation, and representation of context information in a structured and extensible form, in order to enable the development and enhancement of functionality of network
,000,000
%
,00,000
0%
,000,000
%
,00,000
%
,000,000
%
00,000
%
0 00
00
00
00
00
Household penetration
Number of FMC households
Figure 1. Number of fixed mobile convergence households in Spain, the UK, and France 2006-2010 (Hellwig, 2006)
Spain UK France
0%
C
Context-Aware Systems
resources, personalized according to each individual’s needs. The secondary goal is to facilitate rapid development and deployment of services and applications through a defined framework, which can maintain interoperability between different services and domains. An example of a context-aware system would be BT’s proposed Context Aware Service Platform (CASP). In June 2004, NTT DoCoMo, BT, and a number of other incumbent operators from around the world formed the Fixed Mobile Convergence Alliance (2006), with the objective of developing common technology standards and low-cost devices for integrated fixed-mobile services. The CASP middleware mentioned is the interpretation and one of BT’s visions for the development of a converged platform. The salient features of the proposed product include: •
•
•
•
•
0
User-Centered Operability: One important requirement for a heterogeneous network environment is the ability to instantaneously optimize services for individual users without the need for them to perform any annoying operations. CASP aims to provide transparent connectivity between users with devices and surrounding communication resources. It is able to recognize users’ situations and environmental information automatically. Ease of Service Provisioning: The proposed platform and generic framework guidelines in respect to security, data integrity, non-repudiation, registration, subscription, and quality of service (QoS) for all services will be made available. It offers standard interfaces for all services which enable easier access to a less complex network, with common operation and management, maintenance and training, as well as a common environment for services development and delivery. Interoperability of Shared Services: The proposed platform provides a common specification for services to guarantee the interoperability between shared services in the communication networks. Specific context information with respect to specific aspects characterizing a service or entity can be expressed in an eXtensible Markup Language (XML)-based instance document. Unified Identity: In a true seamless access communication world, every user or communication object is represented by a unified identity. A session initiation protocol (SIP) address (e.g.,
[email protected]) can be used to uniquely identify a user or communication object even when it moves across different networks or between different devices. By having identity management, it simplifies mobility management, security management, and unified user profile management. Dynamic User Interface (UI) on Shared Device: Through the proposed platform, the user can have a shared device that can connect and interact with
•
the ubiquitous communication objects nearby. Each networked object or entity such as cameras, scanners, printers, video players, and so forth can be represented by a different UI based on its own dynamic profile and thus can react intelligently to events in the communication space. Context-Enabled Adaptive Service: The heterogeneity of the converged networks, in terms of network capacity and terminal capabilities, is expected to cause unpredictable changes of network condition. The traditional QoS mechanisms, which do not take the presence of mobility and seamless connectivity into consideration, are not sufficient to guarantee a stable service. Thus, the use of adaptive services being able to change their settings to adapt to the available network resources is a must. CASP enables dynamic selection of the settings used by multimedia services and applications during a multimedia session based on the context of the surrounding environments.
In the near future, stakeholders in the industry will move towards IP-based transport, call control, and service creation and delivery platform functionality. They will follow and adopt developments in the 3rd Generation Partnership Project (3GPP) (http://www.3gpp.org/), European Telecommunications Standards Institute (ETSI) (TISPAN) (http://www.etsi. org/), and Internet Engineering Task Force (http://www.ietf. org/) Next-Generation Network (http://www.ngni.org/) to support open interfaces and avoid interconnection and cooperation incompatibilities. In addition, the will: (1) support IP-based signaling and addressing, media negotiation, QoS, and security mechanisms; (2) support a very large variety of multimedia, banking, and mobile office applications seamlessly across different networks; and (3) adopt seamlessly to the network characteristics and device used. The players in the industry must also consider entering new positions in the value network by taking on new roles and working in cooperation with other telecommunication companies (Hellwig, 2006). Figure 2 shows how and when future enterprise telephony services will embrace convergence.
USAGE SCENARIOS Cheryl subscribes to an Internet VoIP service to save money on calls to her family and university friends who are now spread around the globe, but since her mobile operator utilizes unlicensed mobile access (http://www.umatechnology.org/) technology, she is now able to enjoy cheaper calls by using her mobile phone and connecting to her home WLAN or public hotspots. Her mobile device also enables a number of rich services that enable her to communicate with her friends via voice, video, as well as text.
Context-Aware Systems
Figure 2. Future enterprise telephony services will embrace convergence
C
Source: Yankee Group 2005, various vendors Enablers
Enablers Enterprise Class Dual-Mode Devices and Capacity Smart Phones VoWLAN
SIPBased
WWAN Supporting VoIP (HSPA+) Converged End-to-End VoIP - WWAN VoIP - Hosted Mobile P Center - Converged Voice and Data
LMSDriven
Dual-Mode Enterprise Mobile Voice over IP - P BPX Connectivity over WLAN and VoWLAN - Possibly UMA Solutions for Small Business/SOHO Single-Mode Mobile PBX - TDMWIP PBX Extension over Cellular - Least Cost Routing of Cellular
Enterprise Value
CircuitBased
Hosted Mobile Centrex - Spring Mobile - TeliaSonera
FMCDriven
RMSDriven
Business Tariffs - Employee Share Plans - G Flat Rate Bundles 00
When she was on vacation in Hawaii recently, she was able to show the pictures she had taken with her mobile device to a colleague while he was talking to her over a VoIP call, and later sent a “wish you were here” video message to her parents. Since all her communications are unified in a single device, Cheryl’s friends and family can always reach her, either by voice, text, instant message, video call, or any other means, while Cheryl can use presence to broadcast her availability to her contacts (such as “in a meeting” or “traveling”), as well as manage the incoming communication depending on the context of what she is doing. Now, she can keep her long-distance bills lower by using VoIP, but since it is in her mobile device, she does not have to be sitting in front of her personal computer (PC) to
Figure 3. Cheryl is a recent university graduate whose work as a researcher means she has a busy travel schedule. She often travels to several places to attend international conferences and workshops. She uses a multi-radio mobile device and a mobile VoIP service subscription to keep in touch with her friends and family whether she is at home or on the move.
00+
use it. And whereas some of the PC-based services Cheryl previously used were cumbersome to set up, and she had separate providers for her telecom services, she now gets all these functionalities in a bundled offering from a single operator, and all technology takes care of itself, working invisibly to her through a user-friendly interface.
CONCLUSION Context-aware systems, when made available to the end users, will be greatly valued by them. Consumers will be in a more interactive environment that could help them to take care of small yet related matters automatically. Any possible devices around them could be used to bridge any services offered, giving them the familiarity they preferred. Users would always have the option to alter the service execution or switch it off anytime they like. Meanwhile from the network perspective, knowing the situation of the network and each network node’s role could enable an adaptive and intelligent network. The capabilities such as self-healing, autonomous utilization optimization, and self-reconfiguration to adapt for changes could also be enabled with context-sensitive service logic. The CASP provides a stable and robust environment for the context-aware service developers and operators. This stable environment is extremely important for them to have the accurate anticipated outcome and have the flexibility on changes. On top of the stable environment, the platform will be assisting the service execution to reduce the complexity for the service creation.
Context-Aware Systems
REFERENCES Brydon, A., Heath, M., & Pow, R. (2006). Scenarios for the evolution of the wireless industry in Europe to 2010 and beyond. Analysys Research. Camponovo, G. (2002). Mobile commerce business models. Paper presented at the International Workshop on Business Models, Lausanne, Switzerland. European Commission. (1997). Towards an information society approach. Green paper on the convergence of the telecommunications, media and information technology sectors, and the implications for regulation. European Union. Floch, J., Hallsteinsen, S., Lie, A., & Myrhaug, H. I. (2001, November 26-28). A reference model for context-aware mobile services. Tromsø, Norway: Norsk Informatikkonferanse. FMCA. (2006). Retrieved from http://www.thefmca.com/ Hellwig, C. (2006). New business and services by converging fixed and mobile technologies and applications. T-Systems. McQuire, N. (2005). Residential fixed-mobile convergence heats up but SME Is next frontier. Yankee Group. Ollila, M., Kronzell, M., Bakos, M., & Weisner, F. (2003). Mobile entertainment industry and culture: Barriers and drivers. UK: MGAIN. Porter, M. (1980). Competitive strategy. New York: The Free Press. Yankee Group. (2005). European mobile user survey. Yankee Group.
KEY TERMS eXtensible Markup Language (XML): A specification developed by the W3C, XML is a pared-down version of the Standard Generalized Markup Language (SGML), designed especially for Web documents. It allows designers to create their own customized tags, enabling the definition, transmission validation, and interpretation of data between applications and between organizations. Internet Protocol Multimedia Subsystem (IMS): A standardized next-generation networking (NGN) architecture
for telecommunication companies that want to provide mobile and fixed multimedia services. It uses a VoIP implementation based on a 3GPP standardized implementation of session initialization protocol (SIP), and runs over the standard Internet protocol (IP). Existing phone systems (both packetswitched and circuit-switched) are supported. Public-Switched Telephone Network (PSTN): The international telephone system based on copper wires carrying analog voice data. This is in contrast to newer telephone networks based on digital technologies, such as the integrated services digital network (ISDN) and fiber distributed data interface (FDDI). Quality of Service (QoS): A networking term that specifies a guaranteed throughput level. One of the biggest advantages of asynchronous transfer mode (ATM) over competing technologies such as frame relay and fast ethernet is that it supports QoS levels. This allows ATM providers to guarantee to their customers that end-to-end latency will not exceed a specified level. Session Initiation Protocol (SIP): An application-layer control protocol; a signaling protocol for Internet telephony. SIP can establish sessions for features such as audio/videoconferencing, interactive gaming, and call forwarding to be deployed over IP networks, thus enabling service providers to integrate basic IP telephony services with Web, e-mail, and chat services. In addition to user authentication, redirect, and registration services, the SIP server supports traditional telephony features such as personal mobility, time-of-day routing, and call forwarding based on the geographical location of the person being called. Unlicensed Mobile Access (UMA): The technology that provides access to global system for mobile communications (GSM) and general packet radio service (GPRS) mobile services over unlicensed spectrum technologies, including Bluetooth and 802.11. By deploying UMA technology, service providers can enable subscribers to roam and handover between cellular networks and public and private unlicensed wireless networks using dual-mode mobile handsets. Voice Over Internet Protocol (VoIP): The routing of voice conversations over the Internet or any other IP-based network. The voice data flows over a general-purpose packet-switched network, instead of traditional dedicated, circuit-switched telephony transmission lines. Voice over IP traffic might be deployed on any IP network, including those lacking a connection to the rest of the Internet, for instance on a private building-wide LAN.
Category: M-Business and M-Commerce
Contractual Obligations between Mobile Service Providers and Users Robert Willis Lakehead University, Canada Alexander Serenko Lakehead University, Canada Ofir Turel McMaster University, Canada
INTRODUCTION The purpose of this chapter is to discuss the effect of contractual obligations between users and providers of mobile services on customer loyalty. One of the unique characteristics of mobile commerce that distinguishes it from most other goods and services is the employment of long-term contractual obligations that users have to accept to utilize the service. In terms of over-the-counter products, sold in one-time individual transactions in well-established markets, a strong body of knowledge exists that suggests that businesses may enhance loyalty through the improvement of quality and customer satisfaction levels. With respect to mobile commerce, however, this viewpoint may not necessarily hold true given the contractual nature of businesscustomer relationships. In the case of mobile computing, it is suggested that loyalty consists of two independent yet correlated constructs that are influenced by different factors: repurchase likelihood and price tolerance. Repurchase likelihood is defined as a customer’s positive attitude towards a particular service provider that increases the likelihood of purchasing additional services or repurchasing the same services in the future (e.g., after the contract expires). For example, when people decide to purchase a new mobile phone, they are free to choose any provider they want. In other words, repurchase likelihood is not affected by contractual obligations. In contrast, price tolerance corresponds to a probability of staying with a current provider when it increases or a competitor decreases service charges. In this situation, individuals have to break the existing contractual obligations. Currently, there is empirical evidence to suggest that the discussion above holds true in terms of mobile computing. However, there are few well-documented works that explore this argument in depth. This article attempts to fill that void. This article will present implications for both scholarship and practice. In terms of academia, it is believed that researchers conducting empirical investigations on customer
loyalty with mobile services should be aware of the two independent dimensions of the business-customer relationship and utilize appropriate research instruments to ensure the unidimensionality of each construct. With regards to practice, it is suggested that managers and marketers be aware of the differences between repurchase likelihood and price tolerance, understand their antecedents, and predict the consequences of manipulating each one. It is noted that overall loyalty is not the only multidimensional constuct in mobile commerce. Recently, it was emperically demonstrated that perceived value of short messaging services is a second-order construct that consists of several independent yet correlated dimensions (Turel et al., 2007). Theoretical separation of the overall loyalty construct into two dimensions has been already empirically demonstrated in three independent mobile commerce investigations. First, Turel and Serenko (2006) applied the American customer satisfaction model (ACSM) to study mobile services in North America. By utilizing the original instrument developed by Fornell, Johnson, Anderson, Cha, and Bryant (1996), they discovered a low reliability of the overall satisfaction construct, and found that the correlation between two items representing price tolerance and one item reflecting repurchase likelihood was only 0.21 (p K
(10)
(11)
where 0 ≤ K < η. Span':
n
' ( S , ) = ∑ 'match ( pi , )
(12)
i =1
With such an allowance given for noise, the scoring system will be able to pick up the optimum schema that matches the user preference. This is because the noise threshold allows schemas to be credited for partial matches with the selected products.
Interactive Product Catalog for M-Commerce
Figure 3. Genetic encoding
Figure 4. GA pseudocode
:{*,*,1,2,3,*}
0
0
Schema
1
2
3
0
Chromosome
Owing to the noise threshold, ambiguity appears in the assessment of schemas. A schema that takes advantage of the threshold term in an unwarranted context stands to gain a higher coverage. One main reason is the simple definition of coverage as a product of span and order, which gives unnecessary credit to schema values that do not match the actual attribute instance value.
pt
1 ( , )= 0
= else
and
≠∗
n
' ( S , ) = ∑∑ ' match (
Coverage:
i =1 j =1
ij
, j)
(13)
pt
(
ij
, j)
n
( S , ) = −∑ 'match ( pi , ) ( pi , )
Score: Γ( S , ) =
i =1
' (S , ) + ⋅ ' (S , ) + ⋅ (S , )
The optimal emphasis varies in different contexts. Though a high degree of emphasis improves the responsiveness of the system, the tradeoff is poorer overall generalization. It is thus advisable to use moderate values of E(i) in most circumstances. Empirical trial tests must be carried out to investigate the effect of a chosen emphasis.
Global Optimization (14)
With the redefinition of coverage, there is an improvement in the score to give less emphasis to matches that makes use of the noise threshold. Despite having a more equitable score, the redefined coverage is still incapable of differentiating between the sensible use of the noise threshold to accommodate noise or the abuse of it to increase coverage. To correct this error, the approach adopted is the inclusion of a penalty term to penalize the usage of the noise threshold. Penalty:
I
INITIALIZE population with random candidate solutions repeat until TERMINATION CONDITION 1. EVALUATE chromosomes 2. SELECT parents 3. RECOMBINE pairs of parents 4. MUTATE offspring 5. EVALUATE offspring 6. SELECT survivors to next generation
(15) (16)
where 0 < μ < 1, λ > 1.
Emphasis Finally, we recognize that a user’s preference may evolve in the course of browsing. Products that were selected more recently are thus likely to be more in line with the current preference of the user. To take this factor into account, we allow a progressive emphasis to be set on more recent selection. We define E(i) as the emphasis factor on a product pi as a function of the product index in the sequence of user selection S. The function may follow either a linear or a geometric progression depending on the desired degree of emphasis. The emphasis factor is then applied to all application of the match function (12).
Having defined a scoring function to evaluate the relative superiority of each schema, we seek to design an algorithm to search for the best schema given a sequence of user selection. EA was found to be a more appropriate choice in our context. In particular, we chose a genetic algorithm (GA), which is a form of EA for the optimization of our scoring function.
Genetic Algorithm By assigning a value of zero to the wildcard, the η-tuple of positive integer values of a schema is encoded directly into a chromosome as an array of integers. Figure 3 illustrates the encoding process. Having defined the chromosomes, we apply the typical genetic algorithm as summarized in Figure 4. Evaluation is done using the scoring function defined in the previous section.
Performance To determine the performance of the algorithm, we define accuracy and efficiency as the performance measures. Accuracy is the frequency that results produce when the genetic algorithm matches the actual global optimum. We calculate accuracy as the average percentage of such matches. On the other hand, efficiency is the amount of computational effort required to execute the algorithm. We thus calculate efficiency as the average number of generations.
349
Interactive Product Catalog for M-Commerce
CONCLUSION The approach in this study focused on realizing the possibility for a more complete m-commerce environment. This outlook is shared by other researchers who attempt to tackle the same problem with different strategies (Guan, Ngoo, & Zhu, 2000). To the best of our knowledge, a customized catalog for m-commerce has not been conceived. This study shares the same intent to make shopping a more pleasant experience for users. Our approach differs in the absence of a passive viewing mode, as the context of m-commerce makes it unfeasible for users to concentrate on the screen for an extended period of time. Interaction control was greatly simplified in our catalog. Through the usage of recommender technology, we streamlined the browsing process by using a reduced form of feedback. In summary, this article highlighted the need for specialized applications in the domain of m-commerce. In particular, the need for expansive browsing as a complement to existing search and filter functions has been emphasized. As a possible solution, a novel method of product catalog navigation with the aid of a recommender system has been proposed. This approach emphasizes a minimal-attention user interface that allows a user to browse through a catalog quickly with as little cognitive effort as possible. The associated recommender system that has been conceived adopts a best effort strategy that accommodates any level of user participation. It has been shown to be capable of detecting non-linear preferences in a set of incremental feedback, as well as tolerate noisy input produced by a user. One drawback of this design is the danger of using predefined product ontology in the enumeration of attribute instances. This leads to stereotypical preference interpretation whose relevance depends largely on how the product ontology is defined.
FUTURE TRENDS For future improvement, it may be worth investigating the possibility of having the recommender generate the ontology from the collective feedback of an ensemble of users. Another enhancement to the existing system would be to incorporate fuzzy logic into the enumeration process. Doing so eliminates the problem around segment boundaries where similar attribute values may be arbitrarily classified into different clusters.
REFERENCES BizRate. (2004). Retrieved from www.bizrate.com
350
Bryan, D., & Gershman, A. (1999). Opportunistic exploration of large consumer product spaces. Proceedings of the 1st ACM Conference on Electronic Commerce (pp. 41-47). Burke, R. D. (2002). Interactive critiquing for catalog navigation in e-commerce. Artificial Intelligence Review, 18, 245-267. Burke, R. D., Hammond, K. J., & Young, B. C. (1997). The FindME approach to assisted browsing. IEEE Expert: Intelligent Systems and Their Applications, 12(4), 32-40. Feldman, S. (2000). Mobile commerce for the masses. IEEE Internet Computing, 4, 75-76. Guan, S. U., Ngoo, C. S., & Zhu, F. M. (2000). Handy broker: An intelligent product-brokering agent for m-commerce applications with user preference tracking. Electronic Commerce Research and Applications, 1, 314-330. Holland, J. H. (1975). Adaptation in natural and artificial systems. Ann Arbor: The University of Michigan Press. Lee, J. Y., Lee, H. S., & Wang, P. (2004). An interactive visual interface for online product catalogs. Electronic Commerce Research, 4, 335-358. Montaner, M., Lopez, B., & Lluis, J. (2003). A taxonomy of recommender agents on the Internet. Artificial Intelligence Review, 19, 285-330. Schafer, J. B., Konstan, J., & Riedl, J. (2001). E-commerce recommendation applications. Data Mining and Knowledge Discovery, 5, 115-153. Tateson, R., & Bonsma, E. (2003). ShoppingGardenImproving the customer experience with online catalogs. BT Technology Journal, 21(4), 84-91.
KEY TERMS E-Commerce: Consists primarily of the distributing, buying, selling, marketing, and servicing of products or services over electronic systems such as the Internet and other computer networks. Feedback Mechanism: Process whereby some proportion or, in general, function of the output signal of a system is passed (fed back) to the input. Genetic Algorithm: Search technique used in computer science to find approximate solutions to optimization and search problems. Genetic algorithms are a particular class of evolutionary algorithms that use techniques inspired by evolutionary biology such as inheritance, mutation, natural selection, and recombination (or crossover).
Interactive Product Catalog for M-Commerce
Global Optimization: A branch of applied mathematics and numerical analysis that deals with the optimization of a function or a set of functions to some criteria. M-Commerce: Electronic commerce made through mobile devices.
Product Catalog: Organized, detailed, descriptive list of products arranged systematically. Product Ontology: Studies’ being or existence, and their basic categories and relationships, to determine what entities and what types of entities exist.
351
I
352 Category: M-Learning
An Interactive Wireless Morse Code Learning System Cheng-Huei Yang National Kaohsiung Marine University, Taiwan Li-Yeh Chuang I-Shou University, Taiwan Cheng-Hong Yang National Kaohsiung University of Applied Sciences, Taiwan Jun-Yang Chang National Kaohsiung University of Applied Sciences, Taiwan
INTRODUCTION Morse code has been shown to be a valuable tool in assistive technology, augmentative and alternative communication, and rehabilitation for some people with various conditions, such as spinal cord injuries, non-vocal quadriplegics, and visual or hearing impairments. In this article, a mobile phone human-interface system using Morse code input device is designed and implemented for the person with disabilities to send/receive SMS (simple message service) messages or make/respond to a phone call. The proposed system is divided into three parts: input module, control module, and display module. The data format of the signal transmission between the proposed system and the communication devices is the PDU (protocol description unit) mode. Experimental results revealed that three participants with disabilities were able to operate the mobile phone through this human interface after four weeks’ practice.
BACKGROUND A current trend in high technology production is to develop adaptive tools for persons with disabilities to assist them with self-learning and personal development, and lead more independent lives. Among the various technological adaptive tools available, many are based on the adaptation of computer hardware and software. The areas of application for computers and these tools include training, teaching, learning, rehabilitation, communication, and adaptive design (Enders, 1990; McCormick, 1994; Bower et al., 1998; King, 1999). Many adapted and alternative input methods now have been developed to allow users with physical disabilities to use a computer. These include modified direct selections (via
mouth stick, head stick, splinted hand, etc.), scanning methods (row-column, linear, circular) and other ways of controlling a sequentially stepping selection cursor in an organized information matrix via a single switch (Anson, 1997). However, they were not designed for mobile phone devices. Computer input systems, which use Morse code via special software programs, hardware devices, and switches, are invaluable assets in assistive technology (AT), augmentative-alternative communication (AAC), rehabilitation, and education (Caves, 2000; Leonard et al., 1995; Shannon et al., 1981; Thomas, 1981; French et al., 1986; Russel & Rego, 1998; Wyler & Ray, 1994). To date, more than 30 manufactures/developers of Morse code input hardware or software for use in AAC and AT have been identified (Anson, 1997; http://www.uwec. edu/Academic/Outreach/Mores2000/morse2000.html; Yang, 2000; Yang, 2001; Yang et al., 2002; Yang et al., 2003a; Yang et al., 2003b). In this article, we adopt Morse code to be the communication method and present a human interface for persons with physical disabilities. The technology employed in assistive devices has often lagged behind mainstream products. This is partly because the shelf life of an assistive device is considerably longer then mainstream products such as mobile phones. In this study, we designed and implemented an easily operated mobile phone human interface device by using Morse code as a communication adaptive device for users with physical disabilities. Experimental results showed that three participants with disabilities were able to operate the mobile phone through this human interface after four weeks’ practice.
SYSTEM DESIGN Morse code is a simple, fast, and low-cost communication method composed of a series of dots, dashes, and intervals
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
An Interactive Wireless Morse Code Learning System
Figure 1. System schematics of the mobile phone humaninterface ( Phone Call ) GSM Station
Mobile Phone Users
( SMS )
Voice Earphone
UART
Manipulation
A Mobile Phone Human-Interface
Single Switch
in which each character entered can be translated into a predefined sequence of dots and dashes (the elements of Morse code). A dot is represented as a period “.”, while a dash is represented as a hyphen, or minus sign, “-”. Each element, dot or dash, is transmitted by sending a signal for a standard length of time. According to the definition of Morse code, the tone ratio for dot to dash must be 1:3. That means that if the duration of a dot is taken to be one unit, then that of a dash must be three units. In addition, the silent ratio for dot-dash space to character-space also has to be 1:3. In other words, the space between the elements of one character is one unit while the space between characters is three units (Yang et al., 2002). In this article, the mobile phone human interface system using Morse code input device is schematically shown in Figure 1. When a user presses the Morse code input device, the signal is transmitted to the key scan circuit, which translates the incoming analog data into digital data. The digital data are then sent into the microprocessor, an 8051 single chip, for further processing. In this study, an ATMEL series 89C51 single chip has been adopted to handle the communication between the press-button processing and the communication devices. Even though the I/O memory capacity of the chip is small compared to a typical PC, it is sufficient to control the device. The 89C51 chip’s internal serial communication function is used for data transmis-
sion and reception (Mackenzie, 1998). To achieve the data communication at both ends, the two pins, TxD and RxD, are connected to the TxD and RxD pins of a RS-232 connector. Then the two pins are connected to the RxD and TxD of an UART (Universal Asynchronous Receiver Transmitter) controller on the mobile phone device. Then, persons with physical disabilities can use this proposed communication aid system to connect their mobile communication equipment, such as mobile phones or GSM (global system for mobile communications) modems, and receive or send their messages (SMS, simple message service). If they wear an earphone, they might be able to dial or answer the phone. SMS is a protocol (GSM 03.40 and GSM 03.38), which was established by the ETSI (the European Telecommunications Standards Institute) organization. The transmission model is divided into two models: text and PDU (protocol description unit). In this system, we use the PDU model to transmit and receive SMS information through the AT command of the application program (Pettersson, 2000). Structurally the mobile phone human-interface system is divided into three modules: the input module, the control module, and the display module. The interface framework is graphically shown in Figure 2. A detailed explanation is given below.
INPUT MODULE A user’s input will be digitized first, and then the converted results will be sent to the micro controller. From the signal processing circuit can monitor all input from the input device, the Morse code. The results will be entered into the input data stream. When the user presses the input key, the micro-operating system detects new input data in the data stream, and then sends the corresponding characters to the display module. Some commands and/or keys, such as OK, Cancel, Answer, Response, Send, Receive, Menu, Exit, and so forth, have been customized and perform several new functions in order to accommodate the Morse code system. These key modifications facilitate the human interface use for a person with disabilities.
Figure 2. Interface framework of mobile phone for persons with physical disablities Input
Module
Control
Module
Input Device
Recognition Circuit
Key Scan Circuit
Microprocessor
Signal Processing Circuit
Communication Device
Display Module
LCD Circuit
Speaker
353
I
An Interactive Wireless Code Learning System
Figure 3. Block diagram of the Morse code recognition system Dot-Dash Space bj(xi) m1(x1),b1(x1),...,mn(x1), bn(x1),m1(xi+1),...
Space Recognition
Character Space m1(xi),…mn(xi) mj(xi)
CONTROL MODULE The proposed recognition method is divided into three modules (see Figure 3): space recognition, adjustment processing, and character translation. Initially, the input data stream is sent individually to separate tone code buffer and space recognition processes, which are based on key-press (Morse code element) or key-release (space element). In the space recognition module, the space element value is recognized as a dot-dash space or a character space. The dotdash space and character space represent the spaces existing between individual characters and within isolated elements of a character respectively. If a character space is identified, then the value(s) in the code buffer is (are) sent to character translation. To account for varying release speeds, the space element value has to be adjusted. The silent element value is sent into the silent base adjustment process. Afterwards, the character is identified in the character translation process. A Morse code character, xi, is represented as follows: m1(xi), b1(xi), …, mj(xi), bj(xi), …, mn(xi), bn(xi) where bj(xi): jth silent duration in the character xi. n: the total number of Morse code elements in the character xi. mj(xi): the jth Morse code element of the input character
DISPLAY MODULE Since users with disabilities have, in order to increase the convenience of user operations, more requirements for 354
Silent Base Adjustment
Data Input
Morse code digital stream
xi.
bj(xi)
Tone Code Buffer
Character Translation
xi
system interfaces than a normal person, the developed system shows selected items and system condition information on an electronic circuit platform, which is based on LCD (liquid crystal display). The characteristics of the proposed system can be summarized as follows: (1) easy operation for users with physical disabilities with Morse code input system, (2) multiple operations due to the selection of different modes, (3) highly tolerant capability from adaptive algorithm recognition, and (4) system extension for customized functions.
RESULTS AND DISCUSSIONS This system provides two easily operated modes, the phone panel and LCD panel control mode, which allow a user with disabilities easy manipulation. The following shows how the proposed system sends/receives simple message service (SMS) message or make /respond to a phone call.
SMS Receiving Operation First, when users receive a message notification and want to look at the content, this system will provide “phone panel” and “LCD panel” control modes to choose from. In the phone panel mode, users can directly key-in Morse code “…” (as character ‘S’). The interface system will go through the message recognition process, then exchange the message into AT command “AT+CKPD=’S’, 1”, to execute the “confirm” action of the mobile phone. The purpose of this process is the same as users keying-in “yes” on the mobile phone keyboard, then keying-in Morse code “. - - . .” (as key ‘↓’). The system will recognize the message, then automatically send the “AT+CKPD=’↓’, 1” instruction. The message cursor of the mobile phone is moved to the next line, or key-in Morse code “. - . . -” (as key ‘↑’) for moving it to the previous data line. Finally, if users want
An Interactive Wireless Morse Code Learning System
to exit and return to the previous screen, they only need to key-in Morse code “. . - .” (as character ‘F’), and start the c key function on the mobile phone keyboard. If LCD panel mode is selected, one can directly follow the selected items on the LCD crystal, to execute the reception and message reading process.
SMS Transmitting Operation Message transmission services are provided in two modes: phone panel and LCD panel. In the phone panel mode, continually type two times the Morse code “. - . - .” (as key ‘→’). The system will be converted into AT Command and transferred into mobile phone to show the selection screen of the message functions. Then continuing to key-in three times the Morse code “. . .” (as character ‘S’), one can get into the editing screen of message content, and wait for users to input the message text data and receiver’s phone number. The phone book function can be used to directly save the receiver’s phone number. After the input, press the “yes” key to confirm that the message sending process has been completed. In addition, if the LCD panel mode is selected, one can follow the LCD selection prompt input the service selection of all the action integrated in the LCD panel. Then go through the interface and translate to a series of AT command orders, and batch transfer these into the mobile phone to achieve the control purpose. The selection command “Answer a phone,” displays on the menu of the LCD screen, and can be constructed using Morse code. The participants could press and release the switch, and input the number code “. - - - -” (as character ‘1’) or hot key “. -“ (as character ‘A’). The mobile phone is then answered automatically. Problems with this training, according to participants, are that the end result is limited typing speed and users must remember all the Morse code set of commands. Three test participants were chosen to investigate the efficiency of the proposed system after practicing on this system for four weeks. Participant 1 (P1) was a 14-yearold male adolescent who has been diagnosed with cerebral palsy. Participant 2 (P2) was a 14-year-old female adolescent with cerebral palsy, athetoid type, who experiences involuntary movements of all her limbs. Participant 3 (P3) was a 40-year-old male adult, with a spinal cord injury and incomplete quadriparalysis due to an accident. These three test participants with physical impairments were able to make/respond to phone calls or send/receive SMS messages after practice with the proposed system.
tate the use of everyday appliances for people with physical disabilities considerably.
CONCLUSION To help some persons with disabilities such as amyotrophic lateral sclerosis, multiple sclerosis, muscular dystrophy, and other conditions that worsen with time and cause the user’s abilities to write, type, and speak to be progressively lost, it requires an assistive tool for purposes of augmentative and alternative communication in their daily lives. This article presents a human interface for mobile phone devices using Morse code as an adapted access communication tool. This system provides phone panel and LCD panel control modes to help users with a disability with operation. Experimental results revealed that three physically impaired users were able to make/respond to phone calls or send/receive SMS messages after only four weeks’ practice with the proposed system.
ACKNOwLEDGMENTS This research was supported by the National Science Council, R.O.C., under grant NSC 91-2213-E-151-016.
REFERENCES Anson, D. (1997). Alternative computer access: A guide to selection. Philadelphia, PA: F. A. Davis. Bower, R. et al. (Eds.) (1998). The Trace resource book: Assistive technology for communication, control, and computer access. Madison, WI: Trace Research & Development Center, Universities of Wisconsin-Madison, Waisman Center. Caves, K. (2000). Morse code on a computer—really? Keynote presentation at the First Morse 2000 World Conference, Minneapolis, MN. Enders, A., & Hall, M. (Ed.) (1990). Assistive technology sourcebook. Arlington, VA: RESNA Press,. French, J. J., Silverstein, F., & Siebens, A. A. (1986). An inexpensive computer based Morse code system. In Proceedings of the RESNA 9th Annual Conference, Minneapolis (pp. 259-261). Retrieved from http://www.uwec.edu/Academic/ Outreach/Mores2000/morse2000.html.
FUTURE TRENDS
King, T. W. (1999). Modern Morse code in rehabilitation and education. MA: Allyn and Bacon.
In the future, Morse code input device could be adapted to several environmental control devices, which would facili-
Lars Pettersson. (n.d.). Dreamfabric. Retrieved from http:// www.dreamfabric.com/sms 355
I
An Interactive Wireless Code Learning System
Leonard, S., Romanowski, J., & Carroll, C. (1995). Morse code as a writing method for school students. Morsels, University of Wisconsin-Eau Claire, 1(2), 1.
environmental control system for severely disabled individuals. IEEE Transactions on Neural System and Rehabilitation Engineering, 11(4), 463-469.
Mackenzie, I. S. (1998). The 89C51 Microcontroller (3rd ed.). Prentice Hall.
KEY TERMS
McCormick, J. A. (1994). Computers and the Americans with disabilities act: A manager’s guide. Blue Ridge Summit, PA: Wincrest/McGraw Hill. Russel, M., & Rego, R. (1998). A Morse code communication device for the deaf-blind individual. In Proceedings of the ICAART, Montreal (pp. 52-53). Shannon, D. A., Staewen, W. S., Miller, J. T., & Cohen, B. S. (1981). Morse code controlled computer aid for the nonvocal quadriplegic. Medical Instrumentation, 15(5), 341-343. Thomas, A. (1981). Communication devices for the nonvocal disabled. Computer, 14, 25-30. Wyler, A. R., & Ray, M. W. (1994). Aphasia for Morse code. Brain and Language, 27(2), 195-198. Yang, C.-H. (2000), Adaptive Morse code communication system for severely disabled individuals. Medical Engineering & Physics, 22(1), 59-66. Yang, C.-H. (2001). Morse code recognition using learning vector quantization for persons with physical disabilities. IEICE Transactions on Fundamentals of Electronics, Communication and Computer Sciences, E84-A(1), 356362. Yang, C.-H., Chuang, L.-Y. Yang, C.-H., & Luo, C.-H. (2002). An Internet access device for physically impaired users of Chanjei Morse code. Journal of Chinese Institute of Engineers, 25(3), 363-369. Yang, C.-H. (2003a). An interactive Morse code emulation management system. Computer & Mathematics with Applications, 46, 479-492. Yang, C.-H., Chuang, L.-Y., Yang, C.-H., & Luo, C.-H. (2003b, December). Morse code application for wireless
356
Morse Code: Morse code is a transmission method, implemented by using just a single switch. The tone ratio (dot to dash) in Morse code has to be 1:3 per definition. This means that the duration of a dash is required to be three times that of a dot. In addition, the silent ratio (dot-space to character-space) also has to be 1:3. Adaptive Signal Processing: Adaptive signal processing is the processing, amplification and interpretation of signals that change over time through a process that adapts to a change in the input signal. Augmentative and Alternative Communication (AAC): Support for and/or replacement of natural speaking, writing, typing, and telecommunications capabilities that do not fully meet communicator’s needs. AAC, a subset of AT (see below), is a field of academic study and clinical practice, combining the expertise of many professions. AAC may include unaided and aided approaches. Assistive Technology (AT): A generic term for a device that helps a person accomplishes a task. It includes assistive, adaptive and rehabilitative devices, and grants a greater degree of independence people with disabilities by letting them perform tasks they would otherwise be unable of performing. Simple Message Service (SMS): A service available on digital mobile phones, which permits the sending of simple messages between mobile phones. Global System for Mobile Communications (GSM): GSM is the most popular standard for global mobile phone communication. Both its signal and speech channels are digital and it is therefore considered a 2nd generation mobile phone system.
Category: 3G 357
Interworking Architectures of 3G and WLAN Ilias Politis University of Patras, Greece Tasos Dagiuklas Technical Institute of Messolonghi, Greece Michail Tsagkaropoulos University of Patras, Greece Stavros Kotsopoulos University of Patras, Greece
INTRODUCTION The complex and demanding communications needs of modern humans led recently to the deployment of the 3G/ UMTS mobile data networks and the wireless LANs. The already established GSM/GPRS radio access technology can easily handle the voice and low-rate data traffic such as short messages (SMS); however, it is inadequate for the more challenging real-time multimedia exchanges that require higher data rates and ubiquitous connectivity. The UTRAN radio access technology provides wide area coverage and multimedia services up to 2Mbps, while the recently deployed WLANs offer radio access at hotspots such as offices, shopping areas, homes, and other Internet/intranet-connected networks, with very high data rates up to 54Mbps (IEEE 802.11g). Hence, there is a strong need to integrate WLANs and 3G access technologies, and to develop a heterogeneous network based on an all-IP infrastructure that will be capable to offer ubiquitous and seamless multimedia services at very broadband rates. The major benefits that drive towards an all-IP based core network are the following (Wisely et al., 2002): • • •
• •
Cost Saving on Ownership and Management: Network operators need to own and manage one single network, instead of multiple. Cost Saving on Transport: For example, the cost to provide IP transport is lower. Future Proof: It can be claimed that the future of backbone network, both for voice and data, is IP based. An IP-based network allows smooth interworking with an IP backbone and efficient usage of network resources. Smooth integration of heterogeneous wireless access technologies. The IP multimedia domain can support different access technologies and greatly assist towards fix/mobile convergence.
•
•
•
Capacity Increase: The capacity enhancement of an IP-based transport network is quicker and cheaper. The same is also true to service capacity, thanks to the distributed nature of the service architecture. Rich Services: The benefits of VoIP are available for improved and new services, for example, voice/multimedia calls can be integrated with other services, providing a powerful and flexible platform for service creation. Enable peer-to-peer networking and service model.
This hybrid network architecture would allow the user to benefit from the high throughput IP-connectivity in ‘hotspots’ and to attain service roaming across heterogeneous radio access technologies such as IEEE 802.11, HiperLan/2, UTRAN, and GERAN. The IP-based infrastructure emerges as a key part of next-generation mobile systems since it allows the efficient and cost-effective interworking between the overlay networks for seamless provisioning of current and future applications and services (De Vriendt et al., 2002). Furthermore, IP performs as an adhesive, which provides global connectivity, mobility among networks, and a common platform for service provisioning across different types of access networks (Dagiuklas et al., 2002). The development of an all-IP interworking architecture, also referred to as fourthgeneration (4G) mobile data network, requires specification and analysis of many technical challenges and functions, including seamless mobility and vertical handovers between WLAN and 3G radio technologies, security, authentication and subscriber administration, consolidated accounting and billing, QoS, and service provisioning (Tafazolli, 2005). This article discusses the motivation, interworking requirements, and different architectures regarding 3G/ WLAN interworking towards an all-IP hybrid networking environment. Five common interworking techniques and architectures that effectively can support most of the issues addressed previously are presented and discussed. These are
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
I
Interworking Architectures of 3G and WLAN
namely: open coupling, loose coupling, tight coupling, very tight coupling (3GPP, 2004), and the recently developed interworking technology named unlicensed mobile access (UMA), which arises as a very competitive solution for the interworking environment (3GPP-UMAC, 2005). The focus of the article is on a comparison and qualitative analysis of the above architectures.
•
3G AND wLAN INTERwORKING Motivation The main motivation for mobile operators to get involved in the WLAN business (Dagiuklas & Velentzas, 2003) is the following: • • • •
Public WLANs provide the opportunity for mobile operators to increase their revenues significantly from mobile data traffic. WLANs can be considered as an environment for testing new applications at the initial stage. High-demand data traffic from hotspot areas can be diverted from 3G to WLAN, relieving potential network congestion. Location-based services in hotspot areas could be based on WLAN technology rather than using more-complex GPS-like systems.
On the other hand, a shift from WLAN to 3G could take place due to the following reasons: • • •
Poor Coverage: Users may be able to use WLAN services at the airport of departure, but not at the airport of arrival or at the hotel. Lack of Brand Recognition: The service operators are often new start-ups, which causes end-users to hesitate to use the service. Lack of Roaming Agreements: End users are forced to locate different service providers at the places they roam to.
The service provider value proposition for utilizing integrated WLANs with cellular networks includes the following benefits for carriers as well as their subscribers: •
•
358
Extension of current service offering by: • integrating cellular data and WLAN solutions, • positioning for voice phone service in hotspots, and • engaging enterprises with in-building solutions. Improve bottom line with new revenue and lower churn:
•
The carrier provides improved in-building coverage by using intranet bandwidth instead of in-building cell sites to provide coverage. • Cross system/service integration features become a competitive advantage for the carriers offering seamless mobility services. • The cellular provider derives service revenue for authentication services, mobility services, and calls that do not use cellular bearer channels. • The cellular handset becomes an indispensable element. • The handset can operate with more functionality, for example, even as gateway. • The subscriber increases his dependency on the handset. Payload traffic trade-off: • Some calls will hand over from cellular channels to WLAN connections when subscribers enter these coverage areas. • Other calls will hand over to cellular bearer channels when people leave WLAN coverage areas. • A more integrated approach to data traffic will probably increase the use of data transferred over cellular networks.
It becomes evident that as subscribers become more dependent on their much more useful handsets, they will call and be called more and everywhere.
3G and wLAN Architectures The interworking between 3G and WLAN is a trivial issue that is under study by international standardization fora, namely, ETSI, 3GPP, and the UMTS Forum. The undergoing investigation has provided specific requirements that interworking solutions need to meet. The demands include the establishment of some kind of partnership between the 3G operator and the wireless Internet service provider (WISP), a common billing and accounting policy between roaming partners, and a shared subscriber database for authentication, authorization, and accounting (AAA) and security provisioning (Nakhjiri & Nakhjiri, 2005). The work in this article refers to four already established interworking scenarios (Salkintzis, 2004) regarding 3G and WLAN, which are presented and compared to the recently developed UMA architecture. In open coupling interworking architecture, there is no requirement for specific WLAN access, while each of the networks3G and WLANfollows separate authentication procedures. This architecture does not support seamless services, while the user performs a vertical handover from 3G towards WLAN and vice versa.
Interworking Architectures of 3G and WLAN
Figure 1. Open coupling scenario ard nd Sta NIC
I
wISP Server
WLAN
CCBS AAA
Access Point
Access Router
HLR
RAN BSS
RNC
SGSN
GGSN
Billing system The link between 2 independent access networks (WLAN & RAN)
Figure 2. Loose coupling scenario C N I ed ific d u s c e r Sp ca IM S if
w ISP Server
W LAN
CCBS AA A
Access Point
Access Router
AA A-H LR link H LR
RAN RN C
SGSN
G GSN
Figure 3. Tight coupling scenario C NI i fic e d c e r S p equi r
WLAN Access Point
New interface definition Iu (RNC-SGSN ) Iub (RAN-RNC)
IW U/RN C
CCBS
HLR
RAN RNC
In a loose coupling scenario, 3G and WLAN share a common customer database and authentication process. There is no requirement for specific WLAN access. In addition, no load balancing is provided for applications with specific QoS requirements, and the architecture does not support seamless services. Similar to previous architecture, loose coupling does not support seamless services. On the other hand, a tight coupling scenario supports seamless service provisioning for vertical handovers and
SGSN
GGSN
load balancing for QoS demanding applications. However, there is need for definition of the interface interconnection between the WLAN and SGSN node. Similar to the previous scenario, very tight coupling offers the same advantages, although in this case WLAN is considered as part of the UTRAN and a new interface interconnecting the WLAN and the UTRAN/RNC needs specification.
359
Interworking Architectures of 3G and WLAN
Figure 4. Very tight coupling scenario C NI ic c i f ir e d e S p equ r
WLAN
New interface definition Iu (RNC-WLAN), WLAN seen as a cell Management at the RNC level
Access Point
CCBS
HLR
RAN RNC
SGSN
GGSN
Figure 5. UMA architecture
The above interworking scenarios are constrained by the fact that they cover only three wireless technologies. The B3G vision considers the employment of new emerging wireless technologies (e.g., WiMAX, 4G, etc.). The limitation of the approach mentioned above can be alleviated through the approach envisioned by the Unlicensed Mobile Access Consortium (UMAC), where all wireless technologies can be smoothly integrated towards an all-IP-based heterogeneous network. Recently, UMAC produced a multi-access architecture known as unlicensed mobile access (UMA), which rises as an exiting prospect. In the next section, the UMA architecture is presented.
UNLICENSED MOBILE ACCESS The UMA technology provides access to circuit and packetswitched services over the unlicensed spectrum, through technologies which include Bluetooth and IEEE 802.11 WLANs. The standardization procedure has been initiated by the Unlicensed Mobile Access Consortium (UMAC) 360
and is currently continued by the 3rd Generation Partnership Project (3GPP) under the work item “Generic Access to A./Gb interface.” UMA deployment offers to the subscribers a ubiquitous connectivity and consistent experience for mobile and data services as they transition between cellular and public or private unlicensed wireless networks. Compared to other cellular-WLAN interworking solutions, UMA is superior in both the technology aspects and the cost effectiveness for customers and providers (Velentzas & Dagiuklas, 2005). In more detail, UMA technology does not affect the operation of current cellular radio access networks (UTRAN/ GERAN), like cell planning. Hence the investments in existing and future mobile core network infrastructures are preserved. UMA utilizes standard all-IP infrastructure, which provides access to standard packet-switched services and applications. Additionally, the UMA network (UMAN) is independent of the technology used for unlicensed spectrum access, thus it is open to new wireless technologies with no extra requirements. It enables seamless handover between the heterogeneous access technologies, while it ensures service
Interworking Architectures of 3G and WLAN
continuity across the network coverage area. Compared to a loose-coupling WLAN-3G interworking scenario, UMA offers greater control over the authentication, authorization, and accounting procedures and tighter end-to-end security. Moreover, UMA technology supports load balancing for efficient allocation of bandwidth and data rates according to the customer requirements, and provides higher throughput and network capacity to the operator that it is translated to more connected customers, thus producing higher revenues. The UMA technology introduces a new network element, the UMA Controller (UNC), and associates protocols that provide secure transport of GPRS signaling and user plane traffic over IP. UNC acts as a GERAN base station controller (BSC) and includes a security gateway (SGW) that: (1) terminates secure remote access tunnels between the UNC and the mobile station (MS); and (2) provides authentication, encryption, and data integrity for signaling and media traffic. The UNC is connected to a unique mobile switching center (MSC) through the standard A-interface and a serving GPRS support node (SGSN) through the Gb-interface in order to relay GSM and GPRS services respectively to
the core network, while through the Wm-interface allows authentication signaling with the corresponding AAA server. The Up-interface between the UNC and MS operates over the IP transport network, and relays circuit and packet-switched services and signaling among the mobile core network and the MS. The MS, which is capable of switching between cellular radio access networks (RANs) and unlicensed, also has an IP interface to the access point that extends the IP access from the UNC to the MS.
COMPARISON OF DIFFERENT ARChITECTURES AND qUALITATIVE ANALYSIS The following table showcases the comparison between the proposed interworking architectures. The comparison is based on already mentioned parameters that affect not only the technical efficiency of each of the proposed solutions, but also the cost efficiency, as this is perceived from the operator’s and from the user’s point of view.
Table 1. Qualitative comparison of interworking scenarios
Qualitative Parameters
Open Coupling
Loose Coupling
Tight Coupling
Very Tight Coupling
UMA
Service Continuity
The running application will not continue across 3G and WLAN once vertical handover takes place
It is not supported and time sensitive services will be interrupted during handover
Service continuity is supported, although QoS may be degraded during handover
Similar to Tight Coupling
Service continuity is fully supported
Simplicity
The user for the same service may have to subscribe to at least two service providers
The user for the same service may have to subscribe to at least two service providers
One service, one mailbox
One service, one mailbox
One service, one mailbox
Seamless vertical handovers are not supported
Seamless services are not supported by this architecture
Seamless handovers and mobility are supported
Seamless handovers and mobility are supported
Seamless mobility for circuit and packet switched services
The architecture has no capability to divert the services according to their QoS demands
It is no supported, the system cannot select the network that is suitable for the QoS requests of the service
It is not supported, network selection is based on network coverage at current user’s location
It is not supported, network selection is based on network coverage at current user’s location
Supports load balancing decisions based on the required QoS of the application and the application’s requested bandwidth
Common authentication procedures and security provisioning
The architecture supports SIM/EAPAKA authentication and IPSec protocol for the unlicensed mobile part and common GPRS/3G authentication procedure from UNC towards SGSN and MSC
Seamlessness
Load Balancing
Security and Authentication
Separate authentication procedure and security provisioning
Common authentication procedure, the 3G/HLR database is shared between the networks
Common authentication procedures, SGSN is the point of decision, 3G like security scheme
361
I
Interworking Architectures of 3G and WLAN
Qualitative Parameters
Table 1. Qualitative comparison of interworking scenarios, continued
Openness
Additional radio access schemes may be added to the architecture
The architecture may support other heterogeneous wireless technologies
The architecture is proprietary and depends on the WLAN technology used
The architecture is proprietary and depends on the WLAN technology used
The architecture supports interworking between 3G and heterogeneous wireless technologies
Access Control
No specific WLAN access is required, the access control is WLAN based
Independent of access technology used, the access control is WLAN based
Dependent on access technology used due to the new IWU/RNC-SGSN interface, access control is 3G based
Dependent on access technology, management at RNC, 3G based access control
Access is WLAN based over unlicensed radio spectrum
MT with specific NIC
MT are required to support mechanism (software module) for switching between UMA and 3G radio interface
‘Iu’ interface between RNC and WLAN
Further standardization is required, however leverages many already defined 3G and IP based protocols
The operator is required to install new infrastructure at hotspots to interconnect WLAN in the RNC
New infrastructure needs to be installed, UNC, however utilisation of unlicensed spectrum and openness to new technologies compensates for the cost.
MT Complexity
Standard MT are used with some kind of WLAN interface
Standard MT are used with some kind of WLAN interface
MT with specific NIC
Standardization Complexity
No further standardization effort is required
ISP/AAA-3G/HLR link standardization, translation from MAP to RADIUS/ DIAMETER
Significant standardization effort, new interface definition, ‘Iu’ (RNC-SGSN), ‘Iub’ (RAN-RNC)
Cost Efficiency
WLAN and 3G networks run separately and there is no further financial burden for the operator
A very cost efficient solution since it is based on the implementation of well established technologies
In comparison with the loose and tight coupling scenarios, the UMA approach for WLAN 3G interworking has the advantage in both technology and cost efficiency for operators and customers. Specifically, the installation of a UNC controller, although it requires further standardization and protocol definitions, arises as a better solution. It allows easy openness to new wireless technologies with no extra requirements and is able to provide QoS guarantees for multimedia real-time applications and time-sensitive services in general, seamless mobility, and uninterrupted services across the network coverage area. Hence, the operator is able to cover the demands of customers and the customers have access to all services anywhere and anytime. Authentication, authorization, accounting, and security are common between 3G and WLAN and based on already established protocols and standards. Compared to the other solutions, UMA offers greater control on AAA procedures and tighter end-to-end security. Finally, UNC supports load balancing, which providesin addition to the best possible solution in terms of connectivity, bandwidth, and data rates, according to the service that the customer requireshigher throughput and network capacity to the operator so that it is translated to more connected customers, thus producing higher revenues. 362
The operator is required to install new infrastructure at hotspots to interconnect WLAN in the SGSN
FUTURE TRENDS There is no industry consensus on what next generation networks will look like, but as far as the next generation networks are concerned (Kingston, Morita, & Towle, 2005), ideas and concepts include: • • • • • • • • •
transition to an “All-IP” network infrastructure; support of heterogeneous access technologies (e.g., UTRAN, WLANs, WiMAX, xDSL, etc.); VoIP substitution of the pure voice circuit switching; seamless handovers across both homogeneous and heterogeneous wireless technologies; mobility, nomadicity, and QoS support on or above IP layer; need to provide triple-play services creating a service bundle of unifying video, voice, and Internet; home networks are opening new doors to the telecommunication sector and network providers; unified control architecture to manage application and services; and convergence among network and services.
Interworking Architectures of 3G and WLAN
Two important factors have been considered to satisfy all these requirements. The first one regards the interworking of existing and emerging access network under the umbrella of a unified IP-based core network and unified control architecture supporting multimedia services. A proposed solution towards this direction is the unlicensed mobile access (UMA), allowing heterogeneous wireless technologies to interconnect to a core network through a network controller. The second requirement regards IP multimedia subsystem (IMS) evolution in order to cope with requirements imposed by NGN architecture (Passas & Salkintzis, 2005). The initial release of 3GPP IMS was developed only for mobile networks. The increasing demand of interworking between different access devices and technologies led to subsequent releases that defined IMS as a core independent element and a key enabler for applying fixed mobile convergence (FMC). FMC comprises two attributes: using one number, voice/mail and seamless handover of multimedia sessions. In the B3G/4G vision, IMS is required to become the common architecture for both fixed and mobile services. Towards this end the ETSI Telecoms and Internet converged services and protocols for advanced networks (TISPAN) is also producing new functionality extensions for the IMS (ETSI TISPAN, n.d.).
De Vriendt, J. et al. (2002). Mobile network evolution: A revolution on the move. IEEE Communications Magazine, 4, 104-111.
CONCLUSION
3GPP. (2004a, September). Technical specification group services and system aspects: 3GPP system to wireless local area network (WLAN) interworking: System description, 3 TS 23.234. Retrieved from http://www.3gpp.org
The conclusion of the qualitative analysis relates the most suitable solution for an interworking architecture of 3G and WLAN radio access technologies based on an all-IP core network with the UMA network. The most important characteristics of UMA, as they have been discovered during the analysis, are among others: the seamless support for vertical handovers and the QoS guarantees for multimedia and time-sensitive applications due to the load balance capability; and the network continuity, scalability, and cost efficient openness. Moreover, the UMA network solution for integrated 3G and WLAN technologies enables network operators to leverage cost and performance benefits of VoIP, broadband, and Wi-Fi, while it supports all mobile services voice, packet, and IMS/SIP, and utilizes standard interfaces into the all-IP core network.
REFERENCES Dagiuklas, T. et al. (2002). Seamless multimedia services over all-IP network infrastructures: The EVOLUTE approach. Proceedings of the IST Summit 2002 (pp. 75-78). Dagiuklas, T., & Velentzas, S. (2003, July). 3G and WLAN interworking scenarios: Qualitative analysis and business models. IFIP HET-NET03, Bradford, UK.
ETSI TISPAN. (n.d.). NGN functional architecture: Resource and admission control subsystems, release 1. Kingston, K., Morita, N., & Towle, T. (2005). NGN architecture: Generic principles, functional architecture and implementation. IEEE Communications Magazine, (October), 49-56. Nakhjiri, M., & Nakhjiri, M. (2005). AAA and network security for mobile access (pp. 1-23). New York: John Wiley & Sons. Passas, N., & Salkintzis, A. (2005). WLAN/3G integration for next generation heterogeneous mobile data networks. Wireless Communication and Mobile Computing Journal, (September). Salkintzis, A. (2004). Interworking techniques and architectures for WLAN/3G integration towards 4G mobile data networks. IEEE Wireless Communications, (June), 50-61. Tafazolli, R. (2005). Technologies for the wireless future. New York: John Wiley & Sons.
3GPP. (2004b, September). IP multimedia subsystem version 6. 3G TS 22.228. 3GPP-UMAC. (2005, June). UMA architecture (stage 2). Retrieved from http://www.3gpp.org Unified Mobile Access Consortium. (n.d.). Retrieved from http://www.uma.org Velentzas, S., & Dagiuklas, T. (2005, July). Tutorial: 4G/wireless LAN interworking. IFIP HET-NET 2005, Ilkley, UK. Wisely, D. et al. (2002). IP for 3G: Networking technologies for mobile communications. New York: John Wiley & Sons.
KEY TERMS Authentication Authorization Accounting (AAA): Provides the framework for the construction of a network architecture that protects the network operator and its customers from attacks and inappropriate resource management and loss of revenue.
363
I
Interworking Architectures of 3G and WLAN
B3G/4G: Beyond 3G and 4G mobile communications that provide seamless handover between heterogeneous networks and service continuity.
NGN: An ITU standard for Next Generation Networks where cellular mobile 3G systems, WLANs, and fixed networks are integrated over IP protocol.
IP Multimedia Subsystem (IMS): Provides a framework for the deployment of both basic calling and enhanced multimedia services over IP core.
3G: Third generation of cellular mobile communications (GPRS/UMTS).
364
Category: Location and Context Awareness 365
iPod as a Visitor’s Personal Guide Keyurkumar J. Patel Box Hill Institute, Australia Umesh Patel Box Hill Institute, Australia
INTRODUCTION
BACKGROUND
Over the past few years, use of mobile devices for various purposes has increased. Apple released its first iPod on October 23, 2001, a breakthrough MP3 player. Today, Apple’s fifth-generation iPod is available which can be considered as a portable media player that focuses on the playback of digital video, as well as storing and displaying pictures and video (see apple.com). Since then the iPod has been successfully and effectively used for various purposes including as a media player, bootable drive, external data storage device, PDA replacement, and for podcasting. Academia and tourism are two areas where the use of mobile devices are encouraged to gain benefits from the technology. For academic use, the iPod’s recording and storage capabilities have been explored by some educational institutes across the United States. According to the Duke University iPod First-Year Experience Final Evaluation Report, the iPod supports individual learning preferences and needs, and easy-to-use tools for recording interviews, field notes, and small-group discussions. The tourism industry is also identified as a potential area to use mobile technologies. Recently, Dublin Tourism, Ireland discovered the use of the iPod as a portable tourist guide; Ireland’s neighbor Scotland followed (see Physorg, 2006). Sales of interactive portable MP3 players have increased explosively in the last few years. Information Media Group predicts that sales will continue to increase at the rate of 45% for next six years (Macworld UK, 2005). The iPod is currently the world’s best-selling digital audio player and increased its popularity in Australia sevenfold in 2004 (see apple.com). Greg Joswiak, the worldwide vice president of iPod marketing, said: “As of August 2005, market share in Australia is 68% of [the] digital player market.” With the increasing use of digital media together with the handheld devices, this iPod application will eliminate the need for human guides and will provide an entertaining experience to visitors. It will be very useful for landmark tourist destinations such as aquariums and museums, and there will be a huge demand with the increasing flow of tourists in Australia, which according to Tourism Australia (2005) was an increase of 5.4% from 2004 to 2005, with tourists numbering 5.5 million in the latter year.
Tourism is an important activity for human life as a source of pleasure and during the holidays. We visit various places every now and then, including tourist destinations such as a museum, commercial destinations such as a stock market, educational institutes such as a university, or public places such as a shopping mall. Every new visitor suffers from preconceptions and anxiety from their lack of knowledge about the visiting site. This acts as a barrier and must be overcome before an effective visit can take place. As for visitors’ preconceptions, the authors of this article encourage visitors to address their anxiety and introduce excitement before they start the tour. The tourism industry so far has promoted the various communication mediums such as maps on the board, written information about specific locations, and now display video screens. Tourism has been a popular area for mobile information systems (Cheverst, Davies, Mitchell, Friday, & Efstratiou, 2000) and other PDA-based systems. Audio and video has been neglected or underused as a leaning medium in recent years (Scottish Council for Educational Technology, 1994.). The general view is that video is a better tool for leaning than audio. Animation and interactive media like simulations can attract attention, but they proved to be expensive. Hearing is an astoundingly efficient skill according to Clark and Walsh (2004). Portable media players such as PDAs and iPods can provide information anytime and anywhere. These devices come with their own hard drives and eliminate transportation of storage devices, which is a requirement for video communication. The iPod, with built-in speakers and microphone, makes it easier to record and playback information stored into its hard drive. Clark and Walsh (2004) stated that besides its popularity and ease of use, listening to an iPod and similar devices at public places is socially accepted. At Box Hill Institute of TAFE, we realized the use of an iPod as a part of our “Innovation Walk” project. The “Innovation Walk” is developed with the aim of showcasing the institute’s prized innovations. Career teachers, overseas visitors, students, industry and government dignitaries, and member of the community can undertake the walk independently or as a guided tour.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
I
iPod as a Visitor’s Personal Guide
Figure 1. Designed main menu of a prototype device
automatically display the page as an index, rather than providing a list of available files to open. Secondly, the extension “.linx” of the filename defines the method used to display on the iPod screen a link to another text file. The iPod has two methods of displaying a link. The default is to have a link created within a text file appear as a hyperlink, similar to that of an html Web page with the word or sentence underlined. The second method is to display the link as an actual menu item on the iPod. This method would be ideal for the contents of our visitor’s guide. Once the “Main.linx” file has been created and located correctly, the next step will be the contents of the main page. This will create the major links to each of the locations that will contain information. This is achieved by opening up the “Main.linx” file in the Notepad and entering the following:
A prototype device is being developed using an Apple iPod (see Figure 1).
PERSONAL GUIDE DESIGN The visitor’s personal guide itself will be in the form of an iPod, which can be programmed to give details of a defined list of locations, as well as playing an audible narration of each featured location. This will allow visitors to navigate the visitor’s site on their own with the use of the iPod. The following technologies were considered initially to program the iPod as per the requirements: • • •
creating an application in J2ME on the Java platform porting it to the iPod; installing a variant of Linux (more on this later), and modifying its operation to create the system from this platform; and creating a text-based guide on the iPod.
The text-based option is the easiest way, with some limitations and given preference on the basis of the estimated time and skills available. To get multiple text pages to run is a fairly simple concept. It requires a specifically named file located under the “Notes” folder that acts as an index page, from which the menu would be created and all other notes will be created. As discussed earlier, iPods as storage devices can easily be connected to a computer via the USB port, and the drive that is mounted for the iPod can be navigated easily from “My Computer.”
Creating a Content Page Open up the “Notes” folder and create a new file called “Main.linx.” This file name is required for two reasons. The first reason is that by naming the file “Main,” the iPod will 366
Alternative Operating Systems for Apple’s iPod Currently there are two main alternatives to Apples’ iPod Operating System: iPodLinux (an open source venture into porting Linux onto the iPod) and Rockbox (an open source replacement firmware for mp3 players). iPodlinux (www.ipodlinux.org) and Rockbox Operating System (www.rockbox.org) are able to replace Apple’s Operating System and still maintain the same functionality. The alternative operating systems are capable of playing mp3s and other audio formats, videos, and reading notes. The main difference between Apple’s Operating System is that with iPodlinux and Rockbox you can: • • • •
play video games, run applications, simply develop your own applications without requiring commercial development tools, and programmers can develop their own applications or modify/customize existing iPodlinux GPL (General Public Licenxe) applications.
Certain Linux applications are recompiled to run on the iPod without modification. Both alternatives to Apple’s iPod Operating System have a following of enthusiastic programmers and developers who have figured out the workings of the five generations of the iPod. Developers and programmers of the iPodlinux have contributed a lot to an open source operating system by setting up Internet relay chat rooms, news groups, forums, wikis, and Web sites. Sourceforge hosts the source code and development comma separated value (CSV) tree, which is maintained by the iPodlinux core developers. Documentation of the iPod hardware components such as the microcontroller, display, memory, battery, and
iPod as a Visitor’s Personal Guide
so on is now accessible to everyone. Rockbox Operating System developers thank the hard work of the iPodlinux project because if it was not for iPodlinux documentation and developers, the Rockbox Operating System may have never been ported to the iPod.
why Choose an Alternative iPod Operating System
are happy with your application, it can be packaged as a module and inserted into the Podzilla menu structure.
ADVANTAGES This new application and use of iPod will:
The iPod as a visitor’s personal guide project initially was looking at the bleeding edge mobile Java J2ME application technology to fulfill its requirements. After research it was discovered that there are other ways to implement a tour guide on an iPod. The research found iPod Notes, iPodlinux, and Rockbox.
• •
•
•
• •
The iPod Apple Operating System is proprietary and therefore a close source. iPodLinux and Rockbox are open source operating systems written under the GNU General Public License. iPod Apple OS only supports a crippled html language in “Notes” which allows the development of interactive Notes that can contain pictures, video, and text.
how iPod Can be Programmed for iPodLinux Programming for iPodLinux is done in C, and as a prerequisite the standard functions and libraries must be used. Here is an example of the “Hello World” code using the print function from the stdio.h library. Using notepad or a C application, do the following: • • •
•
Start off by including the precompiler derivative includes statement: #include . Next create the main function from which we will put in the code to print Hello World (see Figure 2). Now save the code you have entered into the notepad using the filename of hworld.c. The step is to compile hworld.c using the arm complier tools, arm-elf-gcc hworld.c -o hworld -elf2flt. Executing hworld on the iPod running iPodlinux will display “Hello World” to the iPod screen. Once you
Figure 2. Sample main function created using programming language “C” int main (int argc, char **argv) { printf (“Hello World!\n” ); return 0; }
• •
eliminate the need for a human guide; provide a self-guided tour with entertainment to visitors; lead to interactive customer service; provide flexibility to tourists to tour the area per their own need, time, and interest, which is an important perspective; and achieve great tourist turnaround, as there is no need to wait for some predefined number of tourists.
FUTURE TRENDS This concept can further be considered to provide visitor information in other languages than English, with possible navigation for the use of different languages in a multicultural environment. Also our next application will discover the possibilities of porting iPodLinux platform on the fifth-generation iPod which is not done so far. Further, the possibility of using an iPod—similar to a PDA—in a commercial environment will be investigated. For now, this cost-effective solution can be implemented at various landmark tourist destinations such as mines, aquariums, and museums, and in the near future it will replace existing expensive technologies.
CONCLUSION We have demonstrated that the iPod can be used as an innovative and cost-effective tool. To realize the use of the iPod as a visitor’s personal guide, iPod’s simple user interface designed around a central scroll wheel can be explored for the navigation and recording/ playback facility. It provides the latest information to visitors. Furthermore, iTunes can add an entertaining experience with preferable music while using it as a personal guide.
ACKNOwLEDGMENTS We would like to thank our final-year students Andrew Obersnel, Ben Coster, Randima Sampath Ratnayake, and Pathum Wickasitha Thamawitage for their contributions toward this project. We would also like to thank Rob McAllister, 367
I
iPod as a Visitor’s Personal Guide
general manager of teaching, innovation, and degrees, Box Hill Institute; Stephen Besford, center manager, Center for Computer Technology, Box Hill Institute; and John Couch, administrative officer, Center for Computer Technology, Box Hill Institute.
REFERENCES Cheverst, K., Davies, N., Mitchell, K., Friday, A., & Efstratiou, C. (2000). Developing a context-aware electronic tourist guide: Some issues and experiences. Proceedings of CHI 2000 (pp. 17-24). The Hague: ACM Press. Clark, D., & Walsh, S. (2004). iPod learning. White Paper, Epic Group, Brighton, UK. Duke University. (2005). Duke iPod first year experience. Retrieved February 17, 2006, from http://www.duke.edu/ ipod/ Macworld UK. (2005). Music player sales ‘set to double’. Macworld UK, (July 22). Retrieved April 26, 2006, from http://www.macworld.co.uk/news/index.cfm?NewsID=92 18&pagePos=5 Physorg. (2006). Portable tourist guide now on service. Retrieved February 19, 2006, from http://www.physorg. com/news10338.html
368
Scottish Council for Educational Technology. (1994). Audio. In Technologies and learning (pp. 24-25). Glasgow: SCET. Tourism Australia. (2005). Visitors arrival data. Retrieved February 19, 2006, from http://www.tourism.australia. com/Research.asp?sub=0318&al=2100
KEY TERMS Java2 Macro Edition (J2ME): A Java platform especially for programming mobile devices such as PDAs. ers.
Linux: An open source operating system for comput-
MP3: MPEG-1 Audio Layer 3, of sound or music recordings stored in the MP3 format on computers. Personal Digital Assistant (PDA): A mobile device that can be used both as a mobile phone and a personal organizer primarily. Universal Serial Bus (USB) Port: Port used to connect devices to computers such as PCs, laptops, and Apple Macintosh computers.
Category: Mobile Phone 369
Keyword-Based Language for Mobile Phones Query Services Ziyad Tariq Abdul-Mehdi Multimedia University, Malaysia Hussein M. Azia Basi Multimedia University, Malaysia
INTRODUCTION A mobile system is a communications network in which at least one of the constituent entities—that is, user, switches, or a combination of both—changes location relative to another. With the advancements in wireless technology, mobile communication is growing rapidly. There are certain aspects exuded by mobile phones, which make them a high potential device for mobile business transactions. Firstly, there is a growing statistic on the number of users who own at least one mobile phone. In 2003 alone, the numbers of mobile phone users were as high as 1.3 billion, and this number is growing steadily each year. Secondly, more and more mobile phones are equipped with much better features and resources at a considerably lower price, which make them affordable to a larger number of users. And thirdly, and most importantly, the small size of mobile phones makes them easily transportable and can truly be the device for anywhere and anytime access (Myers & Beigl, 2003). Database querying, which is the interest of this article, is a kind of business transaction that can benefit from mobile phones. In general, database querying concerns the retrieval of information from stored data (kept in a database) based on the query (request) posed by the users. This aspect of the database transactions had been the focus of many database researchers for a long time. The mobile phone aspect of the transaction had only recently gained interest from the database communities, and these interests were mainly targeted to the “fatter” mobile devices. The work on mobile database querying can be grouped into those focused on the technology and techniques to handle the limitations of the mobile transmissions due to the instability of the mobile cellular networks, which were concentrated on developing applications that involved access to databases for the mobile environment, and those that handled both of the above issues. For example, caching (Cao, 2002) and batching (Tan & Ooi, 1998) are some of the popular techniques that were and still are investigated in detail to handle the problems of the mobile transmissions. On the other hand, Hung and Zhang’s (2003) telemedicine application, Koyama, Takayama, Barolli, Cheng, and Kamibayashi’s (2002) education application,
and Kobayashi and Paungma’s (2002) Boonsrimuang transportation application are some examples of the work on mobile database application. These works were observed to be application-specific and supported a very limited and predefined number and type of possible queries. Each of the possible queries, in turn, requires several interactions with the server before a full query can be composed. This article will highlight the framework opted by the authors in developing a database query system for usage on mobile phones. As the development work is still in progress, the authors will share some of the approaches taken in developing a prototype for a relationally complete database query language. This work concentrates on developing an application-independent, relationally complete database query language. The remainder of this article is organized as follows. The next section presents some of the existing work related to the study. We then introduce and describe the framework undertaken in order to develop a database query system for mobile phones, and discuss the prototype of the database query language used by the system. We end with our conclusion.
RELATED wORK Query languages are specialized languages for asking questions, or queries, which involve data in a database (Ramakrishnan & Gehrke, 2000). Query languages for relational databases originated in the 1970s with the introduction of relational algebra and relational calculus by E.F. Codd. Both relations are equivalent in their level of expressiveness or query completeness. These two formal relations had interchangeably been used as the benchmarks for measuring the completeness of the later query languages. Codd originally proposed eight operations to be included in the relational algebra, but out of the eight, five were considered fundamentals as they allowed most of the data retrieval operations. These operations are known as selection, projection, cartesian product, union, and set difference. If a query language supports the five operations, then it is considered as being relationally complete (Connolly, Begg, & Strachan, 1997). Throughout
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
K
Keyword-Based Language for Mobile Phones Query Services
the years, several other measures of query completeness were proposed such as datalog, stratified, computable, and others (Chandra, 1988). However, in the authors’ opinion, these later measures might be too extensive to be considered for mobile phones and their users’ application. Although both relational algebra and relational calculus are complete, they are hard to understand and use. This resulted in a search for other easy-to-use languages that are at least compatible to relational algebra and calculus. Some of these query languages are transform-oriented non-proceduralbased languages, which use relations to transform input data into required outputs. Structured Query Language (SQL) is an example of such a language. Besides non-procedural languages, visual query languages have also gained much acceptance in the database community. Some of the work on visual query languages found in the literature, such as Czejdo, Rusinkiewicz, Embley, and Reddy (1989), Catarci (1991), and Polyviou, Samaras, and Evripidou (2005), used the entity relationship diagram and other data modeling as the basis for query formulation, and some used icons to present pre-defined queries (Massari, Weissman, & Chrysanthis, 1996). Query languages are textual languages that caught the interest of some database query language researchers. Some of these languages were represented in the form of natural language (Kang, Bae, & Lee, 2002; Hongchi, Shang, & Ren, 2001), and some were represented in the form of keywords (Calado, da Silva, Laender, Ribeiro-Neto, & Vieira, 2004; Agrawal, Chaudhuri, & Das, 2001). This type of languages is less restrictive compared to the other types of languages. However, they need extra work in approximating the meaning of the terms or keywords used in a query. Thesaurus and ontology are few approaches used to approximate meanings of terms or keywords (Kimoto & Iwadera, 1991; Weibenberg, Voisard, & Gartmann, 2006) in this type of query language. Even though each type of query language mentioned above has its own advantages, very few of them, except for Polyviou et al. (2005) and Massari et al. (1996), were tested on small devices. SQL, for example, would be too tedious to be entered using mobile phones and too complicated for ordinary users. Visual query languages, on the other hand, would require considerable screen space as well as resource consuming in order to be rendered. Natural language and keyword language would also be difficult to be textually keyed in using mobile phones. There were, however, attempts to use spoken method for query languages (Chang et al., 2002; Bai, Chen, Chien, & Lee, 1998). But this approach leads to another problem in matching the intonation of users. The textual form of query languages (keyword method in particular) might be the most suitable language to be used on mobile phones since they are the least resource consuming and easily extensible. However, there must be a method to ease the input part of the query formulation process. To date, the authors have not been able to find any publication 370
of the investigations of the same method as applied to mobile phones. Therefore, we believe that the keyword-based language is worth some investigation.
Framework Model Polyviou, Samaras, and Evripiou (Kang et al., 2002) laid down several challenges that must be dealt with in order to develop a modern search interface. The challenges specified were: the search interface must be usable, powerful, flexible, and scalable. These challenges are adopted in our approach while developing the database query system for mobile phones. The concept of usable is implemented in our design by providing a language that supports menu-based guidance for the users to form valid queries. The concept of powerful is implemented by making sure that the language is relationally complete. The concept of flexible, on the other hand, is implemented in the language by allowing the language to work with any type of relational databases and any type of applications. Finally, the scalability aspect is handled by allowing the language to accept a database of any size, but at the same time filtering the data to be presented to the users according to some form of user grouping and access patterns. The keyword-based language is developed to answer the above challenges. The reasons for choosing such a language, among others, are due to the ability of such a language to present complex relationships with a minimal number of keywords, and it takes lesser space for presentation. For example, it is possible to access information from two indirectly linked relations, no matter how far apart the relations might be, by simply providing the name of the two relations as query keywords. This ability makes the language scalable and easily extensible. However, keyword-based language does have constraints. Firstly, it is in textual form and therefore is tedious and prone to typing error. Secondly, it requires users to have exact knowledge of the keywords to be entered in order to form valid queries. Therefore, the authors have modified the keyword-based approach by providing users with selectable keywords in a menu form. This approach has another advantage: it allows users to point and click the keywords needed without having to type them manually, which is a way to handle the input mechanism problem of mobile phones (most phones only have keypads as an input mechanism). This approach requires lengthy display space, which is lacking for mobile phones. Therefore, the authors intend to handle this problem by providing only selected keywords to users based on their personal profiles and preferred queries. Figure 1 shows the general framework of the query language, and Figure 2 shows the position of the query language with respect to the rest of the whole database query system. As shown in both figures, the query language basically resides in two locations: in the mobile phones as the query interface, and in the application server as the query engine that transformed the keyword
Keyword-Based Language for Mobile Phones Query Services
Figure 1. General framework for query language
Figure 2. The position for database query system
K
Figure 4. The operational stage of network architecture
Figure 3. Activities during pre-language operational stage
the mobile phones and query processing by the application and database servers.
queries into their relevant Structured Query Language (SQL) queries. The keywords that are presented as options during query formulation are selected metadata (i.e., relations’ names and attributes’ names) of the relations in the database. The selected metadata is based on the user information provided during login (currently personal information; later we will consider results from group training as well) and the accessibility information set by the database administrator prior to the query language becoming operational. Figure 3 shows the activities that are done by the database administrator and users during the pre-language operational stage. The operation of the database query system is depicted in the system architecture, as shown in Figure 4. The prelanguage operational stage will be conducted using a normal computing terminal over the Internet, while the query operations will be handled through query formulation using
EXPERIMENTAL STUDY The prototype of the query language is developed to identify the query interface midlet on the mobile phones and J2SE for the query engine servlets on the application server. The database consists of data on students, subjects, and staff of the university. Students enroll in many subjects that are conducted in many sessions at several venues. The subjects are taught by many lecturers who are of different ranks. The students stay in hostels managed by wardens, and their outings must be approved by a staff member. The schema in each relation in the database is shown in Figure 5 and their relationships in Table 1 respectively. The database is used to prove that the developed query language is relationally complete. As mentioned earlier, there are five fundamental operations that must be satisfied in order for a language to be considered relationally complete. 371
Keyword-Based Language for Mobile Phones Query Services
Table 1. The relationships between attributes ATTRIBUTE
REFER TO
Hostelwarden
Staff ID
Staff position
StaffPost code
Staff ranking
StaffRank code
Student hostel Code
Hostel code
Subject lecturer
Staff ID
Session subjectCode
Subject code
Session venue
Venue code
Outing studentID
Student ID
Outing approver
Staff ID
Enrolment studentID
Student ID
Enrolment subjectCode
Subject code
In order to explain the operations, let us assume that there are two relations R and S which contain m and n numbers of attributes respectively. The five operations are: Selection, Projection, Cartesian Product, Union, and Set Difference. Here we describe how the five operations are handled by the developed query language: •
Selection: Mathematically denoted as σpredicate(R), works on a single relation R and defines a relation that contains only those tuples of R that satisfy the specified condition (predicate). For example, a Selection query of σstudYear=3(Student) will produce as output, tuples from the Student relation which have a value 3 in their studYear attribute. The developed query language handles the Selection operation by allowing a user to select the proper relation name from the list of keywords. This action will allow the user to later choose the name of the attribute that he/she wants to check the value of, to choose the operation he/she wants to perform, and to provide the value he/she is looking for. Figure 6 show
•
•
the screen shots of a sample Selection operation. The query language also allows multiple conditions to be implemented. Projection: Mathematically denoted as πa1,a2,..,am(R), works on a single relation R and defines a relation that contains a vertical subset of R, extracting the values of specified attributes and eliminating duplicates. For example, a Projection query of πsubjCode,subjNam e(Subject) will produce as output all tuples in the Subject relation. For each tuple, the only values associated with the attributes of subjCode and subjName will be returned. The developed query language handles the Projection operation by allowing a user to select the proper relation name which later gives as a list all of the attributes of the relation. A user can then select as many attributes as he/she likes to view as output. Figure 7 shows the screen shots of a sample Projection operation. Cartesian Product: Mathematically denoted as R Χ S, defines a relation that is the concatenation of every tuple of relation R with every tuple of relation S. For example, if a Staff relation contains 100 tuples with 10 attributes each and a Subject relation contains 100 tuples with five attributes each, a Cartesian product query of Staff Χ Subject will produce as output 1,000 times 100 tuples, since each tuple of the Staff relation will be concatenated with each one of the tuples from the Subject relation. Furthermore, each tuple of the output relation will have ten plus five attributes. The result of a Cartesian product operation is less meaningful. Therefore, a more restrictive form of the operation, called Join, is more preferable. A Join operation, mathematically denoted as R predicate S, includes only the combinations of both relations that satisfy certain conditions. For example, a query of Staff staffID=lectID Subject will produce tuples which combine a tuple from the Staff with its associated Subject tuple. The number of the output tuples will be equal to the number of the tuples of Staff. There are several variations to the Join operation such as left-outer join,
Figure 5. The database schema Hostel (code, name, warden) Staff (ID, name, gender, DOB, mobile#, position, ranking, area) StaffPost (code, name) StaffRank (code, name) Student (ID, name, gender, DOB, year, hostelCode, room) Subject (code, name, creditHour, lecturer) Session (subjectCode, day, timeStart, timeEnd, venue) Venue (code, name, capacity) Enrolment (studentID, subjectCode) Outing (studentID, dateOut, tiemout, destination, dateIn, timeIn, approver)
372
Keyword-Based Language for Mobile Phones Query Services
Figure 6. Selection operation
Figure 8.
Figure 7.
K
be possibly implemented by allowing a user to select two relations and a difference operator. The query engine will then check for the compatibility of their schema types. Besides the five operations, the query language is capable of combining multiple operations in one query, and it can also be easily extended to include other operations as needed.
CONCLUSION
•
•
right-outer join, equijoin, and so on. The developed query language handles the Join operation by allowing a user to select as many relations as he/she wants to join and the relations can be indirectly related as well. Figure 8 shows screen shots of a sample Join operation. Union: Mathematically denoted as R ∪ S, concatenates all tuples of R and all tuples of S into one relation with duplicate tuples being eliminated. However, R and S must be union-compatible (i.e., both relations contain the same number of attributes, and each corresponding attribute is of the same domain) in order for the operation to be valid. For example, the query of Lecturing-Staff ∪ Administrative-Employee is valid if they both have the same schema type. This operation is not yet implemented by the query language. But the concept would be possibly implemented by allowing a user to select two relations and the union operator. The query engine will then check for the compatibility of their schema types. Set Difference: Mathematically denoted as R – S, defines a relation consisting of the tuples that are in relation R, but not in S. R and S again must be union-compatible. For example, the query of Staff–Lecturing-Staff will produce all administrative staff, assuming lecturers and administrators are the only two types of staff in the university. This operation is not yet implemented by the query language. Similarly, the concept would
The use of a keyword-based query language with menu-based guidance for formulating database queries using mobile phones is feasible due to its usability, powerfulness, flexibility, and scalability. With the physical constraints of the mobile phones, this type of query language uses minimal space for presentation and a lesser number of interactions to form complex queries. Furthermore, the keyword-based language is robust since it enables users to enter all possible queries by combining relevant keywords. Therefore, it is able to accept unplanned queries; it can be extended, and it is adaptable to other database applications. With further research, especially in the method for recommending the keywords relevant to a user, the keyword-based language could be the answer to access a full-scale database from mobile phones.
REFERENCES Agrawal, S., Chaudhuri, S., & Das, G. (2002, February 26March 1). DBXplorer: A system for keyword-based search over relational databases. Proceedings of the IEEE 18th International Conference on Data Engineering (ICDE’02) (pp. 5-16). Bai, B. R., Chen, C. L., Chien, L. F., & Lee, L. S. (1998). Intelligent retrieval of dynamic networked information from mobile terminals using spoken natural language queries. IEEE Transactions on Consumer Electronics, 44(1), 62-72.
373
Keyword-Based Language for Mobile Phones Query Services
Boonsrimuang, P., Kobayashi, H., & Paungma, T. (2002, July 3-5). Mobile Internet navigation system. Proceedings of the 5th IEEE International Conference on High Speed Networks and Multimedia Communications (pp. 325-328). Calado, P., da Silva, A.S., Laender, A. H. F., Ribeiro-Neto, B. A., & Vieira, R. C. (2004). A Bayesian Network approach to searching Web databases through keyword-based queries. Information Processing and Management, 40(5), 773-790. Cao, G. (2002). On improving the performance of cache invalidation in mobile environments. Mobile Networks and Applications, 7(4), 291-303. Catarci, T. (1991). On the expressive power of graphical query languages. Proceedings of the 2nd IFIP W.G. 2.6 Working Conference on Visual Databases (pp. 404-414). Chandra, A. (1988). Theory of database queries. Proceedings of the 7th ACM Symposium on Principles of Database Systems (pp. 1-9). Chang, E., Seide, F., Meng, H. M., Chen, Z., Shi, Y., & Li, Y. C. (2002). A system for spoken query information retrieval on mobile devices. IEEE Transactions on Speech and Audio Processing, 10(8), 531-541. Connolly, T., Begg, C., & Strachan, A. (1997). Database systems—A practical approach to design, implementation and management. Boston: Addison-Wesley. Czejdo, B., Rusinkiewicz, M., Embley, D., & Reddy, V. (1989, October 4-6). A visual query language for an ER data model. Proceedings of the IEEE Workshop on Visual Languages (pp.165-170). Hongchi, S., Shang, Y., & Ren, F. (2001, October 7-10). Using natural language to access databases on the Web. Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics (Vol. 1, pp. 429-434). Hung, K., & Zhang, Y.-T. (2003). Implementation of a WAPbased telemedicine system for patient monitoring. IEEE
374
Transactions on Information Technology in Biomedicine, 7(2), 101-107. Kang, I.-S., Bae, J.-H. J., & Lee, J. H. (2002, November 68). Database semantics representation for natural language access. Proceedings of the 1st International Symposium on Cyber Worlds (pp. 127-133). Kimoto, H., & Iwadera, T. (1991, July 8-14). A dynamic thesaurus and its application to associated information retrieval. Proceedings of IJCNN-91, the Seattle International Joint Conference on Neural Networks (Vol. 1, pp. 19-29). Koyama, A., Takayama, N., Barolli, L., Cheng, Z., & Kamibayashi, N. (2002, November 6-8). An agent based campus information providing system for cellular phone. Proceedings of the 1st International Symposium on Cyber Worlds (pp. 339-345). Massari, A., Weissman, S., & Chrysanthis, P. K. (1996). Supporting mobile database access through query by icons. Distributed and Parallel Databases (Special Issue on Databases and Mobile Computing), 4(3), 47-68. Myers, B. A., & Beigl, M. (2003). Handheld computing. Computer, 36(9), 27-29. Polyviou, S., Samaras, G., & Evripidou, P. (2005). A relationally complete visual query language for heterogeneous data sources and pervasive querying. Proceedings of IEEE 2005. Ramakrishnan, R., & Gehrke, J. (2000). Database management systems. New York: McGraw-Hill. Tan, K. L., & Ooi, B. C. (1998). Batch scheduling for demand-driven servers in wireless environments. Journal of Information Sciences, 109(1-4), 281-298. Weibenberg, N., Voisard, A., & Gartmann, R. (2006). Using ontologies in personalized mobile applications. Proceedings of the12th Annual International Workshop on GIS. ACM Press.
Category: Mobile Software Engineering 375
Knowledge Representation in Semantic Mobile K Applications Pankaj Kamthan Concordia University, Canada
INTRODUCTION Mobile applications today face the challenges of increasing information, diversity of users and user contexts, and ever-increasing variations in mobile computing platforms. They need to continue being a successful business model for service providers and useful to their user community in the light of these challenges. An appropriate representation of information is crucial for the agility, sustainability, and maintainability of the information architecture of mobile applications. This article discusses the potential of the Semantic Web (Hendler, Lassila, & Berners-Lee, 2001) framework to that regard. The organization of the article is as follows. We first outline the background necessary for the discussion that follows and state our position. This is followed by the introduction of a knowledge representation framework for integrating Semantic Web and mobile applications, and we deal with both social prospects and technical concerns. Next, challenges and directions for future research are outlined. Finally, concluding remarks are given.
BACKGROUND In recent years, there has been a proliferation of affordable information devices such as a cellular phone, a personal digital assistant (PDA), or a pager that provide access to mobile applications. In a similar timeframe, the Semantic Web has recently emerged as an extension of the current Web that adds technological infrastructure for better knowledge representation, interpretation, and reasoning. The goal of the mobile Web is to be able to mimic the desktop Web as closely as possible, and an appropriate representation of information is central to its realization. This requires a transition from the traditional approach of merely presentation to representation of information. The Semantic Webprovides one avenue towards that. Indeed, the integration of Semantic Webtechnologies in mobile applications is suggested in Alesso and Smith (2002) and Lassila (2005). There are also proof-of-concept semantic mobile applications such as MyCampus (Gandon & Sadeh, 2004) and mSpace Mobile (Wilson, Russell, Smith, Owens, & Schraefel, 2005) serving a specific community. However,
Table 1. Knowledge representation tiers in a semantic mobile application Semiotic Level
Semantic Mobile Web Concern and Technology Tier
Social
Trust
Pragmatic
Inferences
Semantic
Metadata, Ontology, Rules
Syntactic
Markup
Empirical
Characters, Addressing, Transport
Physical
Not Directly Applicable
Decision Support
Feasibility
these initiatives are limited by one or more of the following factors: the discussion of knowledge representation is one-sided and focuses on specific technology(ies) or is not systematic, or the treatment is restricted to specific use cases. One of the purposes of this article is to address this gap.
UNDERSTANDING KNOwLEDGE REPRESENTATION IN SEMANTIC MOBILE APPLICATIONS In this section, our discussion of semantic mobile applications is based on the knowledge representation framework given in Table 1. The first column addresses semiotic levels. Semiotics (Stamper, 1992) is concerned with the use of symbols to convey knowledge. From a semiotics perspective, a representation can be viewed on six interrelated levels: physical, empirical, syntactic, semantic, pragmatic, and social, each depending on the previous one in that order. The physical level is concerned with the representation of signs in hardware and is not directly relevant here. The second column corresponds to the Semantic Web“tower” that consists of a stack of technologies that vary across the technical to social spectrum as we move from bottom to top, respectively. The definition of each layer in this technology stack depends upon the layers beneath it. Finally, in the third column, we acknowledge that there are time, effort, and budgetary constraints on producing a
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Knowledge Representation in Semantic Mobile Applications
representation and include feasibility as an all-encompassing factor on the layers to make the framework practical. For example, an organization may choose not to adopt a technically superior technology as it cannot afford training or processing tools available that meet the organization’s quality expectations. For that, analytical hierarchy process (AHP) and quality function deployment (QFD) are commonly used techniques. Further discussion of this aspect is beyond the scope of the article. The architecture of a semantic mobile application extends that of a traditional mobile application on the server-side by: (a) expressing information in a manner that focuses on description rather than presentation or processing of information, and (b) associating with it a knowledge management system (KMS) consisting of one or more domain-specific ontologies and a reasoner. We now turn our attention to each of the levels in our framework for knowledge representation in semantic mobile applications.
Empirical Level of a Semantic Mobile Application This layer is responsible for the communication properties of signs. Among the given choices, the Unicode Standard provides a suitable basis for the signs themselves and is character-by-character equivalent to the ISO/IEC 10646 Standard Universal Character Set (UCS). Unicode is based on a large set of characters that are needed for supporting internationalization and special symbols. This is necessary for the aim of universality of mobile applications. The characters must be uniquely identifiable and locatable, and thus addressable. The uniform resource identifier (URI), or its successor international resource identifier (IRI), serves that purpose. Finally, we need a transport protocol such as the hypertext transfer protocol (HTTP) or the simple object access protocol (SOAP) to transmit data across networks. We note that these are limited to the transport between the mobile service provider that acts as the intermediary between the mobile client and the server. They are also layered on top of and/or used in conjunction with other protocols, such as those belonging to the Institute of Electrical and Electronics Engineers (IEEE) 802 hierarchy.
Syntactic Level of a Semantic Mobile Application This layer is responsible for the formal or structural relations between signs. The eXtensible Markup Language (XML) lends a suitable syntactical basis for expressing information in a mobile application.
376
The XML is supported by a number of ancillary technologies that strengthen its capabilities. Among those, there are domain-specific XML-based markup languages that can be used for expressing information in a mobile application (Kamthan, 2001). The eXtensible HyperText Markup Language (XHTML) is a recast of the HyperText Markup Language (HTML) in XML. XHTML Basic is the successor of compact HTML (cHTML) that is an initiative of the NTT DoCoMo, and of the Wireless Markup Language (WML) that is part of the wireless application protocol (WAP) architecture and an initiative of the Open Mobile Alliance (OMA). It uses XML for its syntax and HTML for its semantics. XHTML Basic has native support for elementary constructs for structuring information like paragraphs, lists, and so on. It could also be used as a placeholder for information fragments based on other languages, a role that makes it rather powerful in spite of being a small language. The Scalable Vector Graphics (SVG) is a language for two-dimensional vector graphics that works across platforms, across output resolutions, across color spaces, and across a range of available bandwidths; SVG Tiny and SVG Basic are profiles of SVG targeted towards cellular phones and PDAs, respectively. The Synchronized Multimedia Integration Language (SMIL) is a language that allows description of temporal behavior of a multimedia presentation, associates hyperlinks with media objects, and describes the layout of the presentation on a screen. It includes reusable components that can allow integration of timing and synchronization into XHTML and into SVG. SMIL Basic is a profile that meets the needs of resource-constrained devices such as mobile phones and portable disc players. Namespaces in XML is a mechanism for uniquely identifying XML elements and attributes of a markup language, thus making it possible to create heterogeneous (compound) documents (Figure 1) that unambiguously mix
Figure 1. The architecture of a heterogeneous XML document for a mobile device CSS Mobile Profile Style Sheet Point
Include
SMIL Basic Animation Link SVG Basic Graphic
XHTML Basic Placeholder Document
Knowledge Representation in Semantic Mobile Applications
elements and attributes from multiple different XML document fragments. Appropriate presentation on the user agent of information in a given modality is crucial. However, XML in itself (and by reference, the markup languages based on it) does not provide any special presentation semantics (such as fonts, horizontal and vertical layout, pagination, and so on) to the documents that make use of it. This is because the separation of the structure of a document from its presentation is a design principle that supports maintainability of a mobile application. The cascading style sheets (CSS) provides the presentation semantics on the client, and CSS mobile profile is a subset of CSS tailored to the needs and constraints of mobile devices. With the myriad of proliferating platforms, information created for one platform needs to be adapted for other platforms. The eXtensible Stylesheet Language Transformations (XSLT) is a style sheet language for transforming XML documents into other, including non-XML, documents. As an example, information represented in XML could be transformed on-demand using an XSLT style sheet into XHTML Basic or an SVG Tiny document, as appropriate, for presentation to users accessing a mobile portal via a mobile device. Representing information in XML provides various advantages towards archival, retrieval, and processing. It is possible to down-transform and render a document on multiple devices via an XSLT transformation, without making substantial modifications to the original source document. However, XML is not suitable for completely representing the knowledge inherent in information resources. For example, XML by itself does not provide any specific mechanism for differentiating between homonyms or synonyms, does not have the capabilities to model complex relationships precisely, is not able to extract implicit knowledge (such as hidden dependencies), and can only provide limited reasoning and inference capabilities, if at all. The combination of the layers until now forms the basis of the mobile Web. The next two layers extend that and are largely responsible for what could be termed as the semantic mobile Web.
Semantic Level of a Semantic Mobile Application This layer is responsible for the relationship of signs to what they stand for. The resource description framework (RDF) is a language for metadata that provides a “bridge” between the syntactic and semantic layers. It, along with RDF Schema, provides elementary support for classification of information into classes, properties of classes, and means to model more complex relationships among classes than possible with XML only. In spite of their usefulness, RDF/RDF Schema suffer from limited representational capabilities and non-
standard semantics. This motivates the need for additional expressivity of knowledge. The declarative knowledge of a domain is often modeled using ontology, an explicit formal specification of a conceptualization that consists of a set of concepts in a domain and relations among them (Gruber, 1993). By explicitly defining the relationships and constraints among the concepts in the universe of discourse, the semantics of a concept is constrained by restricting the number of possible interpretations of the concept. In recent years, a number of initiatives for ontology specification languages for the semantic Web, with varying degrees of formality and target user communities, have been proposed, and the Web Ontology Language (OWL) has emerged as the successor. Specifically, we advocate that OWL DL, one of the sub-languages of OWL, is the most suitable among the currently available choices for representation of domain knowledge in mobile applications due to its compatibility with the architecture of the Web in general; and the Semantic Webin particular benefits from using XML/RDF/RDF Schema as its serialization syntax, its agreement with the Web standards for accessibility and internationalization, well-understood declarative semantics from its origins in description logics (DL) (Baader, McGuinness, Nardi, & Schneider, 2003), and provides the necessary balance between computational expressiveness and decidability.
Pragmatic Level of a Semantic Mobile Application This layer is responsible for the relation of signs to interpreters. There are several advantages of an ontological representation. When information is expressed in a form that is oriented towards presentation, the traditional search engines usually return results based simply on a string match. This can be ameliorated in an ontological representation where the search is based on a concept match. An ontology also allows the logical means to distinguish between homonyms and synonyms, which could be exploited by a reasoner conforming to the language in which it is represented. For example, Java in the context of coffee is different from that in the context of an island, which in turn is different from the context of a programming language; therefore a search for one should not return results for other. Further, ontologies can be applied towards precise access of desirable information from mobile applications (Tsounis, Anagnostopoulos, & Hadjiefthymiades, 2004). Even though resources can be related to one another via a linking mechanism, such as the XML Linking Language (XLink), these links are merely structural constructs based on author discretion that do not carry any special semantics. Explicit declaration of all knowledge is at times not cost effective as it increases the size of the knowledge base, which 377
K
Knowledge Representation in Semantic Mobile Applications
Example 1. Ontological Inferences
becomes rather challenging as the amount of information grows. However, an ontology with a suitable semantical basis can make implicit knowledge (such as hidden dependencies) explicit. A unique aspect of ontological representation based for instance on OWL DL is that it allows logical constraints that can be reasoned with and enables us to derive logical consequencesthat is, facts not literally present in the ontology but entailed by the semantics. We have a semantic mobile portal for tourist information. Let Mont Tremblant, Laurentides, and Québec be defined as regions, and the subRegionOf property between regions be declared as transitive in OWL (see Example 1.) Then, an OWL reasoner should be able to derive that if Mont Tremblant is a sub-region of Laurentides, and Laurentides is a sub-region of Québec, then Mont Tremblant is also a sub-region of Québec. This would give a more complete set of search results to a semantic mobile application user. In spite of its potential, ontological representation of information presents certain domain-specific and human-centric challenges (Kamthan & Pai, 2006). Query formulations to a reasoner for extracting information from an ontology can be rather lengthy input on a mobile device. It is currently also difficult both to provide a sound logical basis to aesthetical, spatial/temporal, or uncertainty in knowledge, and represent that adequately in ontology.
Social Level of a Semantic Mobile Application This layer is responsible for the manifestation of social interaction with respect to the representation. Specifically, ontological representations are a result of consensus, which in turn is built upon trust. The client-side environment in a mobile context is constrained in many ways: devices often have restricted processing capability and limited user interface input/output facilities. The Composite Capabilities/Preference Profiles (CC/PP) Specification, layered on top of XML and RDF, allows the expression of user (computing environment and 378
Example 2. Device Profile ... MyMobileCompany ABC 200 320 16 ...
personal) preferences, thereby informing the server side of the delivery context. In Example 2, CC/PP markup for a device whose processor is of type ABC and the preferred default values of its display and memory as determined by its vendor are given. The namespace in XML is used to disambiguate elements/ attributes that are native to CC/PP or RDF from those that are specific to the vendor vocabulary. CC/PP can be used as a basis for introducing contextawareness in mobile applications (Sadeh, Chan, Van, Kwon, & Takizawa, 2003; Khushraj & Lassila, 2004). One of the major challenges to the personalization based on profile mechanism is the user concern for privacy. The Platform for Privacy Preferences Project (P3P) allows the expression of privacy preferences of a user, which can be used by agents to decide if they have the permission to process certain content, and if so, how they should go about it. This ensures that users are informed about privacy policies of the mobile service providers before they release personal information. Thus, P3P provides a balance to the flexibility offered by the user profiles in CC/PP. The Security Assertion Markup Language (SAML), XML Signature, and XML Encryption provide assurance of the sanctity of the message to processing agents. We note that an increasing number of languages to account for may place an unacceptable load, if it is to be processed exclusively, on the client side. We also acknowledge that these technologies alone will not solve the issue of trust, but when applied properly, could contribute towards it.
FUTURE TRENDS The transition of the traditional mobile applications to semantic mobile applications is an important issue. The previ-
Knowledge Representation in Semantic Mobile Applications
ous section has shown the amount of expertise and level of skills required for that. Although up-transformations are in general difficult, we anticipate that the move will be easier for the mobile applications that are well-structured in their current expression of information and in their conformance to the languages deployed. The production of mobile applications, and by extension semantic mobile applications, is becoming increasingly complex and resource (time, effort) intensive. Therefore, a systematic and disciplined approach for their development, deployment, and maintenance, similar to that of Web engineering, is needed. Related to that, the issue of quality of represented and delivered information will continue to be important. The studies of specific attributes such as usability (Bertini, Catarci, Kimani, & Dix, 2005) and “best practices” for mobile applications from the World Wide Web Consortium (W3C) Mobile Web Initiative are efforts that could eventually be useful in an “engineering” approach for producing future semantic mobile applications. The process of aggregation and inclusion of information in a mobile application is primarily manual, which can be both tedious and error prone. This process could be, at least partially, automated via the use of Web services where mobile applications could be made to automatically update themselves with (candidate) information. Therefore, manifestations of mobile applications through Semantic Webservices (Wagner & Paolucci, 2005; Wahlster, 2005) are part of a natural evolution.
REFERENCES
CONCLUSION
Kamthan, P., & Pai, H.-I. (2006, May 21-24). Human-centric challenges in ontology engineering for the Semantic Web: A perspective from patterns ontology. Proceedings of the 17th Annual Information Resources Management Association International Conference (IRMA 2006), Washington, DC.
For mobile applications to continue to provide a high quality-of-service (QoS) to their user community, their information architecture must be evolvable. The incorporation of Semantic Webtechnologies can be much more helpful in that regard. The adoption of these technologies does not have to be an “all or nothing” proposition: the evolution of a mobile application to a semantic mobile application could be gradual, transcending from one layer to another in the aforementioned framework. In the long term, the benefits of transition outweigh the costs. Ontologies form one of the most important layers in semantic mobile applications, and ontological representations have certain distinct advantages over other means of representing knowledge. However, engineering an ontology is a resource-intensive process, and an ontology is only as useful as the inferences (conclusions) that can be drawn from it. To be successful, semantic mobile applications must align themselves to the Semantic Webvision of inclusiveness for all. For that, the semiotic quality of representations, particularly that of ontologies, must be systematically assured and evaluated.
Alesso, H. P., & Smith, C. F. (2002). The intelligent wireless Web. Boston: Addison-Wesley. Baader, F., McGuinness, D., Nardi, D., & Schneider, P. P. (2003). The description logic handbook: Theory, implementation and applications. Cambridge University Press. Bertini, E., Catarci, T., Kimani, S., & Dix, A. (2005). A review of standard usability principles in the context of mobile computing. Studies in Communication Sciences, 1(5), 111-126. Gandon, F. L., & Sadeh, N. M. (2004, June 1-3). Contextawareness, privacy and mobile access: A Web semantic and multiagent approach. Proceedings of the 1st French-Speaking Conference on Mobility and Ubiquity Computing (pp. 123-130), Nice, France. Gruber, T.R. (1993). Toward principles for the design of ontologies used for knowledge sharing. Formal ontology in conceptual analysis and knowledge representation. Kluwer Academic. Hendler, J., Lassila, O., & Berners-Lee, T. (2001). The semantic Web. Scientific American, 284(5), 34-43. Kamthan, P. (2001, March 20-22). Markup languages and mobile commerce: Towards business data omnipresence. Proceedings of the WEB@TEK 2001 Conference, Québec City, Canada.
Khushraj, D., & Lassila, O. (2004, November 7). CALI: Context Awareness via Logical Inference. Proceedings of the Workshop on Semantic Web Technology for Mobile and Ubiquitous Applications, Hiroshima, Japan. Lassila, O. (2005, August 25-27). Using the Semantic Web in ubiquitous and mobile computing. Proceedings of the 1st International IFIP/WG 12.5 Working Conference on Industrial Applications of the Semantic Web (IASW 2005), Jyväskylä, Finland. Sadeh, N. M., Chan, T.-C., Van, L., Kwon, O., & Takizawa, K. (2003, June 9-12). A Semantic Web environment for context-aware m-commerce. Proceedings of the 4th ACM Conference on Electronic Commerce (pp. 268-269), San Diego, CA. Stamper, R. (1992, October 5-8). Signs, organizations, norms and information systems. Proceedings of the 3rd Australian 379
K
Knowledge Representation in Semantic Mobile Applications
Conference on Information Systems (pp. 21-55), Wollongong, Australia. Tsounis, A., Anagnostopoulos, C., & Hadjiefthymiades, S. (2004, September 13). The role of Semantic Web and ontologies in pervasive computing environments. Proceedings of the Workshop on Mobile and Ubiquitous Information Access (MUIA 2004), Glasgow, Scotland. Wagner, M., & Paolucci, M. (2005, June 9-10). Enabling personal mobile applications through Semantic Web services. Proceedings of the W3C Workshop on Frameworks for Semantics in Web Services, Innsbruck, Austria. Wahlster, W. (2005, June 3). Mobile interfaces to intelligent information services: Two converging megatrends. Proceedings of the MINDS Symposium, Berlin, Germany. Wilson, M., Russell, A., Smith, D. A., Owens, A., & Schraefel, M. C. (2005, November 7). mSpace mobile: A mobile application for the Semantic Web. Proceedings of the 2nd International Workshop on Interaction Design and the Semantic Web, Galway, Ireland.
KEY TERMS Delivery Context: A set of attributes that characterizes the capabilities of the access mechanism, the preferences of the user, and other aspects of the context into which a resource is to be delivered. Knowledge Representation: The study of how knowledge about the world can be represented and the kinds of reasoning can be carried out with that knowledge. Ontology: An explicit formal specification of a conceptualization that consists of a set of terms in a domain and relations among them. Personalization: A strategy that enables delivery that is customized to the user and user’s environment. Semantic Web: An extension of the current Web that adds technological infrastructure for better knowledge representation, interpretation, and reasoning. Semiotics: The field of study of signs and their representations. User Profile: An information container describing user needs, goals, and preferences.
380
Category: Location and Context Awareness 381
Location-Based Multimedia Content Delivery System for Monitoring Purposes Athanasios-Dimitrios Sotiriou National Technical University of Athens, Greece Panagiotis Kalliaras National Technical University of Athens, Greece
INTRODUCTION Advances in mobile communications enable the development and support of real-time multimedia services and applications. These can be mainly characterized by the personalization of the service content and its dependency to the actual location within the operational environment. Implementation of such services does not only call for increased communication efficiency and processing power, but also requires the deployment of more intelligent decision methodologies. While legacy systems are based on stationary cameras and operational centers, advanced monitoring systems should be capable of operating in highly mobile, ad-hoc configurations, where overall situation and users roles can rapidly change both in time and space, exploiting the advances in both the wireless network infrastructure and the user terminals’ capabilities. However, as the information load is increased, an important aspect is its filtering. Thus, the development of an efficient rapid decision system, which will be flexible enough to control the information flow according to the rapidly changing environmental conditions and criteria, is required. Furthermore, such a system should interface and utilize the underlying network infrastructures for providing the desired quality of service (QoS) in an efficient manner. In this framework, this article presents a location-based multimedia content delivery system (LMCDS) for monitoring purposes, which incorporates media processing with a decision support system and positioning techniques for providing the appropriate content to the most suitable users, in respect to user profile and location, for monitoring purposes. This system is based on agent technology (Hagen & Magendanz, 1998) and aims to promote the social welfare, by increasing the overall situation awareness and efficiency in emergency cases and in areas of high importance. Such a system can be exploited in many operational (public or commercial) environments and offers increased security at a low cost.
SERVICES The LMCDS provides a platform for rapid and easy set up of a monitoring system in any environment, without
any network configurations or time-consuming structural planning. The cameras can be installed in an ad hoc way, and video can be transmitted to and from heterogeneous devices using an Intelligent decision support system (IDSS) according to the user’s profile data, location information, and network capabilities. Users can dynamically install ad-hoc cameras to areas where the fixed camera network does not provide adequate information. The real-time transmission of still images or video in an emergency situation or accident to the available operational centers can instantly provide the necessary elements for the immediate evaluation of the situation and the deployment of the appropriate emergency forces. This allows the structure of the monitoring system to dynamically change according to on-the-spot needs. The IDSS is responsible for overviewing the system’s activity and providing multimedia content to the appropriate users. Its functionality lies in the following actions: • • •
identifying the appropriate user or group of users that need access to the multimedia content (either through user profile criteria or topological criteria); providing the appropriate multimedia content in relevance to the situation and the location; and adapting the content to the user’s needs due to the heterogeneity of the users devicesthat is, low bit rate video to users with portable devices.
The LMCDS can evaluate users’ needs and crisis events in respect to the topological taxonomy of all users and provide multimedia content along with geographical data. The location information is obtained through GPS or from GPRS through the use of corresponding techniques (Markoulidakis, Desiniotis, & Kypris, 2004). It also provides intelligent methodologies for processing the video and image content according to network congestion status and terminal devices. It can handle the necessary monitoring management mechanisms, which enable the selection of the non-congested network elements for transferring the appropriate services (i.e., video streaming, images, etc.) to the concerned users. It also delivers the service content in the most appropriate
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
L
Location-Based Multimedia Content Delivery System for Monitoring Purposes
Figure 1. System functionality and services
GSM/GPRS
WLAN WLAN
• • •
WLAN
LMCDS control messages
Low bitrate video, pictures
High bitrate video, pictures
format, thus allowing the cooperation of users equipped with different types of terminal devices. Moreover, the LMCDS provides notification services between users of the system for instant communication in case of emergency through text messaging or live video feed. All of the above services outline the requirements for an advanced monitoring system. The LMCDS functionality meets these requirements, since it performs the following features: • • • • • •
location-based management of the multimedia content in order to serve the appropriate users; differentiated multimedia content that can be transmitted to a wide range of devices and over different networks; lightweight codecs and decoders that can be supported by devices of different processing and network capabilities; IP-based services in order to be transparent to the underlying network technology and utilize already available hardware and operating systems platforms; intelligent delivery of the multimedia content through the LMCDS in order to avoid increased traffic payload as well as information overload; and diverse localization capabilities through both GPS and GPRS, and generation of appropriate topological data (i.e., maps) in order to aid users.
However, the system architecture enables the incorporation of additional location techniques (such as WLAN positioning mechanisms)) through the appropriate, but 382
simple, development of the necessary interfaces with external location mechanisms. In order to describe the above services in a more practical way, a short list of available options and capabilities of the system is given below. The target group of the LMCDS consists of small to medium businesses that seek a low-cost solution for monitoring systems or bigger corporations that need an ad hoc extension to their current system for emergency cases and in order to enhance their system’s mobility. Even though the users of the system consist mainly of security staff and trained personnel that are in charge of security, the system’s ease of use, low user complexity, and device diversity allow access even to common untrained users. The system offers a range of capabilities, most of which are summarized in Figure 1, such as:
•
• •
User registration and authentication. User profile (i.e., device, network interface). Location awareness: • User is located through positioning techniques. • User is presented with appropriate topographical information and metadata. • User is aware of all other users’ locations. • User can be informed and directed from a Center of Security (CS) to specified locations. Multimedia content: • Video, images, and text are transmitted to user in real time or off-line based on situation or topological criteria. • User can provide feedback from his device through camera (laptop, PDA, smart phone) or via text input. • Content is distributed among users from the CS as needed. Ad hoc installation of cameras that transmit video to CS and can take advantage of wireless technology (no fixed network needed). Autonomous nature of users due to agent technology used.
LMCDS ARChITECTURE The LMCDS is designed to distribute system functionality and to allow diverse components to work independently while a mass amount of information is exchanged. This design ensures that new users and services can be added in an ad hoc manner, ensuring future enhancements and allowing it to support existing monitoring systems. Multi-agent systems (MASs) (Nwana & Ndumu, 1999) provide an ideal mechanism for implementing such a heterogeneous and sophisticated distributed system in contrast to traditional software technologies’ limitations in communication and autonomy.
Location-Based Multimedia Content Delivery System for Monitoring Purposes
Figure 2. Architecture overview LMCDS
Location Server
GSM/GPRS/WLAN
Users
GSM/GPRS/WLAN
Application Broker
Geographical Content Server
Center of Security Server
Multimedia Content Server
The system is developed based on MAS and allows diversified agents to communicate with each other in an autonomous manner, resulting in an open, decentralized platform. Tasks are being delegated among different components based on specific rules, which relate to the role undertaken by each agent, and information is being composed to messages and exchanged using FIPA ACL (FIPA, 2002a). An important aspect for the communication model is the definition of the content language (FIPA, 2002b). Since LMCDS targets to a variety of devices, including lightweight terminals, the LEAP language (Berger, Rusitschka, Toropov, Watzke, & Schlichte, 2002) of the JADE technology has been exploited. In addition to the security mechanisms supported by the underlying network components, the JADE platform offers a security model (Poggi, Rimassa, & Tomaiuolo, 2001) that enables the delegation of tasks to respective agent components by defining certificates, and ensures the authentication and encryption of TCP connections through the secure socket layer (SSL) protocol (http://www.openssl.org/). The general architecture of the LMCDS is shown in Figure 2. The platform is composed of different agents offering services to the system which are linked by an application broker agent, acting as the coordinator of the system. These agents are the location server, the center of security server, the application broker, the geographical content server, and the multimedia content server. The latter two are discussed in later sections in more detail, while a brief description of the functionality of the others is given as follows. The location server agent is responsible for the tracking
of all users and the forwarding of location-based information to other components. The information is gathered dynamically and kept up-to-date according to specific intervals. The intelligence lies in the finding of the closest users to the demanded area, not only in terms of geographical coordinates, but also in terms of the topology of the environment. More information on location determination is given in a following section. The center of security server agent monitors all users and directs information and multimedia content to the appropriate users. It is responsible for notifying users in emergency situations, and also performs monitoring functions for the system and its underlying network. The User agent components manage information, including the transmission and reception of image or video, the display of location information, critical data, or other kinds of displays. They are in charge of several user tasks, such as updating users’ preferences and location, decoding the response messages from the application broker, and performing service level agreement (SLA) (ADAMANT, 2003) monitoring functionalities. The user agent can reside in a range of devices, since it is Java based. Finally, the application broker agent acts as a mediator that provides the interface between all of the components. It is responsible for prioritizing user requests according to users’ profiles, performing authentication functions, and acting as a facilitator for SLA negotiations. It also coordinates the communication process between users in order for the multimedia content to be delivered in the appropriate format according to the user’s processing and network capabilities, and also the network’s payload. A closer look into the components of each agent, along with the interaction between them, is shown in Figure 3. Each agent is composed of three main components: the graphical display, which is responsible for user input and information display; the communication unit, in charge of agent communication through ACL messages; and the decision processing unit, which processes all received information. In addition, the user and the CS include a multimedia content component for the capture, playback, and transmission of multimedia data through RTP or FTP channels.
MULTIMEDIA PROCESSING AND ENCODING Video/Image Formats One of the novelties is the ability of the system to perform real-time format conversions of the image or video data and transmit to several heterogeneous recipients, ranging from large PC servers to small personal devices like PDAs or mobile phones. 383
L
Location-Based Multimedia Content Delivery System for Monitoring Purposes
Figure 3. Agent components User
Center of Security Server Graphical Display
Graphical Display
RTP Multimedia Content
Multimedia Content Decision Processing Unit
Decision Processing Unit
FTP
Communication Unit
Communication Unit
ACL
ACL
Communication Unit ACL Communication Unit
Decision Processing Unit
Decision Processing Unit
Graphical Display
Location Server
Application Broker
The LMCDS enables the adoption and support of different video formats, according to the partial requirements of the user terminals and the available networks status. The most commonly used is the M-JPEG (http://www.jpeg .org/index.html) format. It was preferred over other common video formats suitable for wireless networks like MPEG (http://www. m4if.org/) and H.263 (http://www.itu.int/rec/ recommendation.asp), which provide higher compression ratio, because using them can require intense processing power both for the encoder and the decoder. Also, frames in MPEG or H.263 streams are inter-related, so a single packet loss during transmission may degrade video quality noticeably. On the contrary, M-JPEG is independent of such cascaded-like phenomena, and it is preferable for photovideo application temporal compression requirements for smoothness of motion. It is a lossy codec, but the compression level can be set at any desired quality, so the image degradation can be minimal. Also, at small data rates (5-20Kbps) and small frame rates, M-JPEG produces better results than MPEG or H.263. This is important, as the photos or video can be used as clues in legal procedures afterwards, where image quality is more crucial than smooth motion. Another offering feature is the easy extraction of JPEG (http://www.jpeg .org/index.html) images from video frames. The video resolution can be set in any industry-standard (i.e., subQCIF) or any other resolution of width and height dividable of 8, so the track is suitable for the device it is intended for. Video is streamed directly from the camera384
equipped terminals in a peer-to-peer manner. Transmission rates for the video depend on the resolution and the frame rate used. Some sample rates are given in Table 1. Apart from M-JPEG, another set of video formats have been adopted, such as H263 and MPEG-4. The development of these formats enables the testing and evaluation of the LMCDS, based on the network congestion and the current efficiency of the supported video formats and the crisis situations in progress. This means that for a specific application scenario, the encoding with M-JPEG format can lead to better quality on the user terminal side, while the MPEG 4 format can be effective in cases that the network infrastructure is highly loaded, so the variance in bit rate can keep the quality in high values. Image compression is JPEG with resolution of any width and height dividable of 8. For the real-time transmission of video stream, the real-time transfer protocol (RTP, http:// www.ietf.org/rfc/rfc3550.txt) is used, while for stored images and video tracks, the file transfer protocol (FTP) is used.
Video Processing It is important to point out that the output video formats can be produced and transmitted simultaneously with the use of the algorithm shown in Figure 4. Note that the image/video generator can be called several times for the same captured video stream as long as it is fed with video frames from the frame grabber. So, a single user can generate multiple live video streams with variations,
Location-Based Multimedia Content Delivery System for Monitoring Purposes
Table 1. Output video formats for the application Resolution Frame Suitable Rate Network 160 x 120
1
160 x 120
5
232 x 176
2
320 x 240 320 x 240 640 x 480
5 15 5
GPRS WLAN GPRS WLAN GPRS WLAN WLAN WLAN WLAN
Trans. Rate (kbps) 2–3 10 – 15
Figure 4. LMCDS video processing overview
Target Device Smartphone PDA, PC Smartphone, PDA,PC
10 – 15
PDA,PC
30 – 40 90 – 120 45 – 55
PC PC PC
not only in frame rate and size, but also in JPEG compression quality, color depth, and even superimpose layers with handmade drawings or text. The algorithm was implemented in Java with the use of the JMF API (http://java.sun. com/ products/java-media/jmf/). For a captured stream at a frame rate of n frames per second, the frame grabber component extracts from the raw video byte stream n/A samples per second, where A is a constant representing the processing power of the capturing device. Depending on how many different qualities of video streams need to be generated, m Image/Video Generator processes are activated, and each process i handles Ki fps. The following relationship needs to be applied:
Video Capturing
Raw Video Frame Buffer Multimedia Content Generator
File Writer
Video File
∑B
i
•
=m
Frame Grabber
JPEG Frame
JPEG Frame Array
JPEG File
Multimedia Database
•
L
Live Video Flow
JPEG Compression
Time & Place Stamping
K i = Bi ∗ n / A ∗ m , where Bi = (K i ∗ A ∗ m )/ n and
Raw Video Byte Stream
MJPEG Stream Generator
RTP Packet Generator
RTP Streamer JPEG File
FTP Client
N E T W O R K
GPS (Global Positioning System) is a global satellite system based on a group of non-geostatic satellites in middle altitude orbit (12,000 miles). The GPS-enabled devices have the ability to locate their position with a high degree of accuracy by receiving signals from at least six satellites. A GSM/GPRS subscriber can be located upon request, depending on the received power from adjacent cells. More information about these mechanisms can be found in Markoulidakis et al. (2004). The Ekahau position engine (http://www.ekahau. com/pdf/EPE_2.1_datasheet.PDF) is designed to locate 802.11b (at present) wireless LAN users positions in the indoor environment. In the context of radio resource management, the software could apply the access point’s radio coverage map in the indoor environment. It uses a calibration method. Initially it measures a set of sample points’ radio strength. Based on these sample points, the engine can estimate a client WLAN station’s approximate location, given an arbitrary combination of signal strengths.
There is also a latency of m*Ci seconds per video stream, where Ci depends on the transcoding time of the image to each final format. So, this tendency of the LMCDS to keep the frame rate low is inevitable due to this sampling process. However, using low frame rates is quite common in surveillance systems. It also allows long hours of recording, where video size is optimal when minimum both for storage and transmission, and is less demanding in processing power for use with video players running on small devices.
•
GEOGRAPhICAL CONTENT DELIVERY
Geographical Database
Positioning Methods
The geographical database of the LMCDS is storing information about the positions of the users that are registered in the system. The information is obtained regularly by scheduled queries. When the users perform service request messages to the system, their position coordinates are automatically
The LMCDS uses the following techniques for locating the users’ positions inside the served environment.
385
Location-Based Multimedia Content Delivery System for Monitoring Purposes
retrieved (through any of the above methods), so their location in the geographical DB is also updated. Special importance has been given to the ability of the system to serve queries about the relative position of its users and estimation of distances. So, the scheduled queries can be of the following types: • • • •
Find my 3 nearest users and send them a video. Estimate the time that I need to get to Building A. Find the users closest to point B and send pictures to those that operate in GPRS network and high-quality video to those in UMTS or WLAN. When user C enters a specified area, send him a message.
The user is displayed visual information in the form of maps to its device. The map consists of raster data—that is, the plan of the area and also several layers of metadata, showing points of interest, paths, as well as the position of relative users.
FUTURE TRENDS Future steps involve the exploitation of video streaming measurements for providing guaranteed QoS of the video content to the end user, as well as the better utilization of the available radio resources. Furthermore, the incorporation of new trends in video streaming in conjunction with a markup language for multimedia content, such as MPEG-7 or MPEG-21, can offer a higher level of personalized location-based services to the end user and are in consideration for future development.
CONCLUSION This article presented a location-based multimedia content system enabling real-time transfer of multimedia content to end users for location-based services. Based on the general architecture of multi-agent systems, it focused on fundamental features that enable the personalization of the service content and the intelligent selection of the appropriate users for delivering the selected content.
REFERENCES Berger, M., Rusitschka, S., Toropov, D., Watzke, M., & Schlichte, M. (2002). Porting distributed agent-middleware to small mobile devices. Proceedings of the Workshop on
386
Ubiquitous Agents on Embedded, Wearable, and Mobile Devices, Bologna, Italy. FIPA (Foundation for Intelligent Physical Agents). (2002a). FIPA ACL message structure specification. SC00061G. FIPA. (Foundation for Intelligent Physical Agents). (2002b). FIPA SL content language specification. SC00008I. Hagen, L., & Magendanz, T. (1998). Impact of mobile agent technology on mobile communication system evolution. IEEE Personal Communications, 5(4). IST ADAMANT Project. (2003). SLA management specification. IST-2001-39117, Deliverable D6. Markoulidakis, J. G., Desiniotis, C., & Kypris, K. (2004). Statistical approach for improving the accuracy of the CGI++ mobile location technique. Proceedings of the Mobile Location Workshop , Mobile Venue ’04. Nwana, H., & Ndumu, D. (1998). A perspective on software agents research. The Knowledge Engineering Review, 14(2). Poggi, A., Rimassa, G., & Tomaiuolo, M. (2001). Multi-user and security support for multi-agent systems. Proceedings of WOA 2001 Workshop, Modena, Italy.
KEY TERMS Agent: A program that performs some information gathering or processing task in the background. Typically, an agent is given a very small and well-defined task. Application Broker: A central component that helps build asynchronous, loosely coupled applications in which independent components work together to accomplish a task. Its main purpose is to forward service requests to the appropriate components. IDSS: Intelligent decision support system. LMCDS: Location-based multimedia content delivery system. MAS: Multi-agent system. Media Processing: Digital manipulation of a multimedia stream in order to change its core characteristics, such as quality, size, format, and so forth. Positioning Method: One of several methods and techniques for locating the exact or relative geographical position of an entity, such as a person or a device.
Category: Location and Context Awareness 387
Location-Based Multimedia Services for Tourists Panagiotis Kalliaras National Technical University of Athens, Greece Athanasios-Dimitrios Sotiriou National Technical University of Athens, Greece P. Papageorgiou National Technical University of Athens, Greece S. Zoi National Technical University of Athens, Greece
INTRODUCTION
SERVICES
The evolution of mobile technologies and their convergence with the Internet enable the development of interactive services targeting users with heterogeneous devices and network infrastructures (Wang et al., 2004). Specifically, as far as cultural heritage and tourism are concerned, several systems offering location-based multimedia services through mobile computing and multimodal interaction have already appeared in the European research community (LOVEUS, n.d.; Karigiannis, Vlahakis, & Daehne, n.d.). Although such services introduce new business opportunities for both the mobile market and the tourism sector, they are not still widely deployed, as several research issues have not been resolved yet, and also available technologies and tools are not mature enough to meet end user requirements. Furthermore, user heterogeneity stemming both from different device and network technologies is another open issue, as different versions of the multimedia content are often required. This article presents the AVATON system. AVATON aims at providing citizens with ubiquitous user-friendly services, offering personalized, location-aware (GSM Association, 2003), tourism-oriented multimedia information related to the area of the Aegean Volcanic Arc. Towards this end, a uniform architecture is adopted in order to dynamically release the geographic and multimedia content to the end users through enhanced application and network interfaces, targeting different device technologies (mobile phones, PDAs, PCs, and TV sets). Advanced positioning techniques are applied for those mobile user terminals that support them.
AVATON is an ambient information system that offers an interactive tour to the user (visitor) in the area of the Aegean Volcanic Arch (see http://www.aegean.gr/petrified_forest/). The system can serve both as a remote and as an onsite assistant for the visitor, by providing multimedia-rich content through various devices and channels: • •
•
over the Internet, via Web browsers with the use of new technologies such as rich-clients and multi-tier architecture in order to dynamically provide the content; with portable devices (palmtops, PDAs) and 2.5G or 3G mobile phones, which are capable of processing and presenting real-time information relevant to the user’s physical position or areas of interest; and via television channels—AVATON allows users to directly correlate geographic with informative space and conceivably pass from one space to the other, in the context of Worldboard (Spohrer, 1999).
With the use of portable devices equipped with positioning capabilities, the system provides: • • • •
dynamic search for geographical content, information related to users’ location, or objects of interest that are in their proximity; tours in areas of interest with the aid of interactive maps and 3-D representations of the embossed geography; search for hypermedia information relative to various geographic objects of the map; user registration and management of personal notes during the tour that can be recalled and reused during later times; and
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
L
Location-Based Multimedia Services for Tourists
Figure 1. The AVATON architecture
Figure 2. The XML-based technologies in the client side GPS Hardware
Server
Geographical Content DB
Multimedia Content DB GUI INTERFACE
Location Server
Content Server XHTML
Application Server
Web Server
TV Server
SVG
Event Dispatcher
SMIL
MESSAGE INTERFACE
GPS INTERFACE
Application Interfaces
SOAP
GML
XML Formats
Messaging Engine
Positioning Engine
W-LAN, GPRS, UMTS, INTERNET, RF XSL Transformer
GPRS, UMTS
INTERNET
interrelation of personal information with natural areas or objects for personal use or even as a collective memory relative to visited areas or objects.
ThE AVATON ARChITECTURE Overview The AVATON system is based on a client-server architecture composed of three main server components: the application server, the content server, and the location server. The application server combines information and content from the content and location servers, and replies to client requests through different network technologies. The content is retrieved from two kinds of databases, the geographical and multimedia content DBs. The above architecture is shown in Figure 1. In more detail: • • •
388
XML Processing
RF
W-LAN
•
XML Parser
Application Functionality
Multimedia Content Database: This database contains the multimedia content such as images, video, audio, and animation. Geographical Content Database: A repository of geographical content such as aerial photos, high-resolution maps, and relevant metadata. Content Server: The content server supplies the application server with multimedia content. It retrieves needed data from the multimedia content database according to user criteria and device capabilities, and responds to the application server.
•
•
Location Server: Serves requests for geographical content from the application server by querying the geographical content database. The content retrieved is transformed into the appropriate format according to user device display capabilities and network bandwidth available. Application Server: The application server receives requests from different devices through GPRS, UMTS (third-generation mobile phone), W-LAN, Internet (PDA, laptop, PC), and RF (television). The server identifies each device and transmits data in an appropriate format. More precisely, the application server incorporates a Web server and a TV server in order to communicate with PCs and televisions respectively.
Client This section focuses on the mobile-phone and PDA applications. The scope of the AVATON system includes Java-enabled phones with color displays and PDAs with WLAN or GPRS/UMTS connectivity. While all the available data for the application can be downloaded and streamed over the network, data caching is exploited for better performance and more modest network usage. When the users complete their registration in the system, they have in their disposal an interactive map that initially portrays the entire region as well as areas or individual points of interest. For acquiring user position, the system is using GPS. The client also supports multi-lingual implementation, as far as operational content is concerned, for example menus, messages, and help. These files are maintained as XML documents. XML is extensively used in order to ease the load of parsing different data syntaxes. A single process, the XML Parser is used for decoding all kinds of data and an XSL Transformer for transcoding them in new formats. The different XML formats are XHTML, SVG, SMIL, SOAP, and GML, as shown in Figure 2.
Location-Based Multimedia Services for Tourists
Geographical Info Presentation In order to render the geographical data, the client receives raster images for the drawing of the background map, combined with metadata concerning areas of interest and links to additional textual or multimedia information. The raster data are aerial high-resolution photographs of the region on two or three scales. Because of the high resolution of the original images, the client is receiving small portions, in the form of tiles from the raster data processing engine in the server side, which are used to regenerate the photorealistic image layer in a resolution that is suitable for the device used. The attributes of the geographical data are generated in vector ShapeFile (ESRI ShapeFile) format, which is quite satisfactory for the server side but not for lightweight client devices. So, a SHP TO SVG converter at the server side is regenerating the metadata in SVG format that can be viewed properly from a handheld device. As soon as the metadata is downloaded to the client device, a final filtering (XSLT transformation) is done and the additional layer is opposed to the image layer in the SVG viewer. On the SVG data layer, the user can interact with points of interest and receive additional information in the form of text or multimedia objects. The above are shown in Figure 3.
Multimedia Info Presentation The presentation of multimedia information mainly depends on user position. The system is designed to provide audio and video clips, 3-D representations, and also textual information concerning each place of interest. Not all devices, though, receive the same content, since they differ in display, processor, or network speed. For that purpose, for each registered
Figure 3. Client-side map rendering
Location Server The location server is the component that handles the geographical content of the AVATON system. It provides a storage system for all geographical data and allows querying of its contents through location criteria, such as global position and areas of interest. Content management is based on a PostgreSQL (http://www.postgres.org) relational database. A JDBCInterface uses the JDBC APIs in order to provide support for data operations. A GISExtension is also present, based on PostGIS (http://www.postgis.org), in order to enable the PostgreSQL server to allow spatial queries. This feature is utilized through a GISJDBCInterface, a PostGIS layer on top of PostgreSQL.
Figure 4. Client-side map rendering
Client XSLT Transformation - Data Filtering
user, the system decides what kind of content is more suitable for them to receive and the multimedia content server generates the appropriate script. Depending on the available memory of the client’s device, media objects stay resident in the cache memory so that frequently requested content is accessed without delays that occur due to network latency. In Figure 4 the components that are involved in the multimedia presentation are shown. The TourScript Data contains the script which describes the multimedia presentation. It is transcoded inside the SMIL generator to a SMIL message that follows the XML syntax, so that it can be incorporated seamlessly to the messages that are exchanged in the AVATON system. At the client side, the SMIL message is received by the SMIL processor which coordinates the process of fetching the multimedia objects from the client cache memory to the suitable renderer, so that the multimedia presentation can be completed.
Filtered SVG Data
SVG Viewer SVG Data Layer Image Layer
Client SMIL Processor
SMIL Presentation
Object Selection SVG Data
Image Tiles
Multimedia Presentation
Client Cache Memory
XML - SMIL Message Client
Server SHP to SVG Converter SHP Data Attribute Data
Raster Data Processing Engine
SMIL Generator
Raster Data Spatial Data
TourScript Data
Multimedia Objects
389
L
Location-Based Multimedia Services for Tourists
Cartographic Data Concerning the photorealistic information, the user can choose from several distinct zoom levels. The mobiles phones and PDAs in the market that support GPRS or WLAN have displays of different resolutions that, in most cases, are multiples of 16 pixels. Hence, the location server can generate tiles with a multiple of 16p x 16p, which can be presented in the user’s mobile device. The server always holds multiple resolutions for every level of cartographic (photographic) information. The levels of cartographic information define the degree of focus. Apart from the photographic layer, additional layers of vector information also exist, and their size is approximately 5% or 10% of the corresponding photographic. Therefore, in practice, every device will initially request from the server the cartographical information with the maximum resolution this device can support. Hence, the server decides the available resolution that best corresponds to the requested resolution from the device. The size of every tile is approximately 1Κ. The devices with greater resolution per tile receive more files, with greater magnitude, for every degree of focus due to the higher resolution.
Multimedia Content Server The multimedia content server component comprises the major unit that controls the mixing and presentation of different multimedia objects. Its purpose is to upload all the objects necessary and present them in a well-defined controlled order that in general depends on the user position, interactions, and tracking information available. The multilingual audiovisual information scheduled for presentation is coordinated so that several objects may be presented simultaneously. The multimedia content server component is also responsible for choosing “relevant” objects for the user to select among in the case the user requires more information on a topic. The multimedia content server interfaces with the multimedia content database, a relational database storing the multimedia content. The database is organized thematically and allows the creation of hierarchical structures. It also contains a complete list of multimedia material, covering all content of the physical site, such as 3D reconstructed plants, audio narration, virtual 3D models, avatar animations, and 2D images.
Media Objects As mentioned already, the multimedia content server is responsible for mixing the basic units of multimedia information. These elements are hierarchically ordered. At the finest level of granularity, there are atomic objects called MediaObjects with specializations such as AudioMediaObject, 390
Figure 5. Hierarchy formulation of media objects at the multimedia content server Area of Interest 1..N TourScript
1..N MediaObject AudioMediaObject ImageMediaObject 3DMediaObject CompositeMediaObject
1..N Area of Interest
1..N TourScript Script MediaObjects
1..N Point of Interest
1..N TourScript
ImageMediaObject, 3DMediaObject, and CompositeMediaObject. These objects contain the actual data to be rendered along with additional profile metadata characterizing them. At a higher level of complexity, a TourScript represents an ordered sequence of MediaObjects, all of which are to be presented if the script is chosen. According to user requirements, the user will be able to navigate through the site in a geographically based tree. This is made possible through the use of points of interest (PoI) and areas of interest (AoI). A PoI can only contain TourScripts and can be viewed as the end node of the site tree. In contrast, an AoI may contain either another PoI, an AoI, or TourScripts. This allows the system to map the actual site into a hierarchy model containing PoI at the top and MediaObject components at the leaf level. The multimedia content server is also responsible for managing this site-tree for the entire site. Moreover it is responsible for traversing it. The use of the site tree is quite interesting: when a media object, for instance audio object, is presented to the user, it belongs to a node in the site hierarchy. Figure 5 shows the structure of the site in a tree view as described previously. The multimedia content server is responsible for coordinating the rendering components in order to provide a synchronized presentation to the user, according to user preferences, position, and commands.
Deployment and Usage Based on the proposed architecture, the AVATON services are being deployed to physical sites within the Aegean Volcano Arc (such as Santorini and Lesvos islands) and evaluated by real end users under different scenarios. The main air interfaces that will be used by the system (along
Location-Based Multimedia Services for Tourists
Figure 6. Implementation plan for the Lesvos island site
Figure 7. SVG map and interface
L
Hill Access Point
Kiosk
Access Point
Trunk
Hollow
Access Point
with standard wired access through common LANs or the Internet) are: •
•
WLAN: A standard 802.11b wireless access network provides connectivity for users equipped with portable devices (such as PDA with built-in WLAN cards or laptops). GPRS/UMTS: For mobile users and smart phones, access will be provided through the GSM network, using GPRS (Bettsteter, Vögel, & Eberspächer, 1999). This restricts the system from providing video or 3-D animations to such users, and the services offered are focused on text, images (including map information), and short audio. As GPRS is already packet oriented, our implementation can be easily transferred to UMTS, if available.
In Figure 6, the actual implementation plan is given for the location of the Sigri Natural History Museum on Lesvos island. The area consists of an open geological site, the Petrified Forest, where the ash from a volcanic eruption some 15 to 20 million years ago covered the stand of sequoia trees, causing their petrification. Wireless access is provided by the use of a 3 Netgear 54Mbps access point equipped with additional Netgear antennas in order to overcome the physical limitations of the area (hills, trunks, and hollows). Visitors to the site are equipped with PDAs or smart phones (provided at the entrance kiosk) and stroll around the area. A typical scenario consists of the following: The users enter the archaeological site and activate their devices. They then perform a login and provide personal details to the server, such as username, language selection, and device
settings. The client then requests from the server and loads the map of the area in SVG (http://www.w3.org/TR/SVG) format, as seen in Figure 7. The circles on the map present distinct points of interest (yellow indicating trees and blue indicating leafs). The application monitors the users’ location and updates the SVG map in real time, informing the users of their position. The SVG map is interactive, and when the users enter the vicinity of a point of interest, the application automatically fetches and displays (via their browser or media player) the corresponding multimedia content (in the form of HTML pages, audio, or video) at the requested language. The users are also able to navigate manually through the available content and receive additional information on topics of their interest. During the tour, the users’ path is being tracked and displayed (the red line on Figure 7) in order to guide them through the site. They are also able to keep notes or mark favorite content (such as images), which can be later sent to them when they complete the tour.
FURThER wORK The system is currently being deployed and tested in two archeological sites. Users are expected to provide useful feedback on system capabilities and assist in further enhancements of its functionality. Also, as 3G infrastructure is being expanded, incorporation of the UMTS network in the system’s access mechanisms will provide further capabilities for smart phone devices and also use of the system in areas where wireless access cannot be provided. Towards commercial exploitation, billing and accounting functionalities will be incorporated into the proposed architecture. Finally, possible extensions of the system are 391
Location-Based Multimedia Services for Tourists
considered in order to include other cultural or archeological areas.
REFERENCES Bettsteter, C., Vögel, H.-J., & Eberspächer, J. (1999). GSM Phase 2+ General Packet Radio Service GPRS: Architecture, protocols, and air interface. IEEE Communication Surveys, (3rd Quarter). ESRI ShapeFile Technical Description. (1998, July). An ESRI white paper. GSM Association. (2003, January). Location based services. SE.23, 3.10. Karigiannis, J. N., Vlahakis, V., & Daehne, P. (n.d.). ARCHEOGUIDE: Challenges and solutions of a personalized augmented reality guide for archeological sites. Computer Graphics in Art, History and Archeology, Special Issue of the IEEE Computer Graphics and Application Magazine. LOVEUS. (n.d.). Retrieved from http://loveus.intranet. gr/documentation.htm Spohrer, J. C. (1999). Information in places. IBM System Journal, 38(4).
392
Wang, Y., Cuthbert, L., Mullany, F. J., Stathopoulos, P., Tountopoulos, V., Sotiriou, D. A., Mitrou, N., & Senis, M. (2004). Exploring agent-based wireless business models and decision support applications in an airport environment. Journal of Telecommunications and Information Technology, (3).
KEY TERMS Cartographic Data: Spatial data and associated attributes used by a geographic information system (GIS). Content Provider: A service that provides multimedia content. Location-Based Services: A way to send custom advertising and other information to cell phone subscribers based on their curent location. Multimedia: Media that usesmultiple forms of information content and information processing (e.g., text, audio, graphics, animation, video, interactivity) to inform or entertain the (user) audience. Network: A network of telecommunications links arranged so that data may be passed from one part of the network to another over multiple links. Tourism: The act of travel for predominantly recreational or leisure purposes, and the provision of services in support of this act.
Category: Location and Context Awareness 393
Location-Based Services Péter Hegedüs Budapest University of Technology and Economics, Hungary Mihály Orosz Budapest University of Technology and Economics, Hungary Gábor Hosszú Budapest University of Technology and Economics, Hungary Ferenc Kovács Budapest University of Technology and Economics, Hungary
INTRODUCTION The basically two different technologies, the location-based services in the mobile communication and the well-elaborated multicast technology, are joined in the multicast over LBS solutions. As the article demonstrates, this emerging and new management area has many possibilities that have not been completely utilized. Currently an important area of mobile communications is the ad-hoc computer networking, where mobile devices need base stations; however, they form an overlay without any Internet-related infrastructure, which is a virtual computer network among them. In their case the selective, locationrelated communication model has not been elaborated on completely (Ibach, Tamm, & Horbank, 2005). One of the various communication ways among the software entities on various mobile computers is the one-to-many data dissemination that is called multicast. Multicast communication over mobile ad-hoc networks has increasing importance. This article describes the fundamental concepts and solutions, especially focusing on the area of location-based services (LBSs) and the possible multicasting over the LBS systems. This kind of communication is in fact a special case of the multicast communication model, called geocast, where the sender disseminates the data to that subset of the multicast group members in a specific geographical area. The article shows that the geocast utilizes the advantages of the LBS, since it is based on the location-aware information being available in the location-based solutions (Mohapatra, Gui, & Li, 2004). There are several unsolved problems in LBS, in management and low surfaces. Most of them are in quick progress, but some need new developments. The product managers have to take responsibility for the software and hardware research and development part of the LBS product. This is a very important part of the design process, because if the development engineer leaves the product useful out of
consideration, the whole project could possibly be led astray. Another import question is that of LBS-related international and national laws, which could throw an obstacle into LBS’s spread. These obstacles will need to be solved before LBS global introduction. The article presents this emerging new area and the many possible management solutions that have not been completely utilized.
BACKGROUND Location-based services are based on the various distances of mobile communications from different base stations. With advances in automatic position sensing and wireless connectivity, the application range of mobile LBS is rapidly developing, particularly in the area of geographic, tourist, and local travel information systems (Ibach et al., 2005). Such systems can offer maps and other area-related information. The LBS solutions offer the capability to deliver location-aware content to subscribers on the basis of the positioning capability of the wireless infrastructure. The LBS solutions can push location-dependent data to mobile users according to their interests, or the user can pull the required information by sending a request to a server that provides location-dependent information. LBS may have many useful applications in homeland security (HLS). A few of the more significant of these applications are security and intelligence operations, notification systems for emergency responders, search and rescue, public notification systems, and emergency preparedness (Niedzwiadek, 2002). Mobile security and intelligence operatives can employ LBS to aid in monitoring people and resources in space and time, and they can stay connected with emergency operations centers to receive the necessary updates regarding the common operating picture for a situation. Emergency operations centers can similarly coordinate
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
L
Location-Based Services
search and rescue operations. Call-down systems can be employed to notify the public in affected disaster areas. In this application the multicast communication is preferable, since it uses in an efficient way the communication channels, which can be partly damaged after a disaster. Location-based public information services can give time-sensitive details concerning nearest available shelters, safe evacuation routes, nearest health services, and other public safety information (Niedzwiadek, 2002). Location-based services utilize their ability of locationawareness to simplify user interactions. With advances in wireless connectivity, the application range of mobile LBSs is rapidly developing, particularly in the field of tourist information systemstelematic, geographic, and logistic information systems. However, current LBS solutions are incompatible with each other since manufacturer-specific protocols and interfaces are applied to aggregate the various system components for positioning, networking, or payment services. In many cases, these components form a rigid system. If such system has to be adapted to another technology, for example, moving from global positioning system (GPS)-based positioning to in-house IEEE 802.11a-based wireless local-area network (WLAN) or Bluetooth-based positioning, it has to be completely redesigned (Haartsen, 1998). In such a way the ability of interoperation of different resources under changeable interconnection conditions becomes crucial for the end-to-end availability of the services in mobile environments (Ibach, & Horbank, 2004). There are a lot of location determination methods and technologies, such as the satellite-based GPS, which is widely applied (Hofmann-Wellenhof, Lichtenegger, & Collins, 1997). The three basic location determination methods are proximity, triangulation (lateration), and scene analysis or pattern recognition (Hightower & Borriello, 2001). Signal strength is frequently applied to determine proximity. As a proximity measurement, if a signal is received at several known locations, it is possible to intersect the coverage areas of that signal to calculate a location area. If one knows the angle of bearing (relative to a sphere) and distance from a known point to the target device, then the target location can be accurately calculated. Similarly, if somebody knows the range from three known positions to a target, then the location of the target object can be determined. A GPS receiver uses range measurements to multiple satellites to calculate its position. The location determination methods can be server based or client based according to the locus of computation (Hightower & Borriello, 2001). Chen et al. (2004) introduce an enabling infrastructure, which is a middleware in order to support location-based services. This solution is based on a location operating reference model (LORE) that solves many problems of constructing location-based services, including location modeling, positioning, tracking, location-dependent query processing,
394
and smart location-based message notification. Another interesting solution is the mobile yellow page service. The LBS is facing technical and social challenges, such as location tracking, privacy issues, positioning in different environments using various locating methods, and the investment of location-aware applications. An interesting development is the Compose project, which aims to overcome the drawbacks of the current solutions by pursuing a service-integrated approach that encompasses pre-trip and on-trip services, considering that on-trip services could be split into in-car and last-mile services (Bocci, 2005). The pre-trip service means the 3D navigation of the users in a city environment, and the on-trip service means the in-car and the last-mile services together. The in-car service is the location-based service and the satellite broadcasting/multicasting. In this case, the user has wireless-link access by PDA to broadcast or multicast. The last-mile service helps the mobile user with a PDA to receive guidance during the final part of the journey. In order to create applications that utilize multicast over LBS solutions, the middleware platform of LocatioNet Systems provides all required service elements such as end-user devices, service applications, and position determination technologies, which are perfectly integrated (LocatioNet, 2006). The middleware platform of LocatioNet has an open API and a software development kit (SDK). Based on these, the application developers are able to easily implement novel services, focusing on comfortable user interface and free from complex details of the LBS (LocatioNet, 2006). The article focuses on the multicast solutions over the current LBS solutions. This kind of communication is in fact a special case of the multicast communication model, called geocast, where the sender disseminates the data to a subset of the multicast group members that are in a specific geographical area.
MULTICASTING The models of the multicast communication differ in the realization of the multiplication function in the intermediate nodes. In the case of the datalink level, the intermediate nodes are switches; in the network level, they are routers; and in the application level, the fork points are applications on hosts. The datalink-level-based multicast is not flexible enough for new applications therefore it has no practical importance. The network-level multicast (NLM)named IP-multicastis well elaborated, and sophisticated routing protocols are developed for it. However, it has not been deployed widely yet since routing on the whole Internet has not been solved perfectly. The application-level solution gives less efficiency compared to the IP-multicast, however
Location-Based Services
its deployment depends on the application itself and it has no influence on the operation of the routers. That is why the application-level multicast (ALM) currently has an increasing importance. There are a lot of various protocols and implementations of the ALM for wired networks. However, the communication over wireless networks further enhances the importance of the ALM. The reason is that in the case of mobile devices, the importance of ad-hoc networks is increasing. Ad-hoc is a network that does not need any infrastructure. Such networks are Bluetooth (Haartsen, 1998) and mobile ad hoc network (MANET), which comprise a set of wireless devices that can move around freely and communicate in relaying packets on behalf of one another (Mohapatra et al., 2004). In computer networking, there is a weaker definition of this ad-hoc network. It says that ad-hoc is a computer network that does not need routing infrastructure. It means that the mobile devices that use base stations can create an ad-hoc computer network. In such situations, the usage of application-level networking (ALN) technology is more practical than IP-multicast. In order to support this group communication, various multicast routing protocols are developed for the mobile environment. The multicast routing protocols for ad-hoc networks differ in terms of state maintenance, route topology, and other attributes. The speed and reliability of sharing the information among the communication software entities on individual hosts depends on the network model and the topology. Theory of peer-to-peer (P2P) networks has gone through a great development in past years. Such networks consist of peer nodes. Usually, registered and reliable nodes connect to a grid, while P2P networks can tolerate the unreliability of nodes and the quick change of their numbers (Uppuluri, Jabisetti, Joshi, & Lee, 2005). Generally the ALN solutions use the P2P communication model, and multicast services overlay the P2P target created by the communicating entities. The simplest ad-hoc multicast routing methods are flooding and tree-based routing. Flooding is very simple, which offers the lowest control overhead at the expense of generating high data traffic. This situation is similar to the traditional IP-multicast routing. However, in a wireless ad-hoc environment, the tree-based routing fundamentally differs from the situation in wired IP-multicast, where treebased multicast routing algorithms are obviously the most efficient ones, such as in the multicast open shortest path first (MOSPF) routing protocol (Moy, 1994). Though treebased routing generates optimally small data traffic on the overlay in the wireless ad-hoc network, the tree maintenance and updates need a lot of control traffic. That is why the two simplest methods are not scalable for large groups. A more sophisticated ad-hoc multicast routing protocol is the core-assisted mesh protocol (CAMP), which belongs to the mesh-based multicast routing protocols (GarciaLuna-Aceves & Madruga, 1999). It uses a shared mesh to
support multicast routing in a dynamic ad-hoc environment. This method uses cores to limit the control traffic needed to create multicast meshes. Unlike the core-based multicast routing protocol as the traditional protocol independent multicast-sparse mode (PIM-SM) multicast routing protocol (Deering et al., 1996), CAMP does not require that all traffic flow through the core nodes. CAMP uses a receiver-initiated method for routers to join a multicast group. If a node wishes to join to the group, it uses a standard procedure to announce its membership. When none of its neighbors are mesh members, the node either sends a join request toward a core or attempts to reach a group member using an expanding-ring search process. Any mesh member can respond to the join request with a join acknowledgement (ACK) that propagates back to the request originator. In contrast to the mesh-based routing protocols, which exploit variable topology, the so-called gossip-based multicast routing protocols exploit randomness in communication and mobility. Such multicast routing protocols apply gossip as a form of randomly controlled flooding to solve the problems of network news dissemination. This method involves member nodes talking periodically to a random subset of other members. After each round of talk, the gossipers can recover their missed multicast packets from each other (Mohapatra et al., 2004). In contrast to the deterministic approaches, this probabilistic method will better survive a highly dynamic ad hoc network because it operates independently of network topology, and its random nature fits the typical characteristics of the network.
ThE LOCATION-AwARE MULTICAST An interesting type of ad-hoc multicasting is the geocasting. The host that wishes to deliver packets at every node in a certain geographical area can use such a method. In such case, the position of each node with regard to the specified geocast region implicitly defines group membership. Every node is required to know its own geographical location. For this purpose they all can use the global positioning system (GPS). The geocasting routing method does not require any explicit join and leave actions. The members of the group tend to be clustered both geographically and topologically. The geocasting routing exploits the knowledge of location. The geocasting can be combined with flooding. Such methods are called forwarding zone methods, which constrain the flooding region. The forwarding zone is a geographic area that extends from the source node to cover the geocast zone. The source node defines a forwarding zone in the header of the geocast data packet. Upon receiving a geocast packet, other machines will forward it only if their location is inside the forwarding zone. The location-based multicast (LBM) is an example for such geocasting-limited flooding (Ko & Vaidya, 2002). 395
L
Location-Based Services
lbm geocasting and ip multicast Using LBM in a network where routers are in fixed locations and their directly connected hosts are in a short distance, the location of the hosts can be approximated with the location of their router. These requirements are met by most of the GSM/ UMTS, WIFI/WIMAX, and Ethernet networks, therefore a novel IP layer routing mechanism can be introduced. This new method is a simple geocasting-limited flooding, extending the normal multicast RIB with the geological location of the neighbor routers. Every router should know its own location, and a routing protocol should be used to spread location information between routers. The new IP protocol is similar to the UDP protocol, but it extends it with a source location and a radius parameter. The source location parameter is automatically assigned by the first router. When a router receives a packet with empty source location, it assigns its own location to it. The radius parameter is assigned by the application itself, or it can be an administratively defined parameter in routers. Routers forward received packets to all their neighbors except the neighbor the packet arrived from, and the neighbors outside the circle defined by the source location and radius parameters. If a packet arrives from more than one neighbor, only the first packet is handled; the duplicates are dropped. This method requires changes in routing operating systems, but offers an easy way to start geocasting services on existing IP infrastructure without using additional positioning devices (e.g., GPS receiver) on every sender and receiver. The real advantages of the method are that geocasting services can be offered for all existing mobile phones without any additional device or infrastructure.
Product Management The product managers have to take responsibility for the software and hardware part of the LBS product. This is a very important part of the process because, if the development engineer leaves a useful product out of consideration, the whole project could possibly be led astray. There are several product-management systems that help to coordinate the whole process.
Future Trends The multicast communication over mobile ad-hoc networks has increasing importance. This article has described the fundamental concepts and solutions. It especially focused on the area of location-based services (LBS) and the possible multicasting over the LBS systems. It was shown that 396
a special kind of multicast, called geocast communication model, utilizes the advantages of the LBS since it is based on the location-aware information being available in the location-based solutions. There are two known issues of this IP-level geocasting. The first problem is the scalability, the flooding type of message transfer compared to multicast tree-based protocols is less robust, but this method is more efficient in a smaller environment than using tree allocation overhead of multicast protocols. The second issue is a source must be connected directly to the router, being physically in the center position, to become source of a session. The proposed geocasting-limited flooding protocol should be extended to handle those situations where the source of a session and the target geological location are in different places.
Conclusion The basically two different technologies, the location-based services in the mobile communication and the well-elaborated multicast technology, are jointed in the multicast over LBS solutions. As it was described, this emerging new area has a lot of possibilities that have not been completely utilized. As a conclusion it can be stated that despite the earlier predicted slower development rate of the LBS solutions, nowadays the technical possibilities and the consumers’ demands have already met. Furthermore, based on the latest development of the multicast over P2P technology, the one-to-many communication can be extended to the LBS systems. Also, in the case of the emerging homeland security applications, the multicast over LBS is not only a possibility, but it became a serious requirement as well. The geospatial property of the LBS provides technical conditions to apply a specialized type of the multicast technology, called geocasting, which gives an efficient and users’ group targeted solution for the one-to-many communication.
References Bocci, L. (2005). Compose project web site. Retrieved from http://www.newapplication.it/compose Chen, Y., Chen, Y. Y., Rao, F. Y., Yu, X. L., Li, Y., & Liu, D. (2004). LORE: An infrastructure to support locationaware services. IBM Journal of Research & Development, 48(5/6), 601-615. Deering, S.E., Estrin, D., Farinacci, D., Jacobson, V., Liu, C-G., & Wei, L. (1996). The PIM architecture for wide-area multicast routing. IEEE/ACM Transactions on Networking, 4(2), 153-162.
Location-Based Services
Garcia-Luna-Aceves, J. J., & Madruga, E. L. (1999, August). The core-assisted mesh protocol. IEEE Journal of Selected Areas in Communications, 1380-1394. Haartsen, J. (1998). The universal radio interface for ad hoc, wireless connectivity. Ericsson Review, 3. Retrieved 2004 from http://www.ericsson.com/review Hightower, J., & Borriello, G. (2001). Location systems for ubiquitous computing. IEEE Computer, (August), 57-65. Hofmann-Wellenhof, B., Lichtenegger, H., & Collins, J. (1997). Global positioning system: Theory and practice (4th ed.). Vienna/New York: Springer-Verlag. Hosszú, G. (2005). Current multicast technology. In M. Khosrow-Pour (Ed.), Encyclopedia of information science and information technology (pp. 660-667). Hershey, PA: Idea Group Reference. Ibach, P., Tamm, G., & Horbank, M. (2005). Dynamic value webs in mobile environments using adaptive location-based services. Proceedings of the 38th Hawaii International Conference on System Sciences. Ibach, P., & Horbank, M. (2004, May 13-14). Highly-available location-based services in mobile environments. Proceedings of the International Service Availability Symposium, Munich, Germany. Ko, Y-B., & Vaidya, N. H. (2002). Flooding-based geocasting protocols for mobile ad hoc networks. Proceedings of the Mobile Networks and Applications, 7(6), 471-480. LocatioNet. (2006). LocatioNet and Ericsson enter into global distribution agreement. Retrieved from http://www. locationet.com Mohapatra, P., Gui, C., & Li, J. (2004). Group communications in mobile ad hoc networks. Computer, 37(2), 52-59. Moy, J. (1994, March). Multicast extensions to OSPF. Network Working Group RFC 1584. Niedzwiadek, H. (2002). Location-based services for homeland security. Retrieved March 2006 from http://www. jlocationservices.com/LBSArticles Uppuluri, P., Jabisetti, N., Joshi, U., & Lee, Y. (2005, July 11-15). P2P grid: Service oriented framework for distributed resource management. Proceedings of the IEEE International Conference on Web Services, Orlando, FL.
KEY Terms Ad-Hoc Computer Network: Mobile devices that require base stations can create the ad-hoc computer network if they do not need routing infrastructure. Ad-Hoc Network: A network that does not need any infrastructure. One example is Bluetooth. Application-Level Multicast (ALM): A novel multicast technology that does not require any additional protocol in the network routers, since it uses the traditional unicast IP transmission. Application-Level Network (ALN): The applications, which are running in the hosts, can create a virtual network from their logical connections. This is also called overlay network. The operations of such software entities are not able to understand without knowing their logical relations. In most cases these ALN software entities use the P2P model, not the client/server for the communication. Client/Server Model: A communicating method, where one hardware or software entity (server) has more functionalities than the other entity (the client), whereas the client is responsible to initiate and close the communication session towards the server. Usually the server provides services that the client can request from the server. Its alternative is the P2P model. Geocast: One-to-many communication method among communicating entities, where an entity in the root of the multicast distribution tree sends data to that certain subset of the entities in the multicast dissemination tree, which are in a specific geographical area. Multicast: One-to-many and many-to-many communication method among communicating entities on various networked hosts. Multicast Routing Protocol: In order to forward the multicast packets, the routers have to create multicast routing tables using multicast routing protocols. Peer-to-Peer (P2P): A communication method where each node has the same authority and communication capability. The nodes create a virtual network, overlaid on the Internet. The members organize themselves into a topology for data transmission. Product Management: A function within a corporation dealing with the continuous management and welfare of the products at all stages of the production procedure in order to ensure that the products profitably meet the needs of customers.
397
L
398 Category: M-Business and M-Commerce
M-Advertising Michael Decker University of Karlsruhe, Germany
INTRODUCTION According to our comprehension, mobile advertising (also called “wireless advertising” or “mobile marketing”) is the presentation of advertising information on mobile handheld devices with a wireless data link like cellular phones, personal digital assistants and smartphones; however notebooks/laptops and tablet PCs are not considered as mobile devices in this sense, because they are used like stationary devices at different locations. For example SMS-messages with product offers would be a simple form of m-advertising. In this article we discuss the special features of m-advertising, but also the problems involved. Afterwards we name basic methods of m-advertising and compare their general strengths and weaknesses using a set of criteria.
are billed according to a rough estimate of the number of generated contacts. But there are additional features of m-advertising [see also Barnes (2002)]: •
M-ADVERTISING COMPARED TO TRADITIONAL FORMS OF ADVERTISING AND INTERNET-BASED ADVERTISING Conventional media for advertising are newspapers, advertising pillars, TV and radio commercials. Relative new media for advertising are the Internet and mobile devices. Both have some features in common: •
•
•
•
Individually Addressable: The user can be addressed individually, so a high degree of personalization is possible: it is possible to tailor the content of each advert according to the profile of the consumer (mass customization). Interactive: if end users receive an advert, they can immediately request further information, participate in a sweepstake or forward an advert to friends. The last one is especially interesting in terms of “viral marketing.” Multimedia-Capable: Multimedia elements (e.g., pictures, movies, jingles, tunes, sounds) are important to realize entertaining adverts and to generate brand awareness. Countable: Each impression of an ad can be counted; for most conventional methods of advertising like TV/radio/cinema commercials or adverts in print media this can not be done and thus the advertisers
•
•
•
Context: In the sense of mobile computing, context is “[…] any information that can be used to characterize the situation of an entity” (Dey, 2001). This information helps to support a user during an interaction with an application. For mobile terminals with their limited user interface context-awareness is especially important. The most prominent example of context is the location of a user, because it changes often and there are a lot of useful scenarios of how to exploit that information. The location information for these “location-based services” can be retrieved based on the position of the currently used base station, the runtime difference when using more than two base stations (TDOA: Time Differential of Arrival) or using a GPS-receiver (Zeimpekis, Giaglis, & Lekakos, 2003). Other examples of context information also used for m-advertising are “weather” and “time” (Salo & Tähtinen, 2005). Reachability: People carry their mobile terminal along with them most of the day and rarely lend it away or share it with other people, because it is a personal device. Therefore marketers can reach people almost anywhere and anytime. Convenience: Mobile terminals are much simpler to handle than personal computers because they are preconfigured by the mobile network operator and have no boot-up time, so they are a medium for electronic advertising to reach people who don’t want to use a computer. Penetration Rates: Mobile terminals—especially cellular phones—have very high penetration rates, which exceed those of fixed line telephones and personal computers. Mobile terminals are more popular than PCs because they are more affordable and simpler to handle. The current global number of cellular phones is beyond one billion, there are even countries with penetration rates over 100 % (Netsize, 2005).
At first glance, m-advertising seems to be a direct continuation of Internet-based advertising: instead of a fixed
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
M-Advertising
computer with a wired data link a mobile terminal with wireless data link is used. But most forms of advertising in the Internet are based on the idea of showing additional advertising information on the user interface (banners on Web pages, sponsored links in search engines). Due to the limited display size of mobile terminals these forms of advertising can’t be used for m-advertising, so new methods have to be developed. Mobile terminals are much more personal devices than personal computers, so a higher degree of personalization can be obtained than with Internet advertising, which leads to better response rates than with other forms of direct marketing (Kavassalis et al., 2003).
ChALLENGES As shown in the last section, m-advertising has some unique advantages when compared to other forms of advertising. But one shouldn’t conceal the challenges associated with it: •
•
•
•
•
•
Unsolicited Messages: Unsolicited mass-mailing with commercial intention (“Spam”) as well as malware (viruses, trojan horses, spyware, etc.) are a great worriment in the fixed-line Internet; the portion of spam messages in e-mail communication exceeds the 50% mark by far. Unsolicited messages on mobile terminals are a much bigger problem, because mobile terminals have limited resources to handle them and are personal devices. Limited Usability: Due to the limited dimensions of mobile terminals they have only a small display and no real keyboard. This has to be considered when designing adverts for mobile terminals. One cannot ask the user for extensive data input, for example, about his/her fields of interests or socio-demographic particulars. Limited Resources: Mobile terminals have very limited resources, for example, memory, CPU-power and available bandwidth. These have to be considered when designing adverts for mobile terminals, for example, transmission of adverts with a lot of data volume is not adequate, Privacy Concerns: As already mentioned, mobile terminals are personal devices with personal data stored on them; it is also possible to track the location of the users. Cost of Mobile Data Communication: Mobile data communication is still very expensive in some regions, so no consumer wants to cover the costs caused by the transmission of adverts. Technical Heterogeneity: The underlying network infrastructure and the capabilities of mobile terminals are much more heterogeneous than for ordinary
fixed-line computers: one advert might look great on one type of terminal, but isn’t displayable on another one. It might cause significant costs when the creator of an m-advertising campaign has to consider many different types of mobile terminals. Due to the problems with unsolicited e-mails and telephone calls “permission-based marketing” is a generally accepted principle when designing systems and campaigns for m-advertising (Barwise & Strong, 2002): A consumer will only receive advertising-messages on his mobile device if he explicitly gave his permission. The adverts sent to a user will be chosen according to the profile of interests of the user. But there is one problem with this principle: a consumer has to know about an m-advertising campaign or system to give his explicit agreement, so one has to “advertise for m-advertising.” Because of this m-advertising is very often integrated into bigger campaigns along with traditional media, see Kavassalis et al. (2003) for examples.
DIFFERENT APPROAChES FOR M-ADVERTISING We distinguish different approaches for m-advertising by the underlying mode of wireless communication used: •
•
•
When using broadcast communication, messages are sent to all ready-to-receive terminals in the area covered by the radio waves according to their natural propagation. If the area covered is rather restricted we talk about a local broadcast, which allows realizing a certain degree of location-aware adoption; the opposite case is global broadcast. Examples: cell-broadcast to all terminals in a certain network-cell (local broadcast) or digital television standards like DVB-H or DMB (global broadcast). Mobile ad-hoc networks (MANETs) are wireless networks without a dedicated infrastructure or a central administration (Murthy & Manoj, 2004). Two terminals of such a network exchange messages when the distance between them is short enough, a message can also be routed via several terminals to the recipient (multi-hop). Unicast communication provides a dedicated point-topoint connection between a base station and a mobile terminal in an infrastructure-based network like GSM, WLAN or UMTS.
To compare different advertising approaches based on these modes we apply the following set of criteria: •
Which degree of personalization and location-aware adoption can be achieved? 399
M
M-Advertising
• •
• •
•
Will there be costs for the data transmission which have to be borne by the end-user? How reliable is the transmission of the advert? Can it be guaranteed that the advert is received indeed within a certain time span? The last point might be crucial for campaigns with temporary and “last minute” offers. Is the mode capable of direct user interaction? Unsolicited messages aren’t bearable on mobile devices, so a system for m-advertising should be designed to make the dispatching of unsolicited messages impossible. Is it possible to detect the exact number of contacts generated?
BROADCASTING Using broadcast communication, adverts are addressed undirected to an anonymous crowd of people, so there is no possibility of personalization. If the broadcast is a local one at least a certain degree of location-aware adoption can be achieved, for example adverts about shopping facilities and sights located in one cell of a GSM-network. Receiving broadcast messages is free of charge for the end-users, but if the user hasn’t turned on his device or isn’t in the area covered by the broadcast he may miss a message. If a user wants to respond to an advert or forward it to other people he has to resort to unicast communication. Unsolicited messages can only occur if the user is in ready-to-receive state. Using broadcasting-based advertising on mobile terminals we lose a lot of the specific opportunities of m-advertising, for example, the high personalization and the interactive nature, but there are no costs for data transmission for the end-user and his anonymity is protected. Without further measures it is not possible to find out how many users received an advertisement.
MANETs The idea of MANETs can be applied for the distribution of adverts (Ratsimor, Finin, Joshi, & Yesha, 2003; Straub & Heinemann, 2004): following the idea of “word-of-mouth” recommendations mobile terminals exchange adverts when they approach each other, whereas the initial transmission of adverts is performed by fixed “information sprinklers.” If these are positioned at an appropriate place we can obtain a certain degree of location awareness at least for those users whose devices receive the message from the sprinklers directly. A client application installed on the mobile terminal will only display adverts that match the profile of the user, so a certain degree of personalization is possible. There are also incentive schemes: when an advert leads to a purchase a user whose device participated in the recommendation chain may receive a bonus. 400
The multi-hop case requires a special client application on the device. At present only the single hop-case can be found in practice: “sprinklers” installed at appropriate places (e.g., entrance halls of shopping or conference centers or even at billboards) submit adverts to mobile devices using infrared [e.g., “marketEyeTM” by Accinity (2006)] or bluetooth [e.g., “BlipZones TM” by BLIP Systems (2006)] communication, but the adverts are not transmitted to other devices automatically. M-advertising based on MANETs doesn’t cause costs for the end-user. The reliability of this form of advertising is not very high, because one cannot give guarantees how long it takes until a consumer receives—if at all—an ad. For interaction the user has to resort to Unicast communication. In the multi-hop case the local client applications can decide which adverts to display and which not to, so the user won’t be harassed by unsolicited messages; in the single-hop case the user can disable his infrared or bluetooth interface if he doesn’t want to receive ads. The advertisers can only count contacts that led to certain defined actions (e.g., if a digital voucher is redeemed). Receiving ads from unknown mobile terminals all the time may also be very energy consuming and there is the danger of receiving malware which exploits flaws of the mobile device.
UNICAST Unicast-communication can be further divided into “push” and “pull:” in push-mode the consumer receives a message without a direct request, for example, SMS; in pull-mode the consumer has to perform an explicit request for each message, for example, request of WAP-pages. Since Unicast communication provides a dedicated point-to-point connection, the advertiser can deliver a different ad for each user, so there is a high degree of personalization possible and the advertiser can calculate the exact number of contacts generated. The most obvious form of pull-advertising for mobile terminals is Web-pages in special formats like WML or cHTML. When viewing such pages the user has to pay for the data volume, so there is the idea that the advertiser covers the costs if the user’s profile meets certain criteria (Figge, Schrott, Muntermann, & Rannenberg, 2002). Pull-advertising isn’t vulnerable for unsolicited messages, but the reliability is also restricted, because the user might miss offers he is interested in or obtain offers too late if he doesn’t request the right page at the right time. SMS as form of mobile push-messaging is also the most successful data service for mobile phones, but the messages can only consist of text and are bound to a maximum size of 140 bytes or 160 letters for a 7-bit-encoding—and Barwise and Strong (2002) even recommend to use much shorter messages when using SMS for advertising. But it is the data
M-Advertising
service which is supported by almost all cellular phones, so it is an interesting opportunity for m-advertising. The multimedia messaging service (MMS) is a further development of SMS, which is also capable of displaying multimedia content, but the creation of an appealing message on a mobile device is challenging; also in many countries sending MMS is very expensive. “V-Card” (Mohr, Nösekabel, & Keber, 2003) is a platform for the creation of MMS and offers templates and multimedia content, so the user can design easily a personalized message. To keep the service free of charge for the end-user a sponsor bears the costs and thus has his advertising message included in the MMS. The “MoMa” system (Bulander, Decker, Schiefer, & Kölmel, 2005) is a system for context aware push advertising for mobile terminals with a special focus on privacy aspects. The system acts as mediator between end-users and advertisers: end-users put “orders” into the system using a special client application to express that they are interested in advertising concerning a certain product or service (e.g., restaurants or shoe shops in their current surrounding); the possible products and services are listed in a hierarchical catalogue. Depending on their type the orders can be refined through the specification of attributes. On the other side of the system the advertisers put “offers” into the system. A matching component tries to find fitting pairs of orders and offers, in case of a hit a notification via SMS/MMS, e-mail or text-to-speech call will be dispatched. When appropriate the matching process also considers context-parameters like the location of the user or the weather. Push advertising in general provides a high degree of reliability but is also vulnerable with regard to unsolicited messages.
CONCLUSION We highlighted the special features of mobile and wireless terminals as medium for advertising, but saw also the considerable challenges. M-advertising might be one of the first successful m-business applications in the business-toconsumer sector, because it doesn’t require m-payment. The current market share of m-advertising in relation to the total advertising market seems to be less than one percent in most countries, but is expected to reach a few percent-points (like nowadays Internet-advertising) in the medium term.
REFERENCES Accinity. (n.d.). Retrieved March 28, 2006, from http://www. accinity.com
Barnes, J. (2002). Wireless digital advertising: Nature and implications. International Journal of Advertising, 21(3), 399-420. BLIP Systems. (n.d.). Retrieved March 28, 2006, from http://www.blipsystems.com Bulander, R., Decker, M., Schiefer, G., & Kölmel, B. (2005). Advertising via mobile terminals. In Proceedings of the 2nd International Conference on E-Business and Telecommunication Networks (ICETE ’05). Reading, UK. Dey, A.K. (2001). Understanding and using context. Personal and Ubiquitous Computing Journal, 5(1), 4-7. Figge, S., Schrott, G., Muntermann, J., & Rannenberg, K. (2002). Earning M-Oney—A situation based approach for mobile business models. In Proceedings of the 11th European Conference on Information Systems (ECIS). Naples, Italy. Kavassalis, P., Spyropoulou, N., Drossos, D., Mitrokostas, E., Gikas, G., & Hatzistamatiou, A. (2003). Mobile permission marketing: Framing the market inquiry. International Journal of Electronic Commerce, 8(1), 55-79. Mohr, R., Nösekabel, H., & Keber, T. (2003). V-Card: Sublimated message and lifestyle services for the mobile mass market. In Proceedings of the 5th International Conference on Information and Web-Based Applications and Services. Jakarta, Indonesia. Murthy, C., & Manoj, B. (2004). Ad hoc wireless networks—Architectures and protocols. Upper Saddle River, NJ: Prentice Hall. Netsize. (2005). The Netsize guide 2005. Paris, France. Ratsimor, O., Finin, T., Joshi, A., & Yesha, Y. (2003). eNcentive: A framework for intelligent marketing in mobile peer-to-peer environments. In Proceedings of the 5th ACM International Conference on Electronic Commerce (ICEC 2003). Pittsburgh, Pennsylvania. Salo, J., & Tähtinen, J. (2005). Retailer use of permissionbased mobile advertising. In I. Clarke, III & T.B. Flaherty (Eds.), Advances in Electronic Marketing (pp. 139-155). Hershey, PA: Idea Group Publishing. Straub, T., & Heinemann, A. (2004). An anonymous bonus point system for mobile commerce based on word-of-mouthrecommendation. In Proceedings of the ACM Symposium on Applied Computing (SAC ’04). Nicosia, Cyprus. Zeimpekis, V., Giaglis, G., & Lekakos, G. (2003). A taxonomy of indoor and outdoor positioning techniques for mobile location services. ACM SIGecom Exchanges, 3(4), 19-27.
Barwise, P., & Strong, C. (2002). Permission-based mobile advertising. Journal of Interactive Marketing, 16(1), 14-24. 401
M
M-Advertising
KEY TERMS Broadcast: sending data using wireless communication to an anonymous audience according to the natural propagation of radio waves. Context: information available at runtime of a computer system to support the user when interacting with the system. Location-Based Services: most prominent case of context-aware services, which adapt themselves according to the current location of the user. Mobile Advertising (M-Advertising): adverts displayed on mobile and wireless terminals like cellular phones, smartphones or PDAs.
402
Mobile Ad-Hoc Network (MANET): Wireless network without infrastructure and central administration, the nodes (devices) pass messages to other nodes in reach. Pull: A user receives a message only as direct response. Push: A user receives a message without directly requesting it. Unicast: Communication using an infrastructure-based network which provides dedicated point-to-point-connection. Examples are GSM/GPRS, WLAN or UMTS. Viral Marketing: Dissemination of adverts by consumers themselves (“tell a friend!”); is expected to generate exponential rates of contacts and a higher trustability than adverts received from firms directly.
Category: Converging Technology 403
Man-Machine Interface with Applications in Mobile Robotic Systems Milan Kvasnica Tomas Bata University, Zlin, Czech Republic
INTRODUCTION This article focuses on the current state-of-the-art assistive technologies in man-machine interface and its applications in robotics. This work presents the assistive technologies developed specifically for disabled people. The presented devices are as follows: •
• •
•
•
The head joystick works on a set of instructions derived from intended head movements. Five laser diodes are attached to the head at specific points whose light rays’ spots are scanned by a set of CCD cameras mounted at strategic locations (on the ceiling, on the wall, or on a wheelchair). Automatic parking equipment has two laser diodes attached at the back of the wheelchair, and their light rays’ spots are scanned by the CCD cameras. A range-inclination tracer for positioning and control of a wheelchair works on two laser diodes attached onto the front of the wheelchair. A CCD camera mounted on the front of the wheelchair detects the light rays’ spots on the wall. The body motion control system is based on a set of instructions derived from intended body motion detected by a six component force-torque transducer, which is inserted between the saddle and the chassis of the wheelchair. An optoelectronic handy navigator for blind people consists of four laser diodes, the 1-D CCD array (alternatively PSD array), a microprocessor, and a tuned pitch and timbre sound source. The functionality of this system is based on the shape analysis of the structured lighting. The structured lighting provides a cutting plane intersection of an object, and a simple expert system can be devised to help blind people in classification and articulation of 3-D objects. There are two parameters involved: the distance and the inclination of the object’s articulation. The time-profiles of the distance and inclination are used to adjust the frequency and amplitude of the sound generator. The sound representation of a 3-D object’s articulation enables the skill-based training of a user in recognizing the distance and ambient articulation.
The head joystick, the automatic parking equipment, the range-inclination tracer, and the body motion control system for the wheelchair control are suitable for people who have lost the ability to use their own lower limbs to walk or their upper limbs for quadriplegics. The optoelectronic handy navigator is suitable for blind people. The mentioned sensory systems help them to perform daily living tasks, namely to manage independent mobility of electrical wheelchair or to control a robot manipulator to handle utensils and other objects. The customization of described universal portable modules and their combinations enable convenient implementation in rooms and along corridors, for the comfort of the wheelchair user. Smart configuration of the optoelectronic handy navigator for blind people enables the built-in customization into a handy phone, handheld device, or a white stick for blind people.
BACKGROUND Significant progress in human-computer interfaces for elderly and disabled people has been reported in recent years. Some examples for such devices are the eye-mouse tracking system, hand gesture systems, face gesture systems, head controller, head joystick, and human-robot shoulder interface, all presented at recent international conferences. The aim of this article is to publish further progress in the field of assistive technologies like the head joystick, automatic parking equipment, range-inclination tracer, body motion control system, and optoelectronic handy navigator for blind people.
hEAD JOYSTICK AND AUTOMATIC PARKING EqUIPMENT Following parts of the modular sensory system enables the processing of multi-DOF information for the control and the positioning of a wheelchair by means of two types of modules for alternative use, as shown in Figure 1. The module Apc (the module of four laser diodes) is designed for tracking the head motion of the wheelchair user. The ceiling-mounted CCD cameras detect the Apc laser rays. The fifth, auxiliary laser diode with redundant light
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
M
Man-Machine Interface with Applications in Mobile Robotic Systems
Figure 1. The modules Apc and App positioned relative to the plain of the ceiling and the module Bp of the CCD camera mounted in perpendicular view or in perspective view against the light spots on the ceiling
spot is used for the verification of accurate functionality. The module App (the module of two laser diodes) is designed for automatic parking of the wheelchair into a predefined position in the room. The third auxiliary laser diode with redundant light spot is used for the verification of accurate functionality. The modules Apc and App have the presetting control 1 of the angle 2s contained by mutual opposite light rays 2, as shown in Figures 1 and 2. The auxiliary fifth or third laser diode is centered in the axis. The camera with 2-D CCD array can be arranged in two ways: •
•
404
The perpendicular view downwards against the translucent screen, which is mounted parallel to the ceiling. The module Cp for direct sampling is shown in Figure 3. The light spots reflected by light rays of Apc, or App respectively, are sampled by the camera with a 2-D CCD array. The module Cp for direct sampling of the light spots (no. 3) consists of the camera with 2-D CCD array (no. 6) with focusing optics (no. 5) and the flange (no. 1) mounted perpendicular to the translucent screen. The light spots from the laser light rays are projected onto the translucent screen (no. 4). The translucent screen spans the entire ceiling of the room. In larger rooms, four Cp modules are attached onto the ceiling in front of the laser rays. The perspective view of the ceiling and light spots of the laser ray images spots from the modules Apc and App are shown in Figure 1. The module Bp, depicted in Figure 4, is shown in Figure 1 in perspective view on the wall. The camera with 2-D array makes the sampling of the light spot position from the laser light
rays on the ceiling plane. The X-Y coordinate system on the ceiling is used to monitor the parking position of the wheelchair. The module Cp is mounted on the ceiling against the modules Apc and App, respectively. The Apc module is attached to the head of the wheelchair user, and the rays are intersecting the ceiling plain of the Bp or Cp modules respectively, in light spots A, B, C, and D. The intersection of abscises AB and CD is the point S centered by auxiliary laser diode. The lengths of abscises AS, SB, CS, and SD are u1, u2, u3, and u4. The App module is attached to the wheelchair, and the light rays intersect the ceiling plain in light spots E, F in equal distance u from the middle point R centered by auxiliary laser diode. The light spots position and configuration is analyzed and processed for the navigation of a wheelchair. Purposeful head motion of the module Apc represented by the light spots configuration is sampled by means of the modules Bp, or Cp from the ceiling. The following commands are used for three degrees-of-freedom control of the wheelchair with an adjustable operating height of the wheelchair perpendicular to the 2-D coordinate frame on the ground: • • •
• • • • • •
Start: The head movement with the module Apc outwards the dead zone position of the light spots. Stop: The head movement with the module Apc inwards the dead zone position of the light spots. The dead zone is defined by the ratio u1/u2 and u3/u4, for example (only for perpendicular view of the CCD Camera) by the interval and for the angle Ω by the interval . Inside these intervals, following commands are not valid because of physiological trembling of the head. Outwards these intervals are valid following commands. Forward, Backward for the First DOF: The dividing ratio of diagonals u1/u2 > 1,2 respectively u1/u2 < 0,8. Up, Down for the Second DOF: The dividing ratio of diagonals u3/u4 > 1,2 respectively u3/u4 < 0,8. Turn to the Left – Turn to the Right for the Third DOF: Last increment of the angle Ω is positive Ω > +150, respectively, and negative Ω < –150 oriented. The magnitude of the dividing ratios u1/u2, u3/u4, and the angle Ω is assigned to the velocity of the wheelchair motion for each DOF. The motion control system of the wheelchair enables parallel control of all three degrees-of-freedom. The light spot A from the module Apc is recognized by means of the enhanced intensity, color, or shape in contrast with light spots B, C, and D. This is needed for the orientation of the wheelchair against the basic light spot position.
Man-Machine Interface with Applications in Mobile Robotic Systems
Figure 2. The multi-laser configuration modules App, and Apc
Figure 3. The Cp module for direct sampling
12 1 Ap
3 2
s s
Figure 4. The module Bp for the ground plane or for perspective ceiling sampling 3
Bp
2
y 7
•
•
•
5 4
Cp
Figure 5. Navigation by automatic parking system
1 4
x
z
The automatic parking equipment of a wheelchair is devised on the module App, which is attached at the back of the wheelchair:
•
3
5
6
•
6
M
The module App is used for automatic return into the parking position using controls in forward-backward, up-down, and turns right-left. The module App and the module Apc operate independently and sequentially into common modules Bp or Cp. The position E(x1,y1) and F(x2,y2) of the light spots from the module App on the ceiling are sampled by the camera. The information about the light spots position E(x1,y1) and F(x2,y2) is sufficient for the navigation into the points E0(x01,y01) and F0(x02,y02) of the basic position highlighted from the reflexive material on the ceiling. The light spot E from the module App is recognized by means of the intensity, color, or shape in contrast with light spot F. This is needed for the orientation of the wheelchair against the basic light spot parking position. Another technique of recognizing the orientation is to sample every light spot in separate picture synchronized with sequentially switched light rays.
The coordinates of the position x,y of the wheelchair at point R of the ceiling rectangular coordinate frame is given by the relationship (1), and the normal vector n in the
middle point R determines the direction of the wheelchair trajectory.
x= • • •
1 1 ( x1 + x2 ); y = ( y1 + y2 ); 2 2
(1)
The direction n of the wheelchair motion in the ceiling rectangular coordinate frame is given by the relationship (3) according to Figure 5. All coordinates xi, yi, z are with respect to the geometrical center of the light spots. The z coordinate is used for calibration of the magnitude xi, yi according to the position of the light spot image center of the 2-D CCD coordinate frame.
The distance z of the wheelchair from the ceiling in the rectangular coordinate frame (recommended the value s = arctg(1/2)) according to Figures 2 and 5 is given by the relationship (2).
z = ( y2 − y1 ) 2 + ( x2 − x1 ) 2
;
(2)
−1
y −y n = − 2 1 x2 − x1 .
(3) 405
Man-Machine Interface with Applications in Mobile Robotic Systems
The coordinates of the wheelchair position x,y and the normal vector direction n of the wheelchair motion are used for the planning of the wheelchair trajectory as well as for collision avoidance.
Figure 7. Geometrical approach of the activity of the rangeinclination tracer
ThE RANGE-INCLINATION TRACER The range-inclination tracer consists of the camera with 1-D CCD array or linear PSD element and two laser diodes radiating two intersecting light rays against the wall, as shown in Figures 6 and 7. Two light spot positions of the light rays are sampled by a simple sampling algorithm, in order to evaluate the distance D and the inclination β of the wall against the wheelchair. The sampling of every light spot coordinate in separate picture synchronized with sequentially switched light rays enables recognition of the varying orientation of the light spots before and after the crossing point of the laser light rays. Control algorithms are derived in the rectangular coordinate system x,y with an origin in zero point of mutual intersection of light rays (no. 4), so that the y coordinate fuses with optical axis (no. 5) of the camera (no. 3) with 1-D CCD array, and its positive orientation leads into the camera (see Figure 7). The coordinates of the light spot on the wall (no. 6) emitted by the light source (no. 1) are designated by A(r1,s1), and the coordinates of the light spot emitted by the light source (no.,2) are designated by B(r2,s2). Every light ray creates the angle α/2 with optical axis of the camera. The magnitude of the coordinates r1,r2 is derived according to the calibration of the light spot’s position in the image coordinate frame of the 1-D CCD array. Coordinates s1,s2 are computed from the relationship (4).
a/2 a/2 ∆h
wall
s1 =
r1 tg
(r2,s2)
(r1,s1)
A
B β D
;
2
s2 =
r2 tg
2
(4)
where h is the distance between the camera and the point of the intersection of light rays. The Δh is the difference between the point of the light rays’ intersection and the intersection of the optical axis of the camera with the wall, given by the relationship (5). The distance D between the camera and the wall is given by the relationship (6), and the inclination β of the wall against the wheelchair is given by the relationship (7).
D = h ± ∆h ∆h = −
Figure 6. Range-inclination tracer attached to the wheelchair
β
(5)
s1 − s2 r1 + s1 r2 − r1
(6)
r2 − r1 s1 − s2
(7)
= arcctg
The described sensory system is alternatively used for independent control of the wheelchair trajectory in the vicinity of the walls, namely in long corridors, and enables the implementation of a low-cost collision avoidance system.
hUMAN-ROBOT INTERFACE USING ThE BODY MOTION Many people with limited mobility prefer to enjoy life. They particularly enjoy a wheelchair ride at a popular tourist spot with typical leisure activities, and sometimes go on
406
Man-Machine Interface with Applications in Mobile Robotic Systems
high adventures and wilderness expeditions. These expeditions are usually long trips, and active compliant assistance by way of feedback control based on the damping device with a shock absorber is required. This active compliant assistance is able to overcome some barriers and to predict hazardous scenarios in unexpected situations like the quick halt on the stone or the raid on a sharp slope. A basic part of compliant assistance is the six-component force-torque transducer inserted between the saddle and the chassis as shown in Figures 8 and 9. The force-torque transducer used for dynamic weighing of the user and the load on the wheelchair improves the dynamic stability at maneuvering and the comfort of the ride. In addition this force-torque transducer is possible to use for the sampling of the information about the user’s body motion. The control system based on the user’s body motion is based on a set of instructions derived from the body inclination and sampled by means of the six-component force-torque transducer inserted between the saddle and the chassis of the wheelchair. The six-component force-torque transducer is also used for the active compliant assistance. The explanation of the activity is introduced on a simple modification of the six-component force-torque transducer. An example of the six-component force-torque transducer with the acting force -Fz is depicted in Figure 8. Laser diodes (no. 1) emit intersecting light rays (no. 2) creating the edges of a pyramid, intersecting the 2-D CCD array (no. 4) in light spots (no. 3). The beginning of the 3D rectangular pyramid coordinate frame x,y,z is chosen in the crossing point of the light rays (the apex of a pyramid shape). Unique light spots configuration on the 2-D CCD array changes under the force-torque acting between the flanges (no. 5 and no. 6) connected by means of elastic deformable medium (no. 7). The beginning of a floating 2D coordinates frame xCCD,yCCD is chosen in the geometrical center of the 2-D CCD array. An even number of four light rays simplifies and enhances the accuracy of the algorithm for the evaluation of three axial shiftings and three angular displacements. In addition it removes the ambiguity of the imagination at the rotation of the 2-D CCD array around the straight line passing through Figure 8. Six-component force-torque transducer
z
1
D 6
AF
3
- Fz
2
8
9
5
H
x x CCD
7
C
two light spots, where for two inclines of opposite orientation belongs one position of the third light spot. The module of five laser diodes is attached on the outer flange of the force-torque sensor, and the rays intersect the translucent screen in light spots A, B, C, and D. The intersection of the abscises AB and CD is the point S. The length of the abscises AS, SB, CS, and SD is u1, u2, u3, and u4. The light spot configuration is analyzed and processed for the navigation and positioning of the wheelchair. Intentional body inclinations against the saddle are represented by the light spots (no. 3) configuration in the 2-D CCD array (no. 4) of the force-torque transducer, as shown in Figure 8, which is analyzed like the force-torque acting in six components. The following commands are used for two-degrees-of-freedom control of the wheelchair as shown in Figure 9: • • • • • • • •
Start: The body inclination in chosen direction outwards the light spots dead zone position. Stop: The body inclination inwards the light spots dead zone position. The dead zone is defined for the dividing ratio u1/u2 for example by the interval , u3/u4 is symmetric . Forwards, Backwards for the First DOF: The dividing ratio of diagonals u1/u2 > 1.2 respectively u1/u2 < 0.8. Turn to the Left—Turn to the Right: Similar for the u3/u4. Final direction of the wheelchair motion is the vector sum of components u1/u2 and u3/u4. The magnitude of the dividing ratio u1/u2, u3/u4 is assigned to the velocity of the wheelchair motion for two degrees of freedom (DOF). The motion control system of the wheelchair enables parallel control of two degrees-of-freedom. The light spot A is oriented in front direction. This is needed for the orientation of the wheelchair against the basic light spot position.
Some applications of the signal filtering for the elimination of the body inclination are needed for users with muscular trembling at neurological diseases and at the motion of the wheelchair on rough surface. Figure 9 depicts the CCD camera for the sampling of the head joystick’s light spots configuration on the floor. The difference between the sampling from the CCD camera attached to the wall or to the ceiling against the one attached to the wheelchair is in the level of navigation. The CCD camera attached to the wall or to the ceiling enables the positioning of the wheelchair with respect to the shape of the room. The CCD camera attached to the wheelchair enables only the relative positioning of the wheelchair.
4
407
M
Man-Machine Interface with Applications in Mobile Robotic Systems
Figure 9. The force-torque transducer inserted between the saddle and the chassis of the wheelchair and the CCD camera for the sampling of the head joystick’s light spots configuration on the floor
and timbre of sound source. Two parameters on the intersection are computed and evaluated: the distance and the inclination of the object. The time-profiles of the distance and inclination, and their combinations are used to tune the sound generator’s frequency and intensity. The sound representation of a 3-D object’s articulation enables the skill-based training of the user in recognizing the distance and ambient articulation. Smart configuration enables the customizing of navigation device into: • •
OPTOELECTRONIC hANDY NAVIGATOR FOR BLIND PEOPLE Safe and effective mobility of blind people depends largely upon reliable orientation about the articulation of the ambient. The current situation of people who suffer severe visual impairments is that they mostly require active assistance from relatives or occasionally assistance of passing by strangers in order to travel or to use transportation or to manage street traffic in crowded areas. Nevertheless, blind people, just like sighted people, do not like to be dependent on others. There are restricted sources of sound information at the traffic light control or acceptable spoken information at landmarks and significant places in information stations. Independent travel and transportation for blind people involves orienting oneself and finding a safe path through known and unknown articulated environments. Most efforts have been to solve the mobility part of the problem to help the blind traveler detect irregularities on the floor such as boundaries, objects located near or alongside his or her path in order to avoid collisions, and steer a straight and safe course through the immediate environment. Ultrasonic and laser devices for the navigation using active methods for the identification of the environment have been reported. One way for the classification of 3-D object articulation destined for blind people is based on the principle of the twofold range-inclination tracer. The range-inclination tracer enables the shape analysis by means of the structured light-cutting plane intersection with an object. This navigation system consists of four laser diodes with focusing optics used for the light spots imagination 1-D CCD array (alternatively PSD array), microprocessor, and tuned pitch 408
a handy phone that uses only one line from the built-in 2-D CCD array, or a white stick for blind people using 1-D CCD array (alternatively PSD array) or the handheld device.
The twofold range-inclination tracer consists of the camera with 1-D CCD array or linear PSD element and four laser diodes radiating four intersecting light rays against the wall, as shown in Figure 10. Four light spots are sampled by a simple sampling algorithm, which then evaluates the distance D and the inclination β of the plane against the range-inclination tracer. The sampling of left and right light spot coordinates in subsequent pictures synchronized with sequentially switched light rays enables the recognition of the changing difference and orientation of the light spots and their orientation before and after the crossing point. Control algorithms are derived in the rectangular coordinate system x,y with the origin in the point of mutual intersection of four light rays, as described in Figure 7 using the relationships (4), (5), (6), and (7). The described range-inclination tracer is alternatively used for navigation: • •
in single mode, using two crossing light rays with light spots A(r1,s1) and B(r2,s2), with the evaluation of the distances and inclination of the ambient; or in double mode, using four crossing light rays with light spots A(r1,s1), B(r2,s2), E(r3,s3), and F(r4,s4), with the evaluation of the distances and inclination of the ambient including the evaluation of the invariant symptoms of the ambient articulation, like simple plain, parallelism, or perpendicularity of two plains.
The navigation system for blind people is based on the cooperation of four laser diodes with focusing optics used to determine light spot in the 1-D CCD array (alternatively PSD array), microprocessor, and tuned pitch and timbre of sound source. The first couple of laser rays with light spots A,B form an angle α, and the second couple of laser rays with light spots E,F form an angle σ; both are symmetric with the axis of the CCD array. Generally the navigation system consists of two independently working range inclination tracers. The configuration of the navigation system enables, by means of purposeful swept hand motion, simple evalu-
Man-Machine Interface with Applications in Mobile Robotic Systems
Figure 10. Navigation system based on the twofold rangeinclination tracer fastened on the handy device
Figure 12. Indication of parallel plains
ation of invariant symptoms of the object for the indication of the following:
Figure 13. Indication of two plains forming an angle with pointing up of the right angle
• • •
M
simple plane, when the points A,B,E,F belong to one straight line as shown in Figure 11; parallel planes, when the points A,E and B,F create two parallel straight lines as shown in Figure 12; or two plains forming an angle with pointing up of the right angle, when the points A,E and B,F create two mutually intersecting lines, as shown in Figure 13. The navigation device can be used in two ways:
1.
2.
The single mode using only two light spots A,B or E,F on the way of the range-inclination tracer. This way is effective to assign a safe and acceptable range of the distance and inclination using two-tones. For example, when both light spots are on the pavement, a unique tone is played. The loss of safe inclination is signaled by the dissonance, for example at the positions of the light spots on the pavement-wall. The loss of the front light spot is signaled by the interrupted tone of enhanced intensity. The navigation system using two pairs of the light spots A,B and E,F enables, by means of purposeful swept hand motion, a simple indication of the object
Figure 11. Indication of a simple plain
articulation. This method is effective to assign various chord combinations (melodies) for different shapes such as simple plane, parallel planes, or two plains forming an angle. Simultaneously the range inclination resonates in the background. Smart configuration of the handy navigator needs simple built-in customizing into a handy phone, handheld device, or white stick for blind people. Video signal from the 1-D or the 2-D CCD array is processed by a simple method at low cost, and in the form of an embedded system, with a single cheap microprocessor in order to decrease the size of optoelectronic devices. The video signal output is preprocessed in the comparator, which is set up on the white level. It causes the signal from every pixel of the light spot to be indicated like an impulse. Every impulse is assigned to the video dot information sequence in the range between 1 and the maximal number of pixels in the CCD array in order to determine its position in the picture coordinate frame. The run of the dot video information is switched for every picture separately by vertical (picture) synchronization impulse. Every couple of light spots A,B and E,F is evaluated in separate pictures. It enables recognition of the varying orientation of the light 409
Man-Machine Interface with Applications in Mobile Robotic Systems
spot before and after the crossing point of the light rays. An alternative application of the 1-D PSD array is based similarly on the sampling of only one light spot’s position.
EXPERIMENTAL hARDwARE FOR COMPOSING OF SIMPLE TASKS The automatic parking equipment, the range-inclination tracer, the optoelectronic handy navigator, and the wheelchair control by means of the head motion and by means of body motion were implemented by means of following experimental hardware: •
•
• •
•
The Data Translation High-Accuracy, Programmable, Monochrome Frame Grabber Board DT3155 for the PCI Bus: This is suitable for both image analysis and machine vision applications. The Microprocessor-Controlled Programmable Timer PIKRON ZO-CPU2: This used for the timing of the light exposure and asynchronous switching of the laser diodes configuration. The Configuration of Miniature Laser Diodes FLASER 5mW: This includes focusing optics radiating structured light rays. Digital B/W Video Camera SONY KC-381CG: This includes a digital signal processor and high-resolution 795Hx596V 1/3” CCD sensor with high sensitivity (0.02lux at F0.75) and interline transfer, digital light level control system for the back-light compensation, with auto-exposure or manual exposure system, aperture correction, and internal or gen-lock/line-lock external transfer. Zoom Lens Computar MLH-10X: This includes 10x macro zoom, maximal magnification 0.084~0.84x, maximal aperture 1:5.6, maximal image format 6.4×4.8mm (average 8mm), and focus 0.18~0.45m.
The image processing of the light spots from the CCD camera including the control algorithms for the first step of skill-based education was developed on MATLAB. The application of the frame grabber is used only for experiments. The second step of the education is oriented on the processing of the video signal from the 1-D or the 2-D CCD array for the user’s application by a single cheap microprocessor. The activity of the PSD element is based on the sampling of only one light spot’s position. This means that for every light beam, it assigns a separate PSD element. In this way the subsequent sampling of the six-DOF information by means of the PSD element is guaranteed, but causes dynamic distortion dependent on the sampling frequency. Correct activity of the six-DOF sensor based on the CCD
410
element depends on the continuity of sampling. This indicates the verification of every light spot position in regard to responding basic position A,B,C,D,S. The continuous motion of the first light spot can be recognized in various ways, for example by means of: • • • • •
different shape of the light spot, different color of the light spot, different intensity of the light spot, dividing of the picture frame into four quadrants for every light spot, and minimal distance of preliminary position of the light spot.
Having the identity of the first light spot, the other light spots can be recognized clockwise or counter-clockwise. The third step of the educational applications is oriented toward embedded systems in order to decrease the size of developed optoelectronic devices. The structure of the embedded system is oriented to the use of microprocessors in order to enable flexibility in the proof of various algorithmic modules for the enhancement of the accuracy, like the approximation of the light spot center or for the elimination of the nonlinearity, and on the tasks of the dynamic control using C language with subroutines in assembler.
CONCLUSION The modular sensory system design presented here enables easy adaptation of various applications of human-machine interfaces for assistive technologies and mobile robotic systems. In general, this modular sensory system concept is appropriate for a low-cost design and in addition enables the understanding of basic problems concerning the interaction between human and mobile robotic systems.
ACKNOwLEDGMENTS The support from the grant Vyzkumne zamery MSM 7088352102 “Modelovani a rizeni zpracovatelskych procesu prirodnich a syntetickych polymeru” is gratefully acknowledged.
REFERENCES Akira, I. (2002). An approach to the eye contact through the outstaring game Nirammekko. Proceedings of the 11th IEEE International Workshop on Robot and Human Interactive Communication (RO-MAN 2002), Berlin, Germany.
Man-Machine Interface with Applications in Mobile Robotic Systems
Fukuda, T., Nakashima, M., Arai, F., & Hasegawa, Y. (2002). Generalized facial expression of character face based on deformation model for human-robot communication. Proceedings of the 11th IEEE International Workshop on Robot and Human Interactive Communication (RO-MAN 2002), Berlin. Humusoft. (n.d.). Retrieved from http://www.humusoft.cz Kawarazaki, N., Hoya, I., Nishihara, K., & Yoshidome, T. (2003). Welfare robot system using hand gesture instructions. Proceedings of the 8th International Conference on Rehabilitation Robotics (ICORR 2003), Daejeon, Korea. Kim, D.-H., Kim, J.-H., & Chung, M. J. (2001). An eye-gaze tracking system for people with motor disabilities. Proceedings of the 7th International Conference on Rehabilitation Robotics (ICORR 2001), Evry Cedex, France. Kim, J.-H., Lee, B. R., Kim, D.-H., & Chung, M. J. (2003). Eye-mouse system for people with motor disabilities. Proceedings of the 8th International Conference on Rehabilitation Robotics (ICORR 2003), Daejeon, Korea. Kvasnica, M. (1999, July). Modular force-torque transducers for rehabilitation robotics. Proceedings of the IEEE International Conference on Rehabilitation Robotics, Stanford, CA.
Kvasnica, M. (2003). Six-DOF sensory system for interactive positioning and motion control in rehabilitation robotics. International Journal of Human-Friendly Welfare Robotic Systems, 4(3). Kvasnica, M. (2004, September). Six-DOF force-torque transducer for wheelchair control by means of body motion. Proceedings of the 2nd International Conference on Smart Homes and Health Telematics (ICOST 2004), Singapore. Kvasnica, M. (2005). Assistive technologies for man-machine interface and applications in education and robotics. International Journal of Human-Friendly Welfare Robotic Systems, 6(3). Daejeon, Korea: KAIST Press. Kvasnica, M., & Vasek, V. (2004, May). Mechatronics on the human-robot interface for assistive technologies and for the six-DOF measurements systems. Proceedings of the 7th International Symposium on Topical Questions of Teaching Mechatronics, Rackova Dolina, Slovakia. Kvasnica, M., & Van der Loos, M. (2000). Six-DOF modular sensory system with haptic interaction for robotics and human-machine interaction. Proceedings of the World Automation Congress (WAC 2000), Maui, HI. Mathworks. (n.d.) Retrieved form http://www.mathworks. com/
Kvasnica, M. (2001, April). A six-DOF modular sensory system with haptic interface for rehabilitation robotics. Proceedings of the 7th International Conference on Rehabilitation Robotics (ICORR 2001), Paris-Evry, France.
Min, J. W., Lee, K., Lim, S.-C., & Kwon, D.-S. (2003). Human-robot interfaces for wheelchair control with body motion. Proceedings of the 8th International Conference on Rehabilitation Robotics (ICORR 2003), Daejeon, Korea.
Kvasnica, M. (2001). Algorithm for computing of information about six-DOF motion in 3-D space sampled by 2-D CCD array. Proceedings of the 7th World Multi-Conference (SCI’2001-ISAS) (Vol. XV, Industrial Systems, Part II), Orlando, FL.
Moon, I., Lee, M., Ryu, J., Kim, K., & Mun, M. (2003). Intelligent robotic wheelchair with human-friendly interfaces for disabled and the elderly. Proceedings of the 8th International Conference on Rehabilitation Robotics (ICORR 2003), Daejeon, Korea.
Kvasnica, M. (2002). Six DOF measurements in robotics, engineering constructions and space control. Proceedings of the 8th World Multi-Conference on Systemics, Cybernetics and Informatics (SCI’2002-ISAS 2002) (Ext. Vol. XX), Orlando, FL.
Neovision. (n.d.). Retrieved from http://www.neovision.cz
Kvasnica, M. (2003, September). Head joystick and interactive positioning for the wheelchair. Proceedings of the 1st International Conference on Smart Homes and Health Telematics (ICOST 2003), Paris. Kvasnica, M. (2003). Six-DOF sensory system for interactive positioning and motion control in rehabilitation robotics. Proceedings of the 8th International Conference on Rehabilitation Robotics (ICORR 2003), Daejeon, Korea.
KEY TERMS α/2: The angle of the optical axis of the camera with the laser light ray. Β: The inclination of the wall against the wheelchair. D: The distance between the camera and the wall. Δh: The difference between the point of the light rays’ intersection and the intersection of the optical axis of the camera with the wall.
411
M
Man-Machine Interface with Applications in Mobile Robotic Systems
h: The distance between the camera and the point of the intersection of light rays. N: The direction of the wheelchair motion. (r1,s1), (r2,s2): The coordinates of the light spots A, B. 2s: Presetting control of the angle contained by mutual opposite light rays of the modules Apc and App. x,y: The coordinates of the position of the wheelchair. (x1,y1), (x2,y2): Coordinates of the points E, F. z: The distance of the wheelchair from the ceiling in the rectangular coordinate frame.
412
Category: Human Factor 413
M-Commerce Technology Perceptions on Technology Adoptions Reychav Iris Bar-Ilan University, Israel Ehud Menipaz Ben-Gurion University, Israel
INTRODUCTION This article presents a tool for assessing the probability of adopting a new technology or product before it is marketed. Specifically, the research offers managers in firms dealing with mobile electronic commerce a way of measuring perceptions of technology usage as an index for assessing the tendency to adopt a given technology. The article is based on an ongoing study dealing with m-commerce in Israel and internationally. It is centered on creating a research tool for predicting the usage of m-commerce in Israel, based on the PCI model. The suggested model is based on a questionnaire presented to the potential consumer, containing questions linking the consumer’s perception of the various aspects of the technological innovation offered, together with his tendency to buy and therefore adopt it. The tool was found to possess high reliability and validity levels. The average score in the questionnaire is used to predict the probability of adoption of the mobile electronic commerce technology. Implications related to m-commerce technology in Israel and worldwide are discussed.
BACKGROUND The main purpose of this study is to assess the tendency to adopt mobile electronic commerce technologies, prior to actually launching a new product or service based on cellular technology. The focus in the present study is on the general population and not on organizations. Contrary to products or services sold to end users, the mobile electronic commerce field offers an innovative system of business ties with the client, by utilizing mediating tools such as the cellular device. The focus in the mobile electronic commerce field is on influencing consumers’ preferences. This study presents a tool that examines the perceived utilization of technology advances in the field. The focus in this study on the characteristics of using cellular phones for mobile electronic commerce is based on findings from a wide range of studies dealing with the characteristics of the perception of innovativeness itself.
Rogers (1983) studied thousands of cases of diffusion and managed to define five characteristics of innovativeness affecting its diffusion: relative advantage, compatibility, complexity, visibility, and trialability. While Rogers’ characteristics were based on the perception of innovativeness itself, Ajzen and Fishbein (1980, p. 8) claimed that the attitude toward the object is different in essence from the attitude towards a certain behavior related to the object. Innovation penetrates because of accumulating decisions by individuals to adopt it. Therefore, not perceiving the efforts of innovation itself, but the perception of using innovation is the key to its diffusion. In the diffusion studies, the subject of perceptions was treated in relation to innovation itself. Nevertheless, the characteristics of the perceptions of innovation can be redesigned in terms of perceiving the use of the innovation (Moore, 1987). Rewriting the characteristics of perceptions of innovativeness into characteristics of the perceptions of using the innovation was the basis for the PCI (perceived characteristics of innovating) model, developed by Moore and Benbasat (1991) and used as a tool for studying the adoption of information technologies. The PCI model expands the conceptual framework designed by Rogers, by adding additional characteristics that may influence the decision to adopt a new technology. The tool was presented as reliable and valid.
MOTIVATION AND PURPOSE OF ThE STUDY The motivation for creating a tool for measuring the perceptions regarding cellular phone usage for mobile electronic commerce of the potential adopters of the technology originated from three main factors. First were findings from previous research, which focused on adoption patterns of Internet and cellular technologies in various countries, in an attempt to present a methodology for analyzing diffusion (Reychav & Menipaz, 2002). Second, while carrying out the above-mentioned study, the researchers realized there was a lack in theoretical background for studying the initial adoption process of innovative technologies
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
M
M-Commerce Technology Perceptions on Technology Adoptions
such as m-commerce, as well as the understanding of how to successfully assimilate innovative technology. Third, an opportunity presented itself to examine the research model in Israel, a country in which the usage of cellular technology is widespread.
METhOD The study took place in Israel, where the knowledge constraint towards cellular technology does not exist and apprehension on the part of the general population from adopting unknown technologies due to this constraint is nearly unknown. Outsland (1974, p. 28) suggested that perceptions of innovations by potential adopters of innovative technology might be an effective prediction measure for adoption of the innovation, more than personal factors. Based on this assumption, the current research focused on the university student population, which represents a segment in society that is essentially aware of computer and Internet technologies, and therefore its apprehension from adopting unknown technologies is relatively low. In addition, in Israel the penetration percentage of cellular technology has already reached its full potential, and the interest of the current research is to examine the tendencies to adopt usage of cellular phones for mobile electronic commerce. In order to do so, a research questionnaire was constructed, including 35 items dealing with perceptions regarding the use of mobile electronic commerce technology. Each item in the questionnaire was assessed on three time scalesperceptions of usage in the past, at present, and an estimate of usage perceptions in future. The questionnaire was distributed amongst students from various departments at Ben Gurion University in the Negev. The distribution included most university departments. A total of 1,300 questionnaires were distributed. They were completed in the presence of the researcher and handed in directly.
ChARACTERISTICS OF ThE MODEL The model is based on the characteristics of the perceptions of innovativeness, which have been identified in previous studies, with a change in wording from “perception of innovativeness” to “perception of the use of innovation,” as suggested in the PCI model (Moore & Benbasat 1991). The characteristics are as follows: • •
414
Relative Advantage: The extent to which the use of an innovation is perceived as better than the use of its predecessor (based on work by Roger, 1983). Compatibility: The extent to which the use of an innovation is perceived as being persistent with other
•
•
• •
•
existing values, needs, and experiences of the potential adopters (Roger, 1983). Ease of Use: The extent to which individuals believe that the use of a specific system does not require investment of physical and emotional efforts (Davis, 1986). Results Demonstrability: The extent to which the results of using an innovation are tangible and presentable (Rogers, 1983, p. 232). Research has shown that merely being exposed to a product can in itself create a positive attitude toward it among individuals (Zajonc & Markus, 1982). Image: The extent to which the use of an innovation is perceived as improving the individual’s status in society. Visibility: The extent to which the results of the use of an innovation are visible to others. The characteristic “Observability,” which was mentioned by Rogers (1983), is presented in the PCI model via two variables (Results Demonstrability and Visibility). Trialability: The extent to which the use of an innovation can be experienced prior to its adoption.
RESULTS Out of 1,300 distributed questionnaires, 1,005 were completed correctly (11.49% missing data). The study results point to 55.6% potential users of cellular phones for m-commerce in two years’ time, compared to 34.8% two years ago and 38.9% users today. This is indicative of the fact that the questionnaire reflects the target population studied in the current research. The percentage of explained variance obtained in the model runs is 67.732% in the past, 65.896% at present, and 61.470% in the future. It is safe to say that the model for this study has been validated and verified. Therefore, we can conclude that in order to assess the probability of actual usage of cellular phone for m-commerce, the model for perceptions of use of cellular phone for m-commerce presented in this study may be used.
Predicting the Probability of Using M-Commerce After verifying the model, a test was carried out to identify which variables assist in predicting the probability for using cellular phones for mobile electronic commerce. The testing method used was Forward Stepwise Logistical Regression (Hosmer & Lemeshow, 2000), which first brings into the equation variables having the highest level of significance, and then re-calculates the level of significance
M-Commerce Technology Perceptions on Technology Adoptions
Table 1. Logistical regression equation constants in test run on perceptions categories Variable
B (past)
B (current)
B (future)
Relative Advantage
0.117
0
0
Image
0.096
0
0
Compatibility
0
0.182
0.385
Visibility
0.305
0.273
0.185
Cellular Phone Use
0
0
1.021
Method of Connecting with Internet Provider
0
0
-0.533
Equation Constant
-1.770
-1.501
-2.310
of each of the variables in the equation. For the past timeframe, the results were given after running the first step of the regression, while for the present and future timeframes, the results were given after three steps of the regression. The results show that for the period of two years ago, the average perceptions of potential adopters of the technology is a significant variable (p (u1,u2):{k14, k18}k12 KDC => (u5,u6,u7,u8):{k18}k58
For user3 leaving: KDC → u4:{k34, k14, k18}k4 KDC => (u1,u2):{k14, k18}k12 KDC => (u5,u6,u7,u8):{k18}k58
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Secure Group Communications in Wireless Networks
GROUP KEY MANAGEMENT ALGORITHM
Þ A => B: {x}: A sends message x to B via broadcast or multicast
The Proposed Logical Key Structure
Group key management algorithm is core part of multicast security. It maintains the logical key structure and performs the procedures to assign, distribute and update the group key and other KEKs.
Generally, a large multicast group is comprised of several smaller subgroups. Each member not only joins the group communications but also participates in one or some subgroup communications. Nevertheless the current proposed key tree cannot reflect such a group organization structure. Meanwhile a separate key tree must be constructed for each subgroup. To improve the performance of group key management under such scenario, we propose a new logical group keying structure shown in Figure 2. From Figure 2, we can see that the proposed structure is a multi-tier model. The root node instigates the group communication session, and holds the group key. Each subgroup represents a multicast session and is associated with two keys: subgroup key and subgroup KEK (key encryption key). Users in the subgroup are divided into several clusters. Each cluster has its own cluster key to distribute other keys in the cluster. In the user level, each user shares a secret key with base station. It is a common scenario that a user subscribes multiple subgroups simultaneously. To improve the efficiency, we introduce a new concept: super-subgroup, which is a virtual container to accommodate users who participate in multiple subgroups simultaneously. Each super-subgroup is a combination of several subgroups. For example, as shown in Figure 2, super-subgroup (A_B) is a super-subgroup combining subgroup A and B, where users 9 to 11 participate in the subgroup A and B concurrently. Super-subgroup can be a combination of any subgroups. According to the theory of permutation and combination, for a group having ns subgroups, the total number of super-subgroups is
Notation In this section, we depict the notations that we will use in the following sections: Þ u: user Þ bs: base station Þ n: the number of members in the subgroup Þ nc: the number of the cluster in the subgroup Þ ns: the number of subgroups Þ m: the number of users in the cluster Þ j: the number of multiple subgroups which user joins and leaves simultaneously Þ a: degree of the balance tree Þ d: the height of the balance tree (d = logan) Þ k: the encyption key Þ BS = {s1, s2, s3, ..., sn}: the set of the base stations
ns
∑c .
Þ {x}k: message x encrypted by the key k
i =2
i n
Þ A → B: {x}: A sends message x to B via unicast
Figure 2. Logical key management structure Group level key: Group key
Group
Subgroup A
Cluster 1
u1
u2
Subgroup B
Cluster 2
Cluster 3
u3
u4
Cluster 4
u5
u6
Subgroup level key: Subgroup key and subgroup KEK
super-subgroup (A_B)
Cluster 5
u7
u8
Cluster 6
u9
u 10
Cluster 7
Cluster level key: Cluster key
u 11
Key shared by base station and user
833
S
Secure Group Communications in Wireless Networks
The Proposed Group Key Management Algorithm There are three main operations in the wireless group key management (multiple subgroups): member join, member leaving and hand-off. The rekeying procedures of these operations occur independently in each wireless cell. We now illustrate our algorithm in each of these operations in the following subsections.
Member Join Join is a procedure that is invoked by a user who wants to become a member of a group. Backward secrecy should be achieved to prevent the new member from accessing the previous group communication contents. In the proposal, join is comprised of three steps: registration, key grant and key update. In the registration step, a user submits a join request to the local base station, and the base station forwards this request to upper layer area key controller (AKC) for authentication. u → bs: {JOIN_REQUEST} bs → AKC: {JOIN_REQUEST} In the second step, after AKC verifies the user, AKC assigns the user into a cluster and sends the relevant keys to the user. AKC → bs: {kcluster, ksubgroup, ksubgroup_kek, kgroup}
k (AKC_user)
bs → u: {kcluster, ksubgroup, ksubgroup_kek, kgroup} k (AKC_user) Finally, AKC invokes a join-key-update procedure to update the affected cluster, subgroup and group key. This updating can be achieved without multicasting rekeying message (Waldvogel, Caronni, Sun, Weiler, & Plattner, 1999). AKC informs the sender to use the new keys to encrypt the data traffic; the members will perceive the key change in ordinary data packets and update their keys locally by passing through a one-way hash function. AKC → sender: {USING NEW DATA ENCRYPTION KEY} When a user wishes to join multiple subgroups at the same time, the procedure is the same as mentioned above. The user is assigned into a suitable super-subgroup.
Member Leaving Leaving is invoked by a user who wants to quit a group or the AKC evicts the user from the group communications. 834
Forward secrecy must be achieved to prevent the departing user obtaining the forthcoming group communications. There are two steps in this operation: cancellation and key update. In the cancellation, user submits a leaving request to AKC: u → bs: {LEAVE_REQUEST} bs → AKC: {LEAVE_REQUEST} In the rekeying stage, according to the “bottom to top” principle, rekeying procedure starts from the affected cluster where the departure took place then to the remaining clusters in the affected subgroup, and finally in the entire group. First, AKC generates new keys, and constructs a single message containing all new keys by the group-oriented approach (Wong, Gouda, & Lam 2000). AKC => BS: {new kcluster, ksubgroup_KEK, ksubgroup, kgroup} Each base station searches for the affected cluster members in its cell and delivers the new keys to them. Base station employs user-oriented rekeying approach (Wong et al., 2000), that is, for each user the base station constructs a rekeying message that contains precisely the new keys needed by the user. bs ui:{new kcluster, ksubgroup_KEK, ksubgroup, kgroup}k(bs_user) i(i=1, 2...m) To update keys for the remaining clusters in the affected subgroup, the base station multicasts the rekeying message to the clusters members. The new keys are encrypted by the old group key: bs =>(cluster)i:{new ksubgroup_KEK, ksubgroup, kgroup} k(cluster)i(i=1,
nc ) 2…
Finally, the base station updates the group key for the remaining subgroups in the group. bs => (subgroup)i: {new kgroup} k (subgroup_KEK) i ( i = 1, 2, … ,
ns
)
If a multi-subgroup user leaves, the procedure is same. bs => (subgroup)i:{new ksubgroup, kgroup}k (subgroup_KEK) i ( i = 1, 2, … , j)
Handoff Handoff is a unique operation in the wireless group communications. Several approaches (DeCleene, Dondeti, Griffin, Hardjono, Kiwior, Kurose, Towsley, Vasudevan, & Zhang, 2001; Sun, Trappe, & Liu 2002) have been proposed to address the group key management during the handoff.
Secure Group Communications in Wireless Networks
We propose a simple and efficient handoff management scheme, which contains three steps: handoff registration, authentication and switching. We assume that the mobile devices can detect signals of the two adjacent base stations on the border of the wireless cell (Chen & Chao, 2004). In the registration step, user switches to the new base station and sends a handoff join message. Then user switches back to the old base station quickly. u → bsnew: {HANDOFF_JOIN} In the authentication stage, new base station contacts the old base station to verify the user: bsnew → bsold: {AUTHENTICATION_REQUEST} bsold → bsnew: {AUTHENTICATION_REPLY} In the last step, after a predefined time, user switches to the new base station to finish the handoff. Because the group keying structure is identical in the whole domain, users can move from one cell to another without invoking rekeying operation.
PERFORMANCE DISCUSSION In measuring the performance of a group key management system, many parameters need to be taken into consideration (Moyer, Rao, & Rohatgi, 1999; Rafaeli & Hutchison, 2003). However, efficiency is one of the most significant criteria. Communication, computation and storage overheads related to rekeying are the major efficiency measures for a key management algorithm. We take the popular tree-based key management scheme: logical key hierarchy (LKH) (Wong et al., 2000; Wallner, Harden, & Agee, 1999) as the benchmark in our performance analysis.
Communication Efficiency In order to achieve the best result, the communication during the rekeying and handoff combines the unicast and multicast. Communication overhead is recorded in big-O notation as a measure of the number of rekeying messages transmitted per operation. We evaluate the communication overhead for the following three operations: join, leave and handoff.
Join As described above, when a user joins a group having ns subgroups, there is no need to multicast rekeying message. The system only needs to inform the sender to apply the new keys. So the communication cost in this situation is
zero. According to the LKH algorithm, the scheme needs to construct separate ns trees to present all the subgroups. The communication cost of join is d (Wong et al., 2000; Wallner, Harden, & Agee, 1999). When a single subgroup join occurs, there is a join operation and group key updating for the rest (ns-1) subgroups. So the cost of rekeying communication for LKH is O(d+ns). When a user wants to join multiple subgroups simultaneously, with the concept of super-subgroup, the cost of communication in the proposal is still zero. As for the LKH algorithm, multiple trees are affected by the join operation and ns–j subgroups need to update their group key. So in such scenario, the communication overhead of LKH is the summation of cost of multiple join operation and ns–j group key updating, that is, j
O( ∑ di + ns − j ). i =1
Leaving When a single subgroup member leaves, the proposed algorithm needs to rekey four keys: cluster, subgroup, subgroup KEK and group key. The overhead of rekeying communication is shared by AKC and base station network. AKC is responsible for the generation of new keys. By using the group-oriented approach, all the new keys are contained in a single rekeying message, so the communication cost of AKC is O(1). Base station is in charge of the key distribution in its cell. There are two scenarios in the cell: the leaving user present in the cell or the leaving user not in the cell. In the latter situation, the base station can use multicast for rekeying, the rekeying communication is just one or two messages. Therefore we focus our attention on the former scenario. According to our proposal, first, base station sends the rekeying message to the affected cluster users via unicast, so the cost is m-1. Then, base station multicasts the rekeying message to the members of the remaining clusters in the affected subgroup. The overhead of this communication is nc-1. Finally, base station needs to update the group key for the rest of subgroups in the group and the cost is ns -1. Thus the total cost of rekeying communication of base station is O(m +nc+ ns). As for the LKH, the leaving cost is (a – 1)d (Wong et al., 2000; Wallner, Harden, & Agee 1999). When single subgroup departure happens, the communication cost of LKH is one leaving cost plus group key updating for (ns1) subgroups. Therefore the overhead of LKH for single subgroup leaving is O ((a – 1)d + ns). When a member quits multiple subgroups simultaneously, multiple subgroups are affected. Under such scenario, AKC’s communication cost is still O(1), because all the new keys are still in one message. By using the subgroup KEK, base station has the same cost as that of single subgroup leav835
S
Secure Group Communications in Wireless Networks
Table 1. The communication cost comparison of our proposal and LKH join Single
Leaving
Multi-subgroup
our scheme
zero
zero
LKH
O(d+ns)
O(
Single AKC: O(1)
AKC: O(1)
BS: O(m +nc+ ns)
BS: O(m +nc+ ns)
O( ( − 1)d + ns )
O( ∑ ( − 1)di + ns − j )
j
∑ di + n s − j ) i =1
handoff
Multi-subgroup zero
j
O(d)
i =1
Note: nc = the number of cluster of the subgroup; ns: = the number of the subgroups; M = the size of the cluster; α = the degree of the balance tree; j = the number of multiple subgroups which user joins or leaves simultaneously; d = the height of the tree
ing: O(m +nc+ ns). As we described above, LKH algorithm needs multiple leaving operations. Additionally, the group key of (ns–j) subgroup needs to be updated. So the overhead of communication of LKH is: j
O( ∑ ( − 1)di + ns − j ). i =1
Handoff Because of the identical logical keying structure in all the areas, users can freely move from one cell to another without invoking rekeying procedure. As for the LKH algorithm, some handoff procedures at least need a join operation (DeCleene et al., 2001). So the rekeying communication cost is O(d), d is the height of the balance tree.
Summary We tabulate the efficiency evaluation in Table 1. From the table, we can see that our proposal has advantages over the LKH, especially in the handoff and multi-subgroup join and leaving.
Computation Efficiency Computation efficiency is to measure the cost of computation during the rekeying process in the wireless devices. Because of the computation limitation of the wireless device, it can not afford the expensive computation function, such as exponentiation, PKI calculation and so on. In our proposal, there are only two kinds of computations performed in the wireless devices: one-way hash function and symmetric encryption/decryption. In reference to the current mobile technology, these calculations were confirmed that they could be operated in a fast and efficient manner in the wireless devices. So the proposed system can achieve good computation efficiency which makes it suitable for the wireless network environment. 836
Storage Efficiency The storage efficiency is to measure the number of keys stored in AKC and mobile devices. We compare the key storage cost of our proposal with that of LKH (Wong et al., 2000; Wallner, Harden, & Agee 1999) in Table 2. From the table, we can see that, on the server side, the keys stored in LKH algorithm are linearly proportional to the number of group users. In the proposal, the key storage cost is proportional to the number of clusters, which is much less than the number of users. On the user side, there are only 4 keys stored for a single subgroup membership and j+3 keys for the multi-subgroup memberships in our proposal. For the LKH, in the same scenarios, users need to store d + 1 and j
∑ di + j + 1 i =1
keys in single and multiple subgroup scenario respectively.
Table 2. The key storage cost comparison of our proposal and LKH Our proposal ns
LKH ns
ni − 1 +1 −1
KDC/AKC
1 + ∑ (nc )i + ns
Users (single subgroup)
4
d+1
Users (multi-subgroup)
j+3
∑ di + j + 1
i =1
∑ i =1
j
i =1
Note: N = the number of subgroup members; nc = the number of cluster in the subgroup; ns: = the number of the subgroups; α = the degree of the balance tree; j = the number of multiple affected subgroups; d = the height of the tree
Secure Group Communications in Wireless Networks
FUTURE TRENDS Along with the fast development of wireless communications and fast capacities improvement on mobile devices, more and more multicast group applications and services will be emerging on wireless networks. The future research will focus on the efficiency and security of group key management system. Additionally, new architecture and framework will be proposed to address the wireless multicast security.
CONCLUSION Here, we present a new group key management solution to secure multicast communications in wireless networks. This proposed solution has distributed two-tier architecture and clustered hierarchical keying structure based on the group organization chart. The group key management system can perform the multi-subgroup access control in an efficient way. Compared with the existing tree-based key management scheme, this proposed scheme can significantly reduce the overhead associated with communication, computation and storage.
REFERENCES Amir, Y., Kim, Y., NitaRotaru, C., Schultz, J. L., Stanton, J., & Tsudik, G. (2004). Secure group communication using robust contributory key agreement. IEEE Transactions on Parallel and Distributed Systems, 15(5), 468-480 Banerjee, S., & Bhattacharjee, B. (2002). Scalable secure group communication over IP multicast. IEEE Journal on Selected Areas in Communications, 20(8), 1511-1527 Chen, J., & Chao, T. (2004). IP-based next-generation wireless networks. NJ: John Wiley & Sons. DeCleene, B., Dondeti, L. D., Griffin, S., Hardjono, T., Kiwior, D., Kurose, J., et al. (2001). Secure group communications for wireless networks. Military Communications Conference. Communications for Network-Centric Operations: Creating the Information Force. IEEE, 1, 113-117. Harney, H., & Muckenhirn, C. (1997). Group key management protocol (KGMP) architecture. RFC 2094. Kostas, T., Kiwior, D., Rajappan, G., & Dalal, M. (2003). Key management for secure multicast group communication in mobile networks. In Proceedings of DARPA Information Survivability Conference and Exposition (Vol. 2, pp. 41-43).
Mittra, S. (1997). Iolus: A framework for scalable secure multicasting. Processing of the ACM SIGCOMM, 27(4), 277-288 Moyer, M. J., Rao, J. R., & Rohatgi, P. (1999). A survey of security issues in multicast communications. IEEE Network, 13, 12-23 Pahlavan, K., & Krishnamurthy, K. (2002). Principles of wireless networks: A unified approach. NJ: Pearson Education, Inc. Perrig, A., Song, D., & Tygar, D. (2001). ELK, A new protocol for efficient large-group key distribution. In Proceedings of IEEE Symposium on Security and Privacy (pp. 247-262). Rafaeli, S., & Hutchison, D. (2003). A survey of key management for secure group communication. ACM Computing Surveys, 35(3), 309-329. Sherman, A. T., & McGrew, D. A. (2003). Key establishment in large dynamic group using one-way function trees. IEEE on Software Engineering, 29(5), 444-458 Steiner, M., Tsudik, G., & Waidner, M. (1996). Diffie-hellman key distribution extended to group communication. In Proceedings of the 3rd ACM Conference on Computer and Communications Security (pp. 31-37). Sun, Y., Trappe, M., & Liu, K. J. R. (2002). An efficient key management scheme for secure wireless multicast. In Proceedings of IEEE International Conference on Communication (pp. 1236-1240). Waldvogel, M., Caronni, G., Sun, D., Weiler, D., & Plattner, B. (1999). The Versakey framework: Versatile group key management. IEEE Journal on Selected Areas in Communications, 17(8), 1- 15. Wallner, D. M., Harden, E. J., & Agee, R. C. (1999). Key management for multicast: Issues and architecture. RFC 2627.
KEY TERMS Backward Secrecy: To prevent new group members from accessing previous group communications, which they may have recorded. Forward Secrecy: To prevent departing members from decoding future group data traffic. Handoff: In a cellular wireless network, handoff is the transition of signal for any given user from one base station to a geographically adjacent base station as the user moves around.
837
S
Secure Group Communications in Wireless Networks
Key Encryption Key (KEK): This kind of keys is used to encrypt the other keys for distribution in the multicast group. Key Management Algorithm: In the group key management system, an algorithm is applied to maintain the logical key structure held by the group members and other entities.
838
Logical Key Hierarchy (LKH): This algorithm is a tree structure for efficient group rekeying. Each node of the tree represents a key, with the root node representing the group key. Each leaf node represents a group member, and each member knows all the keys in its path to the root. Multicast: Multicast is communication mechanism to delivery a single message to multiple receivers on a network. The message will be duplicated automatically by routers when multiple copies are needed.
Category: Security 839
Security Architectures of Mobile Computing Kaj Grahn Arcada Polytechnic, Finland Göran Pulkkis Arcada Polytechnic, Finland Jonny Karlsson Arcada Polytechnic, Finland Dai Tran Arcada Polytechnic, Finland
INTRODUCTION Mobile Internet users expect the same network service quality as over a wire. Technologies, protocols, and standards supporting wired and wireless Internet are converging. Mobile devices are resource constrained due to size, power, and memory. The portability making these devices attractive also causes data exposure and network penetration risks. Mobile devices can connect to many different wireless network types, such as cellular networks, personal area networks, wireless local area networks (WLANs), metropolitan area networks (MANs), and wide area networks (satellitebased WANs). Wireless network application examples are e-mailing, Web browsing, m-commerce, electronic payments, synchronization with a desktop computer, network monitoring/management, and reception of video/audio streams.
BACKGROUND Major security threats for mobile computing devices are (Olzak, 2005): • • •
theft/loss of the device and removable memory cards, wireless connection vulnerabilities, and malicious code.
Mobile computing devices are small, portable, and thus easily lost/stolen. Most mobile platforms only include support for simple software-based password login schemes. These schemes are easily bypassed by reading information from the device without login. Memory cards are also easily removed from the device. Mobile devices support wireless network connections such as Bluetooth and WLAN. These connections are typi-
cally by default unprotected and thus exposed to eavesdropping, identity theft, and denial-of-service attacks. Malware has constituted a growing threat for mobile devices since the first Symbian worm (Cabir) was detected in 2004. Mobile devices can be infected via MMS, Bluetooth, infrared, WLAN, downloading, and installing from the Web. Current malware is focused on Symbian OS and Windowsbased devices. Malware may result in (Olzak, 2005): • • • •
loss of productivity, exploitation of software vulnerabilities to gain access to resources and data, destruction of information stored on a SIM (subscriber identity module) card, and hi-jacking of airtime resulting in increased costs.
WIRELESS SECURITY PRINCIPLES Security Policy Examples of rules proposed for mobile device end users are: • •
•
I agree to make sure my device is password protected and that latest security patches are installed. I agree to keep a firewall/anti-virus client with latest anti-virus signatures installed, and to use a remote access VPN client, if I will connect to the corporate network. I agree to use the security policies recommended by the corporate security team.
Examples of rules proposed for administrators of mobile devices in corporate use are:
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Security Architectures of Mobile Computing
•
End-users get mobile network access after agreeing to the end-user rules of behavior. Handheld firewalls shall be configured to log security events and send alerts to security-manager@company. com. Handheld groups and Net groups shall have restricted access privileges and only to needed services.
•
Handheld security policies should be automated by restrictive configuration settings for handhelds, firewalls, VPNs, intrusion detection systems, and directory servers (Handheld Security, 2006).
•
• •
Storage Protection Mobile device storage protection is online integrity control of all stored program code and all data, optional confidentiality of stored user data, and protection against unauthorized tampering of stored content. Protection should include all removable storage modules used by the mobile device. The integrity of the operating system code, the program code of installed applications, and system and user data can be verified by checksums, cyclic redundancy codes (CRCs), hashes, message authentication codes (MACs, HMACs), cryptographic signatures, and so forth. However, only hardware protection of verification keys needed by MACs, HMACs, and signatures provide strong protection against tampering attacks. Online integrity control of program and data files must be combined with online integrity control of the configuration of a mobile device for protection against malware intrusion attempts. User data confidentiality can be granted by file encryption software. Such software also protects integrity of stored information, since successful decryption of an encrypted file is also an integrity proof.
Security Layers Mobile computing security layers are based on the OSI (Open Systems Interconnection) Security Model. Defined security services are authentication, access control, non-repudiation, data integrity, confidentiality, assurance/availability, and notarization/signature (ISO/IEC 7498-1, 1994; ISO 7498-2, 1989). Specific wireless security architecture issues include Mobile IP security features, and link-level and physicallevel security protocols of wireless access technologies like WLAN, GPRS, and Bluetooth Mobile IP security means that: •
840
a mobile node, which is a mobile device, has the same connectivity and security in a visited foreign network as in its home network; and
the home network and visited foreign networks have protection against active/passive attacks. These security goals require:
•
•
that Mobile IP registration and location update messages have data integrity protection, data origin authentication, and anti-replay protection; access control to foreign network resources used by visiting mobile nodes; and that IP packet redirecting tunnels provide data integrity protection, data origin authentication, and data confidentiality.
Moreover, mobile nodes should have location privacy and anonymity (Zao et al., 1999). Replay prevention with timestamps or nonces for all mobile IP messages is specified in Perkins and Calhoun (2000). Other mobile IP security solutions are authentication schemes and protection of data communication (Calhoun et al., 2005; Barun & Danzeisen, 2001; Hwu, Chen, & Lin, 2006).
Identification Hardware Identification hardware contains user information and cryptographic keys used to authenticate users to mobile devices, applications, networks, and network services. The following identification hardware types are used: • • • •
subscriber identity module (SIM), public key infrastructure SIM (PKI SIM), universal SIM (USIM), and IP multimedia services identity module (ISIM).
SIM A basic SIM card is a smartcard securely storing a key (Ki) identifying a GSM network user. A SIM card is a microcomputer executing cryptographic operations with Ki. The SIM card also stores SMS (short message service) messages, MMS (multimedia messaging system) messages, and a phonebook. The use and content of a SIM card is PIN protected (Rankl & Effing, 2003).
PKI SIM A PKI SIM card is a basic SIM card with added PKI functionality. An RSA co-processor is added for public key-based encryption and signing with private keys. The PKI SIM card stores private keys and certified public keys needed for digital signatures and encryption (Setec, 2006).
Security Architectures of Mobile Computing
USIM A USIM card is a SIM used in 3G mobile telephony networks. The physical size is the same as for a GSM SIM card, but hardware is different. USIM is actually an application running on a UICC (universal integrated circuit card) storing a pre-shared secret key (Lu, 2002).
ISIM An ISIM card consists of an application (ISIM) residing on a UICC. ISIM provides secure authentication of handheld users to IMS (IP multimedia system) services (Dietze, 2005).
Wireless Security Protocols Security protocols are—for wired networks—implemented by (Perelson & Botha, 2004): authentication services, confidentiality services, non-repudiation services, and authorization. Four wireless security protocol types are needed: • • •
access control to mobile devices, local access control to networks and network services, remote access control to networks and network services, and protection of data communication to/from mobile devices.
and-error attacks due to their limited length and alphabet. Passwords are more secure than PINs since their length and alphabet are larger (Jansen, 2003).
Visual and Graphical Login Visual authentication means that a user must remember image sequences to authenticate to a mobile device. A picture password system can be designed to require a sequence of pictures or objects matching a certain criteria and not exactly the same pictures. For example, the user must find a certain number of objects with four sides. This makes the shoulder surfing quite difficult (Duncan, Akhtari, & Bradford, 2004).
Biometrics Biometric user authentication is a hardware solution for examining one or more physical attributes of an authorized user. Biometric controls, such as fingerprints, are becoming more common in handheld devices (Perelson & Botha, 2004).
Authorization
Different protocols are presented in Markovski and Gusev (2003).
Usually mobile devices are personal, and authentication infers that the user is authorized. A corporate handheld device may however be used by several employees and may contain confidential company information. Needed user authorization features for such mobile devices include (Perelson & Botha, 2004):
Access Control to Mobile Devices
•
Access control must be implemented on a mobile device itself to prevent unwanted access to confidential data stored in the device (see Figure 1). Authentication confirms a claimed user identity.
•
PIN and Password Authentication
Local Network Access
A PIN is four digits from a 10-digit (0-9) keypad. However, PINs are susceptible to shoulder surfing or to systematic trial-
Local network access protocols depend on the wireless access network type (WLAN, Bluetooth, Cellular Network, etc.). A
•
•
File Masking: Some files cannot be viewed by unauthorized users. Access Control Lists: User-related object permissions. Role-Based Access Control: User role-related permissions.
Figure 1. The access control principle
1. The user presents an identity (e.g., password or biometric) 2. The user’s identity is confirmed 3. An authenticated user is allowed access to a resource
841
S
Security Architectures of Mobile Computing
WLAN is usually an access network to a LAN. Authentication for LAN resources is thus also needed unless WLAN authentication is integrated in a single-sign-on scheme. Local network access protocols are described in later sections.
Remote Network Access Secure remote network access from a mobile device requires a VPN (virtual private network), which is a protected data path in an existing unsecured network to a private LAN. VPNs can be based on different protocols: IPSec (IP security), SSL (secure socket layer), or SSH (secure shell).
IPSec VPN IPSec operates at the network layer of the OSI model. IPSec protocols are: • • •
ESP (encapsulating security payload) for authentication, data confidentiality, and message integrity; AH (authentication header) for authentication and message integrity; and IKE (Internet key exchange protocol) for encryption key exchange.
IPSec VPNs require VPN client software in mobile devices (Davis, 2001).
SSL VPN The encrypted tunnel is established at the session layer of the OSI model. SSL VPN clients communicate with the VPN gateway using an SSL-supported application such as a Web browser or e-mail client. No separate VPN client software is therefore needed (Steinberg & Speed, 2005).
SSH SSH (secure shell) is a protocol for login to and executing commands on a remote UNIX computer. SSH provides between two communicating hosts an encrypted communication channel, which can be used for port forwarding with VPN functionality (Barret et al., 2005).
Protection of Data Communication Security protocols for protection of wireless data communication are integrated in protocols for local and remote access to networks/network services. In a cellular network a shared secret session key created by the authentication protocol is used for encryption/decryption of data communication. In a WLAN, the TKIP (temporal key integrity protocol) is integrated in the WPA security protocol, and AES (advanced 842
encryption standard) is integrated in the WPA2 security protocol. The remote access protocols IPSec, SSL/TLS, and SSH also provide end-to-end protection of data communication with secure symmetric encryption algorithms and shared secret session keys created during authentication.
PLATFORMS FOR INTEGRATED ARCHITECTURES Software signing and binary trust-models do not provide adequate protection against third-party programs. Fine-grained software authorization is emerging into mobile units. Typical examples include Java sandboxing and Symbian platform security. Software-based mobile platform examples are Java Mobile Environment, Symbian OS, Embedded Linux, Windows Mobile, Brew (Binary Runtime for Wireless), Blackberry OS, and Palm OS. OS implementation vulnerabilities still remain a challenge. Integrated solutions have been proposed for executing trusted code and for secure boot. Standardization efforts are under development (e.g., Trusted Computing Group and Trusted Mobile Platform). There are different embedded on-chip security solutions, but mostly the security solution relies on combining hardware and software. Platform security examples are Texas Instruments OMAPTM Platform (Sundaresan, 2003) and Intel Wireless Trusted Platform (Intel Corporation, 2006b). The TI platform relies on three layers of security: application layer security, operating system layer security, and on-chip hardware security. The main security features are: • • • •
A secure environment provides secure execution of critical code and data by secure mode, secure keys, secure ROM, and secure RAM. Secure boot/flash prevents security attacks during device flashing/booting. Run-time security is included for security-critical tasks like encryption/decryption, authentication, and secure data management. A hardware crypto engine is also included for DES/ 3DES, SHA1/MD5, and RNG with two configuration modes: secure mode and user mode.
Intel platform building blocks are performance primitives (hardware) and cryptographic primitives (optimized software) for security services. Platform components include • • • •
trusted boot ROM integrity validation and booting to a correct configuration; wireless trusted module processing secrets; security software stack enabling access to platform resources through standard cryptographic APIs; protected storage in system flash for secrets; and
Security Architectures of Mobile Computing
•
physical protection by security hardware in a single device and discrete components in a single physical package.
WIRELESS APPLICATION SECURITY The risks described above should be addressed in wireless application design. Wireless application security includes (Umar, 2004): application access control, client/server communications security, and anti-malware protection.
such as credit card information stored in memory by wireless applications. Time and space for sensitive data in memory should be minimized (Intel Corporation, 2006a).
SECURITY OF MOBILE TECHNOLOGIES A taxonomy of mobile technologies is: •
Application Access Control
•
Many mobile platforms lack support for individual user accounts and for operating system-level logon. Mobile applications handling confidential data should require user authentication before application access is granted. In case a mobile device is lost or stolen while the device user is logged in to an application, the application should also support “session timeout.” This means that a limited inactive time is specified for an application before re-authentication is required (Intel Corporation, 2006a).
•
Client/Server Communication Security Typical wireless Internet connections are: 1. 2.
the wireless connection between a mobile device and an access device, and the Internet connection between the mobile device and the Internet host/server via the access device.
Internet connection security should be provided at the application level. For Web-based client/server applications, the SSL protocol provides encryption and signing of transmitted data. SSL application examples are: • • •
Web browsers for secure communications with Web servers, e-mail client software for secure reading of e-mail messages on e-mail servers, and SETs (secure electronic transactions) for secure financial transactions with credit cards.
For applications using customized protocols, security protocols are also customized. Alternatively, VPN techniques can be used.
Anti-Malware Protection Most current mobile operating systems lack memory space protection. Malware can access and steal application data,
•
wireless cellular networks (GSM, DECT, GPRS, and UMTS), wireless long-range networks (WiMax, Satellite Communication Technology), wireless local area networks (WLAN, ZigBeeTM), and wireless short-range networks (Bluetooth, Wireless USB).
Wireless Cellular Networks First Generation First-generation cellular systems, such as AMPS (advanced mobile phone system) introduced in the early 1980s, use analog transmission and provide no security.
Second Generation 2G cellular systems, such as GSM (Global System for Mobile Communications) introduced in the late 1980s and DECT, use digital transmission. GSM security is based on a unique IMSI (International Mobile Subscriber Identity) and a unique secret key (Ki) stored in the SIM card of each subscriber. The Ki is never transmitted over the network. Every GSM network has: • • •
•
AUC (authentication center), a protected database containing a copy of Ki; HLR (home location register) for subscriber information; VLR (visitor location register) for information of each mobile station currently located in the geographical area controlled by the MSC (Mobile Station Controller); and EIR (equipment identity register) for lists of mobile stations on the network. Stations have unique IMEI (International Mobile Equipment Identity) numbers.
When a mobile station enters a GSM network for the first time, the IMEI is transmitted for determination in which AUC/HLR subscriber data is stored. The MSC/VLR of the visited network asks for and stores a security triplet 843
S
Security Architectures of Mobile Computing
(a unique random number RAND, a signed response SRES, a ciphering key Kc) from the AUC/HLR. SRES and Kc are calculated from RAND with Ki. Subscriber authentication: • • • •
RAND is sent to the mobile station. SRES’ and Kc’ are calculated from RAND with Ki. SRES’ is sent back to MSC/VLR. Authenticated if SRES=SRES’.
Kc=Kc’ is used for radio link encryption/decryption. After the initial registration, IMSI is stored in the VLR. A TMSI (temporary mobile subscriber identity) is generated, transmitted back to the mobile station, stored in the SIM card, and used for future subscriber identification in the visited network. DECT is a cellular system and a common standard for cordless telephony, messaging, and data transmission standardized by ETSI (European Telecommunications Standards Institute). DECT is similar to GSM, but cell ranges are shorter (DECT, 2006). DECT uses several advanced digital radio techniques for efficient radio spectrum utilization. It enables high speech quality and security with low radio interference risks and low-power technology. Mobility management, responsible for DECT communication security, consists of procedures for identity, authentication, location, access rights, key allocation, parameter retrieval, and ciphering (Umar, 2004).
2.5 Generation The GPRS (2.5G) infrastructure equals GSM. TSMI is replaced by P-TMSI (packet TMSI) and by P-TMSI signature as alternate identities. Mapping between IP addresses and IMSI is generated in the HLR GPRS Register. GPRS authentication is performed by SGSN (serving GPRS support node). As a consequence, user data and signaling are encrypted all the way from the mobile station to the SGSN. Tunneling, firewalls, and private IP techniques are used. IP addresses are assigned after authentication and encryption algorithm negotiations.
Third Generation UMTS, Universal Mobile Telecommunications System, a standard for third-generation (3G) systems for mobile communication, referred to as International Mobile Telecommunications 2000 (IMT-2000) and initiated by the International Telecommunication Union (ITU), is presently being developed by the Third Generation Partnership Project (3GPP). The UMTS security architecture is based on 2G/2.5G security. Some GSM security features have been improved and some new features have been added. The UMTS security 844
Figure 2. UMTS functional security architecture; UE is user equipment and RNC is radio network controller
UIC SN
UIC UE
AKA USIM
AKA SN
DC UE
DC RNC
DI UE
DI RNC
AKA HLR
mechanisms are (see Figure 2): user identity confidentiality (UIC), authentication and key agreement (AKA), confidentiality of user and signaling data (DC), and integrity of signaling data (DI). See Lu (2002) for UMTS security details.
WAP WAP (wireless application protocol) is an open mobile device application standard. WAP security protocols and specifications are being developed by the WAP Forum (Open Mobile Alliance, 2006). The evolution of WAP security specifications is shown in Figure 3.
Figure 3. The development of WAP security specifications
WAP 1.0
WAP 1.1
WAP 1.2
WAP 1.2.1
WAP 2.0
WTLS (April, 98)
WTLS (Feb., 99)
WTLS (Nov., 99)
WTLS (Feb., 00)
WTLS (Apr., 01)
WMLSCrypt (Nov. 99)
WMLSCrypt (Nov. 99)
WIM (Nov. 99)
WIM (Feb., 99)
E2E Sec (Jun., 01) TLS (Apr., 01) WMLSCrypt (Jun., 01) WIM (Jul., 99) WPKI (Apr., 01) WAPCert (May., 01)
Apr. 98
Jun. 99
Dec. 99
Jun. 00
Jul. 01
Security Architectures of Mobile Computing
WTLS/TLS/SSL SSL/TLS are TCP-based security protocols for communication in client/server applications. WAP 2.0 adopts TLS as security protocol and supports the tunneling of SSL/TLS sessions through a WAP/WAP proxy. TLS/SSL in WAP 2.0 is a complement to the similar UDP-based WTLS protocol in earlier WAP versions. Server authentication and mutual authentication are options in WTLS/TLS/SSL-protected WAP applications.
WMLScript Crypto Library, WIM, and WPKI The lack of non-repudiation services and end-user authentication was addressed in WAP 1.2. The WMLScript (Wireless Markup Language Script) Crypto Library provides cryptographic functionality for WAP clients. WAP identity module (WIM) is used in WTLS and application-level security functions. A WIM stores and processes user authentication information, such as private keys. A WIM implementation example is a mobile phone S/WIM card (combined SIM and WIM). WPKI (wireless public key infrastructure) is a mobile environment PKI supported since WAP 2.0 (Open Mobile Alliance, 2006).
i-mode i-mode is a Japanese competitor to WAP for m-commerce. i-mode security features are: •
Table 1. Bluetooth service levels Authorization
Authentication
Encryption
Trusted
Yes
Yes
Yes
Untrusted
No
Yes
Yes
Unknown
No
No
Yes
Authentication Bluetooth device authentication is a unidirectional or mutual challenge/response process. Secret keys, called link keys, are generated either dynamically or by pairing. For dynamic link key generation, a passkey—the same passkey—must be entered in both connecting devices each time a connection is established. In pairing, a long-term stored link key is generated from a user-entered passkey, which can be automatically used in several connection sessions between the same devices.
Authorization In authorization, a Bluetooth device determines whether or not another device is allowed access to a particular service. Levels of trust are trusted, untrusted, or unknown. Service levels are shown in Table 1.
Encryption
protection of the radio link between the i-mode handset and the base station, encryption/authentication of data transmitted between i-mode mobile devices and Web sites, and protection of private network links between the i-mode center and special service providers like banks.
Bluetooth data transmission uses 128-bit encryption. Encrypted data can only be viewed by a device owning the proper decryption key. The encryption key is based on the link key.
The radio link is protected using SSL and other protocols, which are not public. Security of Web site connections and private network links are based on SSL. Mutual certificate authentication is supported (Umar, 2004).
ZigBee is a low-cost, low-power communications standard for wireless data communication in home and building automation. The ZigBee stack architecture is based on the standard OSI model. The IEEE 802.15.4-2003 standard defines the physical (PHY) layer and the medium access control (MAC) sub-layer. The ZigBee Alliance builds on this foundation by providing the network (NWK) layer and a framework for the application layer with: the application support sub-layer (APS), ZigBee device objects (ZDO), and manufacturer-defined application objects. Security services are defined for key establishment, key transport, frame protection, and device management. The MAC, NWK, and APS layers are responsible for the secure transport of their respective frames. Data encryption uses the symmetric key 128-bit AES algorithm. Frame integrity is protected, since frames cannot be modified by parties without cryptographic
• •
Bluetooth Bluetooth provides wireless short-distance transmission of data and voice signals between electronic devices. The specifications are defined by Bluetooth SIG (2006). The security is based on authentication, authorization, and encryption. The security modes are: 1. 2. 3.
no security measures, security measures based on authorization, and authentication and encryption.
ZigBee
845
S
Security Architectures of Mobile Computing
keys. Replayed data frames are rejected by a frame freshness verification function of the NWK layer. Furthermore, the APS sub-layer establishes and maintains security relationships. ZDO manages the security policies and the security configuration of a device. Access control uses a list of trusted devices maintained by a ZDO (ZigBee Alliance, 2004).
WLAN
Temporal key integrity protocol (TKIP) to provide dynamical and automatically changed encryption keys, and IEEE 802.1X and EAP (extended authentication protocol) to provide strong user authentication.
•
CCMP (cipher block chaining message authentication protocol) is an IEEE 802.11i protocol that uses the AES (advanced encryption standard) to provide stronger encryption than TKIP.
WiMax WiMax is a new technology for wireless broadband Internet access. The MAC layer of the WiMax network stack has a security sub-layer with (Puthenkulam & Yin, 2005): Figure 4. IEEE 802.11i features in WPA
802.11i 802.1X Other Features
Basic Service Set Independent Basic Service Set Preauthentication Key Hierarchy Key Management Cipher & Authentication Negotiation
Data Privacy Protocols TKIP CCMP (AES)
846
• • •
Broadband mobile communication is supported by a WLAN, which gives mobile users LAN connectivity through a high-speed radio link. Major WLAN security standards are (Pulkkis, Grahn, Karlsson, Martikainen, & Daniel, 2005): IEEE 802.11/WEP, WPA, and IEEE 802.11i. WEP is not recommended due to security flaws. Data encryption is based on static encryption keys, and no user authentication mechanisms are specified. WPA addresses the WEP vulnerabilities and is based on IEEE 802.11i (see Figure 4). The main features of WPA are: •
•
Wi-Fi Protected Access
a base station device and mobile user authentication capability based on the EAP protocol, X.509 certificates, and AAA servers (Radius, Diameter); encryption key management using the privacy key management protocol (PKM) v2; AES-CCM authenticated encryption of all data communication—the Encryption Key Refresh Mechanism supports high data rates; and CMAC (cipher-based message authentication code) and HMAC (hash-based message authentication code), which handle control message integrity protection.
Wireless USB An USB wire provides two security services: (1) a wanted interconnection of two devices is created, and (2) all data in transit is protected from casual observation or malicious modification by external parties. The goal of Wireless USB security is to provide analogous security services. Hosts and wirelessly connected devices are required to authenticate each other to avoid man-in-themiddle attacks. Data communication between a host and a wirelessly connected device is confidential and integritychecked by AES-128/CCM encryption. Secret encryption keys are shared by mutually authenticated hosts and wirelessly connected devices (Wireless, 2005).
Satellite Communication Technology A communication satellite permits two or more earth stations to send radio messages to each other over far distances. For satellite communication security it is necessary that earth stations have significant physical security, and RF (radio frequency) communication channels between satellites and earth stations are protected. Satellite communications are normally secured by scrambling satellite signals using cryptography or transmitting same signals over several frequencies. The data bits are basically transmitted on different signals based on a secret scheme. The receiver of a signal must thus be aware of the secret scheme. Additional security protocols like IPSec can be used to encrypt radio messages. However, such protocols slow down data transmission. The main challenge is thus to find a good balance between performance and security (Umar, 2004).
FUTURE TRENDS Privacy, security, and trust issues are and will be of major importance. The growth of the Internet and m-commerce will dramatically increase the amount of personal and corporate information that can be captured or modified. In the near
Security Architectures of Mobile Computing
future ubiquitous computing systems will accentuate this trend. We can likewise expect an increase in privacy and security risks, not only with the emergence of mobile and wireless devices, but also with sensor-based systems, wireless networking, and embedded devices. Ubiquitous computing technologies will probably suffer from the same sorts of unforeseen vulnerabilities that met the Internet society.
CONCLUSION Mobile terminals face security threats due to openness. Platforms are open for external software and content. Malicious software, like Trojan horses, viruses, and worms, has started to emerge. Fine-grained software authorization has been proposed. Downloaded software may then access particular resources only through user authorization. OS implementation vulnerability still remains a challenge because of difficulties in minimizing OS code running in privileged mode. Integrated hardware solutions may be the solution. Wireless security architectures have many options, and many standards/protocols addressing wireless security are quite recent, especially standards/protocols based on public key cryptography. Therefore more practical experience from the use of these protocols/standards in mobile computing is needed for reliable estimation of the provided security.
REFERENCES Barrett, J.D., Silvermann, E.R., & Byrnes, G.R. (2005). SSH, the secure shell: The definitive guide (2nd ed.). O’Reilly. Barun, T., & Danzeisen, M. (2001). Secure mobile IP communication. Proceedings of the IEEE 26th Annual Conference on Local Computer Networks (pp. 586-593).
Duncan, M. V., Akhtari, M. S., & Bradford, P. G. (2004). Visual security for wireless handheld devices. JOSHUA—Journal of Science & Health at the University of Alabama, 2. Handheld Security. (2006). Laura Taylor, part I-V (20042005). Retrieved August 8, 2006, from http://www.firewallguide.com/pda.htm Hwu, J.-S., Chen, R.-J., & Lin, Y.-B. (2006). An efficient identity-based cryptosystem for end-to-end mobile security. IEEE Transactions on Wireless Communication. Intel Corporation. (2006a). Wireless application security: What’s up with that? Retrieved August 8, 2006, from http://www.intel.com/cd/ids/developer/asmo-na/eng/57399. htm?page=1 Intel Corporation. (2006b). Intel wireless trusted platform: Security for mobile devices. Retrieved August 8, 2006, from http://www.intel.com/design/pca/applicationsprocessors/whitepapers/300868.htm ISO/IEC 7498-1. (1994). Information technology—Open systems interconnection—Basic reference model: The basic model, 1994. ISO 7498-2. (1989). Information processing systems—Open systems interconnection—Basic references model—Part 2: Security architecture, 1989. Jansen, W. A. (2003, May 12-15). Authenticating users on handheld devices. Proceedings of the 15th Annual Canadian Information Technology Security Symposium (CITSS), Ottawa, Canada. Retrieved August 8, 2006, from http://csrc. nist.gov/mobilesecurity/publications.html#MD Lu, W.W. (2002). Broadband wireless mobile, 3G and beyond. New York: John Wiley & Sons.
Bluetooth SIG. (2006). The official Bluetooth wireless info site. Retrieved August 8, 2006, from http://www.bluetooth. com
Markovski, J., & Gusev, M. (2003, April). Application level security of mobile communications. Proceedings of the 1st International Conference Mathematics and Informatics for Industry (MII 2003) (pp. 309-317), Thessaloniki, Greece.
Calhoun, P., Johansson, T., Perkins, C., Hiller, T., & McCann, P. (2005, August). Diameter mobile IPv4 application. IETF, RFC 4004.
Olzak, T. (2005). Wireless handheld device security. Retrieved August 8, 2006, from http://www.securitydocs. com/pdf/3188.PDF
Davis, C. (2001). IPSec: Securing VPNs. New York: McGraw-Hill.
Open Mobile Alliance. (2006). WAP forum. Retrieved August 8, 2006, from http://www.wapforum.org/
DECT Forum. (2006). Retrieved August 8, 2006, from http://www.dect.org
Perelson, S., & Botha, R. (2004, July). An investigation into access control for mobile devices. In H. S. Venter, J. H. P. Eloff, L. Labuschagne, & M. M. Eloff (Eds.), Proceedings of the ISSA 2004 Enabling Tomorrow Conference on Information Security, South Africa.
Dietze, C. (2005). The smart card in mobile communication: Enabler of next-generation (NG) services. In M. Pagani (Ed.), Mobile and wireless systems beyond 3G: Managing new business opportunities. Hershey, PA: IRM Press.
Perkins, C., & Calhoun, P. (2000). Mobile IPv4 challenge/ response extensions. IETF, RFC 3012. 847
S
Security Architectures of Mobile Computing
Pulkkis, G., Grahn, K., Karlsson, J., Martikainen, M., & Daniel, D. E. (2005). Recent developments in WLAN security. In M. Pagani (Ed.), Mobile and wireless systems beyond 3G: Managing new business opportunities. Hershey, PA: IRM Press. Puthenkulam, J., & Yin, H. (2005). 802.16e: A mobile broadband wireless standard. Broadband Wireless Division, Mobility Group, Intel Corporation. Retrieved August 8, 2006, from http://www.ewh.ieee.org/r6/scv/comsoc/0512.zip
Zao, J., Kent, S., Gahm, J., Troxel, G., Condell, M., Helinek, P., Yuan, N., & Castineyra, I. (1999). A public-key based secure Mobile IP. Wireless Networks, 5(5), 393-390. ZigBee Alliance. (2004, December 14). ZigBeeTM Specification v1.0. Retrieved August 8, 2006, from http://www. zigbee.org
KEY TERMS
Rankl, W., & Effing, W. (2003). Smart card handbook (3rd ed.). New York: John Wiley & Sons.
Bluetooth: A technology standard for wireless short distance communication.
Setec Portal. (2006). Retrieved August 8, 2006, from http:// www.setec.fi
DECT: A cellular system and a common standard for cordless telephony, messaging. and data transmission standardized by ETSI (European Telecommunications Standards Institute).
Steinberg, J., & Speed, T. (2005). SSL VPN: Understanding, evaluating and planning secure, Web-based remote access. Birmingham, UK: Packt Publishing. Sundaresan, H. (2003). OMAP platform security features. Retrieved August 8, 2006, from http://focus.ti.com/pdfs/ wtbu/omapplatformsecuritywp.pdf TM
Umar, A. (2004). Mobile computing and wireless communications. Middlesex, NJ: Nge Solutions. Wireless Universal Serial Bus Specification. (2005, May 12). Revision 1.0. Retrieved August 8, 2006, from http://www. usb.org/developers/wusb/docs/WUSBSpec_r10.pdf
848
Mobile IP: Mobile Internet protocol for IP number preservation of a mobile computer. USIM: A SIM used in 3G mobile telephone networks. WiMax: A technology standard for wireless broadband Internet access. ZigBeeTM: A low-cost, low-power communication standard for wireless data communication in home and building automation.
Category: Mobile Software Engineering 849
Semantic Caching in a Mobile Environment Say Ying Lim Monash University, Australia
INTRODUCTION
BACKGROUND
Mobile computing environments enable the database servers to disseminate data via wireless channels to multiple mobile clients (Chung & Kim, 2001). It has increased popularity with the emerging trend of wireless network and usage of handheld devices, such as PDAs and other portable electronic devices. The typical nature of a mobile environment would include low bandwidth and low reliability of wireless channels, which causes frequent disconnection to the mobile users. Hence due to the constraints of the nature of mobile environment it is important to enhance the performance of the query processing, as well as improve the availability of querying particular data items especially during disconnection (Imelinkski & Korth, 1996; Malladi & Davis, 2002). Often, mobile devices are associated with low memory storage and low power computation and with a limited power supply (Myers & Beigl, 2003). Hence, it is important to help mobile clients to save the usage of its battery. By introducing data caching into the mobile environment, it is believed to help improve data availability in case of disconnection by retrieving data that has been previously cached in the local memory and to be able to save power by having a lower data transmission. Generally, in data caching, it means the data is cached in the memory storage of the mobile device, and whenever the mobile users want to issue a query, it will first search its cache and if there exists a valid copy in the cache it returns the results immediately. Otherwise the mobile users would attempt to obtain the data from the server either using the server or broadcast strategy. Caching has emerged as a fundamental technique, especially in distributed systems, as it not only helps reduce communication costs but also off loads shared database servers. In this article, we describe the use of caching, which allows coping with the characteristic of the mobile environment. We concentrate particularly on semantic caching, which is basically a type of caching strategy that is contentbased reasoning ability with the ability to—in addition of caching query results—remember the queries that generated these results. Semantic caching provides accurate, semantic description of the content of the cache.
The effect of having the ability to cache data is of great importance, especially in the mobile computing environment than in other computing environments. This is due to the reason that contacting the remote servers for data is expensive in the wireless environment and, with the vulnerability to frequent disconnection, can further increase the communication costs (Leong & Si, 1997). There are many different types of caching strategies that serve the purpose to improve query response time and to reduce contention on narrow bandwidth (Zheng, Lee, & Lee, 2004). Caching mechanisms need to retain the frequently accessed data locally in the mobile device storage to be able to allow users to access server database queries at least partially in cases of disconnections. Hence, the more effective the caching mechanism is in keeping the frequently accessed data will result in more queries that can be served during disconnection. Due to limitations such as cache space, cache replacement and cache granularity, as well as cache coherence, are the three main issues that characterize caching mechanism. In traditional cache replacement, the most important factor affecting cache performance is the access probability. This refers to replacing the data with the least access probability to free up more cache space for the new data. There is a large variety of caching replacement policies and most of them utilize access probability as the primary factor in determining which data items are to be replaced. Cache granularity relates to determining a physical form of cached data items. It appears to be one of the key issues in caching management systems. There are three different levels of caching granularities in object-oriented databases, which includes: (a) attribute caching, (b) object caching and (c) hybrid caching (Chan, Si, & Leong, 1998). Attribute caching refers to frequently accessed attributes that are stored in the client’s local storage. As for object caching, instead of the attribute itself being cache, the object is cached. In attribute caching, it creates undesirable overheads due to the large number of independent cache attributes. Thus, hybrid caching, which appears to be a better approach, comprises of the combinations of both granularities. Cache coherence—or known as invalidation strategy—involves cache invalidation and update schemes to invalidate
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Semantic Caching in a Mobile Environment
Figure 1. Overlapping results from two queries issued
P1 P
2
P7 P5
P3
P6
User move from Location A towards Location B
Query 1
P4
and update out-dated or non valid cached items (Chan, Si, & Leong, 1998; Cao, 2003). After a certain period, a cached data may appear as no longer valid and therefore mobile users should obtain a newer cache before retrieving the data (Xu, Tang, & Lee, 2003). There are several techniques that have been proposed to overcome this issue. These include (a) stateful server, (b) stateless server, (Barbara & Imielinski, 1994) and (c) leases file caching mechanism (Lee, Leong, & Si, 2001). Stateful server refers to the server having an obligation to its clients, which means the server has the responsibility in notifying the users about changes, if there are any. In contrast, stateless server refers to the server not aware for its clients, whereby the server broadcasts a report that contains the updated item either asynchronously or synchronously. The leases files mechanism, which is also known as lazy invalidation approach, assigns each mobile user to be responsible for invalidating its cached items. Consequently, a good caching management strategy is needed to deal with the critical caching issues, such as caching replacement, caching granularity and caching coherence.
SEMANTIC CACHING A better way of query processing specifically for use in a mobile environment is by allowing the users to specify precisely what data items are missing from its local storage to server the query. This could be achieved by having the previously evaluated query results being cached (Dar et al., 1996; Roussopoulos, 1991).
Using Semantic Caching in a Mobile Environment A semantic cache is defined as consisting of a set of distinct semantic segments, which can be decomposed into separate components or come together as a whole of the query 850
P8
Query 2
results. A semantic segment S can be specified by having whereby SR and SA define the base relation of relation and attributes in the creation of the semantic segment respectively. SP is to indicate the criteria that S satisfies, and SC indicates the actual content of S, which is represented by pages. (Ren, Dunham, & Kumar, 2003) Semantic caching stores semantic descriptions and associated answers of the previous queries in the mobile client (Dar et al., 1996). The main feature of semantic caching is the content-based reasoning ability as well as the fact that only the required data, as opposed to a file or pages of data, is transmitted over the wireless channel. When a new query exists, the mobile client can determine whether should it be totally answered by how much can it be answered and what data are missing. With these abilities, the wireless traffic can be greatly reduced because only the needed data are transferred. This helps with disconnection too, since total or partial results may be obtained even when the server is unreachable (Lee, Heong, & Si, 1999). As a result, if a query can be partially answered from the cache, the volume of missing data requested from the server as well as the wireless bandwidth consumed can be reduced. And if the query could be answered completely based on the cache, then no communication between the client and the server is required at all. This ability is of particular significance during disconnection, which is the main constraint the mobile environment is currently facing. This also leads to reduction of overhead due to redundant computation as the amount of data transferred over the wireless channel can be substantially reduced. Example 1: Suppose a mobile user who is traveling from one location to another location suddenly wished to find a nearby rest place. So the user issue a query while he is in Location A and the server returns the nearest rest place which is P1, P2, P3, P4, P5. But the user is not satisfied with the results. So he re-issued another query while he is moving
Semantic Caching in a Mobile Environment
Figure 2. Issuing a 3NN query
S Query 2
P6
P1
P7
P2 P3
P5
P6
User move from Location A towards Location B P4
P8
Query 1
towards Location B. And the query returns another set of results which may contain some overlapping results such as the same P5 as the user previously received. Hence, the cached results are immediately returned since it has been previously cached. Thus, the user actually only needs to submit the complement of the new query in order to obtain only results that are not the same as the one previously obtained. This example can be illustrated as in Figure 1. This shows that semantic caching not only saves the wireless bandwidth due to less retransmission, but also reduces the query response time since some cached results can be immediately returned.
Benefits and Limitations There are several advantages that can be gained by using semantic caching, with the main reason that because only required data are being transferred communication cost between the client and the servers would be reduced. Moreover, cache space overhead is low for semantic caching since only the data that satisfy previous queries are being stored. With the ability of semantic caching to keep semantic information, it enables missing data to be exactly determined, which causes easy parallel query processing. Hence, semantic caching is very efficient to be used in the mobile environment since more autonomy is given to the clients and partial results can be derived when disconnections from the wireless channels occur (Ren, Dunham, & Kumar, 2003). Besides all the benefits semantic caching brings in, there are also limitations and drawbacks that semantic caching brings. Generally, semantic caching captures the semantics of the queries only, and ignores the semantics of the cached objects. Therefore, the granularity is at a query level that helps answering similar queries faster, but cached objects from different types of queries become difficult. (Hu et al.,
2005) In addition, the types of spatial queries supported by semantic caching are rather limited to simple range query and nearest neighbor (NN) query (Ren & Dunham, 2003; Zheng & Lee, 2001). It is difficult to support complex queries such as k-nearest-neighbor (kNN). Besides that, it also demonstrates complicated cache management. For example, when a new query to be cached overlaps some cached query, a decision has to be made whether to bring these two queries or to trim either of them. When the cache size grows, all these drawbacks would become more remarkable. Example 2: The same scenario as described in Example 1 but, instead of the user issuing a query 2 in a new location after the query 1 which has been sent, he would like to issue query 2 that is comprised of 3 nearest neighbor (3NN) query. Due to the limitation of semantic caching that is not able to trim a 3NN query from the first query, the user would have to sent a full complete query 2 to the server even though the results data P1, P3, P4 have been cached as a result of the first query that has been issued earlier on. They are actually partial results of the new query and should have been returned immediately to the user but are not able to do so. This example can be illustrated as in Figure 2. The retransmission of this result to the mobile users has wasted the wireless bandwidth as well as unnecessary transfer and has prolonged the response time (Hu et al., 2005).
Semantic Caching in Query Processing In order to process a query from a semantic cache, first of all we would check whether the query can be answered locally by the cache. If the answer can be obtained from the cache, the results are locally processed from the cache in the mobile device. However, in cases where the answers are not fully obtainable from the cache, but can only be partially 851
Semantic Caching in a Mobile Environment
Figure 3. Semantic caching query processing mechanisms
Probe Query
Remainder Query
Results
answered, we will trim the original query by either removing or annotating those parts that are answered and send it over the wireless channel to the server for further processing (Godfrey & Gryz, 1999). In other words, if the answer can be totally answered from the cache a probe query is being issued, whereas a remainder query is being issued to the server when only partial answers are obtainable. Figure 3 shows the semantic caching query processing mechanisms (Waluyo, Srinivasan, & Taniar, 2005). Example 3: Consider a mobile user who previously issued queries in getting movie information. This information had been cached into his local memory storage previously. Now he would like to send another query obtaining another list of movie information. So, it first processes the query locally via the semantic cache and determines if any of the semantic segments contribute to the query results from the cache index. If the result can be partially answered by the semantic segment then a probe query will retrieve the results that can be obtained locally, and a remainder query will be issued to the server for evaluation to define the other partial results. And then the result for the query is obtained by integrating all the partial results into a single result, which may consist of the results from probe query and remainder query. Due to the fact that mobile users in a typical mobile environment move around frequently by changing location has opened up a new challenge of answering queries that is dependent on the current geographical coordinates of the users (Barbara, 1999). This is known as location dependent queries (Park, Song, & Hwang, 2005). An example of a location dependent query can be, “Find the nearest restaurants from where I am standing now.” This is an example of static object whereby restaurants are not moving. An example of a dynamic object would be, “What is the nearest taxi that will pass by me?” Queries should be processed in a way that minimizes the consumption of bandwidth and battery power. The problem is challenging because the user location is changing and the results would also change accordingly (Shi, Li, & Wang, 2002). And with the keep on changing location that causes changing results to be downloaded would cause high com852
munication costs if excessive communication is needed to and from the server several times (Seydim, Dunham, & Kumar, 2001). Hence, caching plays a role in location dependent query processing. This allows the queries to be answered without connecting to the server. In summary, there are a series of steps that can be carried out in answering a query. First of all, when a query has been issued, the local cache is checked to see whether the results can be obtainable locally or not. If there are no suitable answers to the queries that are issued that correspond to the location of the user in cases of location dependent queries, then the information of the location of the user will be transmitted to the server to answer the query and returned back to the user. Otherwise, if there is some related data from the cache itself, then the data can be retrieved directly from the cache and the answer to the query has been completed. For a location dependent situation, by giving the current location of the user as well as the speed, the time when this user will move to another location can be computed and determined. However, before the whole answering process ends, after getting the results from the server, a new cache is inserted into the local cache memory for future use (Ren & Dunham, 2000).
FUTURE TRENDS There have been several researches done in the area of semantic caching in a mobile environment. The usage of semantic caching has obviously provoked extensive complicated issues. There are still many limitations of the nature of the mobile environment that generate a lot of attention from research in finding a good cache strategy that is specifically designed for use only in the mobile computing environment. In the future, it is critical to design algorithms for using semantic caching to cope with the low bandwidth of the wireless channel as well as the vulnerable disconnection problem. Applying semantic caching to several different scenarios of location dependent queries, including in a multiple cell environment and a cooperative strategy between multiple clients, is also beneficial. Caching management strategies, which focus on semantic mechanisms that are designed for real mobile queries that will utilize space more efficiently, are also needed in the future. Besides these, further investigation on other cache replacement policies as well as granularities issues, which exploit the semantics of the caching data in terms of size or access pattern, is desirable. Employing the semantic caching into a broadcasting environment, which reduces the size of a broadcast cycle and improves the tuning time, is necessary.
Semantic Caching in a Mobile Environment
CONCLUSION Although there are significant increases in the popularity of mobile computing, there are still several limitations that are inherent, be it the mobile device itself or the environment itself. These include limited battery power, storage, communication costs, and bandwidth problem. All these have become present challenges for researchers to address. In this article we have described issues of caching in a mobile environment with its advantages and disadvantages focusing mainly on semantic caching. We include adapting semantic caching in both location and non-location dependent queries. This article serves as a valuable starting point for those who wish to gain some introductory knowledge about the usefulness of caching, particularly semantic caching.
REFERENCES Barbara, D., & Imielinski, T. (1994, November). Sleepers and workaholics: Caching strategies in mobile environments. MOBIDATA: An Interactive Journal of Mobile Computing, 1(1). Chan, B. Y., Si, A., & Leong, H. V. (1998). Cache management for mobile databases: Design and evaluation. In Proceedings of the International Conference on Data Engineering (ICDE) (pp. 54-63). Chung, Y. D., & Kim, M. H. (2001). Effective data placement for wireless broadcast. Distributed and Parallel Databases, 9(2), 133-150. Dar, S., Franklin, M., Jonsson, B., Srivastava, D., & Tab M. (1996). Semantic data caching and replacement. In Proceedings of 22nd International Conference on Very Large Data Bases (pp. 330-341). Deshpande, P. M., & Ramasamy, K. (1998). Caching multidimensional queries using chunks. In Proceedings of the ACM SIGMOD Conference on Management of Data (pp. 259-270). Dong Jung, Y., You, H., Lee, J. W., & Kim, K. (2002). Broadcasting and caching policies for location dependent queries in urban areas. In Proceedings of the 2nd International Workshop on Mobile Commerce (pp. 54-60). Franklin, M., Carey, M., & Livny, M. (1992). Global memory management on client-server architecture. In Proceedings on International Conference on Very Large Databases (pp. 596-609). Ganguly, S., & Alonso, R. (1993). Query optimization in mobile environments. In Proceedings of Fifth Workshop on Foundation of Models and Languages for Data and Objects (pp. 1-17).
Godfrey, P., & Gryz, J. (1999). Answering queries by semantic caches. In Proceedings of Database and Expert Systems Applications (DEXA) (pp. 485-498).
Häkkilä, J., & Mäntyjärvi, J. (2005). Combining locationaware mobile phone applications and multimedia messaging. Journal of Mobile Multimedia, 1(1), 18-32. Hurson, A.R., & Jiao, Y. (2005). Data broadcasting in mobile environment. In D. Katsaros, A. Nanopoulos, & Y. Manolopaulos (Eds.), Wireless information highways (Chapter 4). Hershey, PA: IRM Press. Imielinski, T., & Badrinath, B. (1994). Mobile wireless computing: Challenges in data management. Communications of the ACM, 37(10), 18-28. Imielinski, T., Viswanathan, S., & Badrinath, B. R. (1994). Energy efficient indexing on air. In Proceedings of the ACM Sigmod Conference (pp. 25-36). Imielinski, T., Viswanathan, S., & Badrinath, B. R. (1997). Data on air: Organisation and access. IEEE Transactions on Knowledge and Data Engineering, 9(3), 353-371. Jayaputera, J., & Taniar, D. (2005). Data retrieval for location-dependent queries in a multi-cell wireless environment. Mobile Information Systems, 1(2), 91-108. Keller, A. M., & Basu J. (1996). A predicate-based caching scheme for client-server database architectures. The VLDB Journal, 5(2), 35-47. Kottkamp, H.-E., & Zukunft, O. (1998). Location-aware query processing in mobile database systems. In Proceedings of ACM Symposium on Applied Computing (pp. 416-423). Lee, K. C. K., Leong, H. V., & Si, A. (2000). A semantic broadcast scheme for a mobile environment based on dynamic chunking. In Proceedings of the IEEE International Conference on Distributed Computing Systems (ICDCS) (pp. 522-529). Lee, C. K. K., Leong, H. V., & Si, A. (2001). Adaptive semantic data broadcast in a mobile environment. In Proceedings of the 2001 ACM Symposium on Applied Computing (pp. 393-400). Lee, K. C. K., Leong, H. V., & Si, A. (2002). Semantic data access in an asymmetric mobile environment. In Proceedings of the 3rd Mobile Data Management (pp. 94-101). Lee, K. C. K., Leong, H. V., & Si, A. (1999, April). Semantic query caching in a mobile environment. ACM Mobile Computing and Communications Review, 3(2), 28-36.
853
S
Semantic Caching in a Mobile Environment
Lee, D. L., Hu, Q., & Lee, W. C. (1998). Indexing techniques for data broadcast on wireless channels. In Proceedings of the 5th Foundations of Data Organization (pp. 175-182). Lee, W. C., & Lee, D. L. (1996). Using signature techniques for information filtering in wireless and mobile environments. Journal on Distributed and Parallel Databases, 4(3), 205-227. Lee, G., Lo, S-C., & Chen, A. L. P. (2002). Data allocation on wireless broadcast channels for efficient query processing. IEEE Transactions on Computers, 51(10), 1237-1252. Leong, H. V., & Si, A. (1997). Database caching over the air-storage. The Computer Journal, 40(7), 401-415. Lee, D-L., Zhu, M., & Hu, H. (2005). When location-based services meet databases. Mobile Information Systems, 1(2), 81-90. Lee, D. K., Xu, J., Zheng, B., & Lee, W-C. (2002). Data management in location-dependent information services. IEEE Pervasive Computing, 2(3), 65-72. Liberatore, V. (2002). Multicast scheduling for list requests. In Proceedings of IEEE INFOCOM Conference (pp. 11291137). Malladi, R., & Davis, K. C. (2002). Applying multiple query optimization in mobile databases. In Proceedings of the 36th Hawaii International Conference on System Sciences (pp. 294 -303). Myers, B. A., & Beigl, M. (2003). Handheld computing. IEEE Computer Magazine, 36(9), 27-29. Park, K., Song, M., & Hwang C-S. (2004). An efficient data dissemination scheme for location dependent information services. In Proceedings of the First International Conference on Distributed Computing and Internet Technology (ICDCIT 2004) (Vol. 3347, pp. 96-105). Springer-Verlag. Ren, Q., Dunham, M. H., & Kumar, V. (2003). Semantic caching and query processing. IEEE Transactions on Knowledge and Data Engineering, 15(1), 192-210. Ren, Q., & Dunham, M. H. (2000). Using semantic caching to manage location dependent data in mobile computing. In Proceedings of the 6th International Conference on Mobile Computing and Networking (pp. 210-221). Triantafillou, P., Harpantidou, R., & Paterakis, M. (2001). High performance data broadcasting: A comprehensive systems ”Perspective.” In Proceedings of the 2nd International Conference on Mobile Data Management (MDM 2001) (pp. 79-90). Waluyo, A. B., Srinivasan, B., & Taniar, D. (2005). Research on location-dependent queries in mobile databases. 854
International Journal on Computer Systems: Science and Engineering, 20(3), 77-93. Waluyo, A. B., Srinivasan, B., & Taniar, D. (2005). Indexing schemes for multi-channel data broadcasting in mobile databases. International Journal of Wireless and Mobile Computing, 1(6). Xu, J., Hu, Q., Lee, D. L., & Lee W.-C. (2000). SAIU: An efficient cache replacement policy for wireless on-demand broadcasts. In Proceedings of the 9th International Conference on Information and Knowledge Management (pp. 46-53). Xu, J., Tang, X., & Lee D. L. (2003). Performance analysis of location-dependent cache invalidation schemes for mobile environments. IEEE Transactions on Knowledge and Data Engineering (TKDE), 15(2), 474-488. Xu, J., Zheng, B., Lee, W-C., & Lee, D. L. (2003). Energy efficient index for querying location- dependent data in mobile broadcast environments. In Proceedings of the 19th IEEE International Conference on Data Engineering (ICDE ‘03) (pp. 239-250). Xu, J., Hu, Q., Lee, W.-C., & Lee, D. L. (2004). Performance evaluation of an optimal cache replacement policy for wireless data dissemination. IEEE Transaction on Knowledge and Data Engineering (TKDE), 16(1), 125-139. Yajima, E., Hara, T., Tsukamoto, M., & Nishio, S. (2001). Scheduling and caching strategies for correlated data in push-based information systems. ACM SIGAPP Applied Computing Review, 9(1), 22-28. Zheng, B., Xu, J., & Lee, D. L. (2002, October). Cache invalidation and replacement strategies for location-dependent data in mobile environments. IEEE Transactions on Computers, 51(10), 1141-1153.
KEY TERMS Caching: Techniques of temporarily storing frequently accessed data designed to reduce network transfers and therefore increase speed of download Cache Management Strategy: A strategy that relates to how client manipulates and maintains the data that has been cached in an efficient and effective way. Location-Dependent Queries: A type of query whose results depend on the location of the issuer whereby when the client moves around the results may change accordingly. Mobile Computing: Mobile computing implies wireless transmission, which enables users to use a computing device while in transit anywhere, anytime.
Semantic Caching in a Mobile Environment
Mobile Query Processing: Database query is being sent by mobile users through wireless communication and being processed by a station server.
Semantic Caching: A type of caching technique that has content-based reasoning ability.
855
S
856 Category: Location and Context Awareness
Semantic Enrichment of Location-Based Services Vassileios Tsetsos University of Athens, Greece Christos Anagnostopoulos University of Athens, Greece Stathes Hadjiefthymiades University of Athens, Greece
INTRODUCTION
BACKGROUND
Location-based services (LBS) are considered the most popular mobile telecommunication services besides the traditional ones, for example, SMS and MMS. They are believed to constitute the killer applications for next generation mobile networks, since they enable adaptive location-driven content provision. Such services can be provided wherever the location of mobile users can be determined. Nowadays, there is a wide range of methods for estimating the location of users in both indoor (i.e., in-building areas) and outdoor environments (Schiller & Voisard, 2004). Outdoor LBS are more developed than their indoor counterparts due to the existence of positioning and topological information systems, GPS (global positioning system) and GIS (geographic information system) respectively. However, almost all known LBS provide their functionality irrespectively of the actual user context, which may consist of user’s location, physical capabilities, and/or cognitive status. Furthermore, most services ignore the semantic information of the spatial elements (e.g., stairs, elevators, and physical obstacles), other than the Euclidean distance. In this article, we describe issues related to the development of intelligent and human-centered LBS for indoor environments. We focus on the navigation service. Navigation is probably the most challenging LBS since it involves relatively complex algorithms and many cognitive processes (e.g., combining known paths for reaching unknown destinations, minimizing path length). With the proposed system, we try to incorporate intelligence to navigation services by enriching them with the semantics of users and navigation spaces. Such semantic information is represented and reasoned using state-of-the-art semantic Web technologies (Berners-Lee, Hendler, & Lassila, 2001).
LBS offer location-aware content provision. Apparently, a key enabler of LBS is the positioning infrastructure. As far as outdoor environments are concerned, the most commonly used positioning method is GPS, which provides spatial information with high accuracy and availability at low cost. On the other hand, there exist many alternative positioning solutions for indoor spaces, but with none of them having been standardized yet. Among these solutions are: WLAN (wireless local area network) triangulation, dead-reckoning techniques (implemented with accelerometers and digital compasses), RFID (radio frequency identification) tags, and infrared/ultrasound beacons. The authors in Hightower and Borriello (2001) provide an extensive survey of indoor positioning techniques. A basic assumption for developing our system is that we have an indoor positioning system at our disposal. Such system can locate users with “adequate” accuracy. This article deals with indoor navigation. Former indoor navigation research focused on robot navigation. As the positioning systems have matured, more effort has been put on developing indoor navigation services for pedestrians, such as museum guides aiding the sightseeing of tourists. An indicative system in this category is CyberGuide (Abowd, Atkeson, Hong, Long, Kooper, & Pinkerton, 1997). Another, more recent and more sophisticated, navigation system is Navio (Gartner, Frank, & Retscher, 2004). Navio aims at developing a route modeling ontology, which provides both outdoor and indoor routing instructions to humans by identifying and formally defining the criteria, the actions and the reference objects used by pedestrians in their reasoning for routes. However, Navio research emphasizes on location fusion (i.e., the aggregation of location information from multiple sensing elements) and user interfaces and, thus, does not contribute significantly to the issue of path selection. This latter issue is of utmost importance for human-centered LBS,
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Semantic Enrichment of Location-Based Services
but it is often ignored or handled in trivial ways. In general, such systems focus on the path presentation to users and on the hardware/positioning infrastructure used. Additionally, some systems have been developed for addressing the special needs of certain user categories, such as navigation for blind people. Such systems, however, lack a holistic approach to the navigation process. This means that their approaches are not considered general enough to address the whole range of potential application requirements. This drawback of existing solutions, as well as their deficiencies that will be identified in the following subsections, have motivated the present research in user- and space-modeling, path selection and navigation algorithms.
Navigation Algorithms Since navigation is a path-searching algorithmic problem, the decision on the path-searching algorithm used is vital for the quality of the provided service. Most of the existing navigation systems, either indoor or outdoor, make use of traditional shortest path algorithms (e.g., Dijkstra, A-star), thus, recognizing the minimization of Euclidian distance as the only objective in the path selection process. However, such approach overlooks the significance of other objectives more relevant to the context of the user. Hence, significant research on that topic has identified that pedestrian navigation needs more sophisticated and human-centered path-searching algorithms. Authors in Duckham and Kulik (2003) have proposed the “simplest path algorithm.” In this algorithm, the selected path is the one with the lowest possible complexity in navigation instructions. This work belongs to the category of approaches that introduce modifications of well-known graph routing algorithms like the aforementioned shortest path algorithms. A rather similar approach is discussed in Grum (2005), where the proposed navigation algorithm computes the “least risk path.” The term risk refers to the possibility of the user getting lost. The aforementioned algorithms, although providing more “intuitively-correct” paths than the conventional shortest paths algorithms, do not take into consideration the user semantics, as dictated by the modern design paradigm “Design for All” (European Institute for Design and Disability, 2005) (a.k.a., inclusive design). This paradigm promotes the design and implementation of services and products so that they can be used by any user, without any further adaptation. The implementation of such a paradigm, in the LBS domain, would lead to services that can be consumed (in an optimal way) by any user, regardless of her special characteristics.
have been proposed for spatial modeling with different data representations and expressiveness. Specifically, geometric models represent the navigation space using a certain coordination system and mainly support geometric queries (e.g., where is the nearest coffee machine?). On the other hand, symbolic models represent the navigation space through sets of symbols (i.e., names) and inter-symbol relationships capturing the topological semantics (e.g., part-of and overlaps spatial relations). Finally, hybrid models are combinations of the two former categories, aiming at maximizing the overall expressiveness of the spatial model. An interesting comparison of spatial models is presented in Leonhardt (1998). As far as indoor navigation is concerned, only a few researchers have proposed practical, yet expressive, models. To our opinion, the most important is presented in Hu and Lee (2004). It is a hybrid model, which represents the space as semantic hierarchies of “locations” and “exits” that also carry geometric information (e.g., coordinates). The use of semantics-based spatial models results in what has been called semantic location-based services. However, we claim that actual semantic LBS should not only exploit semantically enriched spatial models (symbolic or hybrid), but also take into consideration the navigation context (i.e., user context and instantiation of spatial model). Hence, we propose a refinement of the term semantic LBS, or better, human-centered LBS, so that it supports the following requirements: • • • •
Awareness of spatial semantics (e.g., hybrid model) Awareness of navigation context Adherence to the Design-for-All paradigm Reaction to dynamic user or space status changes
As will be shown in a following section, such services can be built with a knowledge-based system architecture. This architecture exploits knowledge representation methods to model the various components, and reasoning/inference techniques in order to implement the actual path selection process. The most popular and practical technology for representing models is the ontologies. Ontology is defined as “an explicit and formal specification of a shared conceptualization” (Studer, 1998, p. 185). In other words, it is a method for describing models of application domains that can be understood by machines. Knowledge reasoning is the process of inferring new implied knowledge from explicit knowledge assertions. Such reasoning can be based either on logic-based methods (i.e., resolution) or production rules (Brachman & Levesque, 2004).
Spatial Models and Ontologies
SEMANTIC INDOOR NAVIGATION
The quality of path-searching algorithms also depends on the spatial modeling of the navigation space. Many approaches
In this article, we propose a framework for human-centered semantic indoor navigation, which meets the aforementioned 857
S
Semantic Enrichment of Location-Based Services
User Model: UNO
Figure 1. OntoNav architecture
In order to describe user context (i.e., profile, capabilities, constraints and navigational preferences), we have developed a minimalistic ontology, named user navigation ontology (UNO). The concepts of such ontology represent user navigation classes. Hence, each user profile is classified into one or more navigation classes according to her characteristics. Indicative user classes are: HandicappedUser (users who cannot walk), BlindUser (users who cannot see), and LazyUser (users who always prefer elevators than stairs). Both INO and UNO have been modeled through the Web Ontology Language-OWL (McGuiness & Harmelen, 2004).
OntoNav Indoor Navigation Ontology Navigation Service
Indoor Positioning System
User Models (UNO)
Indoor Spatial Model
Path-Selection Rules Routing Algorithm
Path-Selection Rules requirements. Such framework, named OntoNav, is based on a novel combination of practical ontology-based knowledge representation and reasoning technologies, as well as Euclidian path-searching algorithms. The architecture of the implemented navigation system is illustrated in Figure 1 and can be decomposed to the following basic components:
Navigation Ontology: INO This spatial ontology, named indoor navigation ontology (INO), describes the basic spatial and structural concepts of indoor environments, as well as the relationships between them. Specifically, it provides a semantic spatial model for reasoning about the selected paths. An extract of the INO taxonomy is depicted in Figure 2, illustrating a hierarchy of path elements.
The path-selection process is performed through sets of production rules. The definition of such rules involves both the spatial semantics (expressed through INO) and the user semantics (expressed through UNO). The rules are applied to the INO instances in order to assert and infer which paths are considered accessible and appropriate for each user request. Moreover, the path-selection rules are further analyzed to physical navigation rules, perceptual navigation rules and navigation preferences. Actually, the physical navigation rules are applied first, in order to discard any paths that are not physically accessible by the user. The perceptual navigation rules are related to the user’s cognitive status (e.g., age, education). Finally, paths that are proposed regarding the user preferences (e.g., paths containing elevators) are identified after the application of the navigation preferences. The rules are described through the Semantic Web Rule Language (SWRL) (Horrocks et al., 2004).
Figure 2. An extract of Indoor Navigation Ontology
owl:Thing is-a Path_Element is-a Path_Point is-a Navigational_Point is-a Junction
858
is-a Turn_Point
is-a Elevator_Exit
is-a Room_Exit
is-a Building
Stairway_Exit
is-a Stairway
is-a is-a Floor
is-a
Room
Ramp
Horizontal_Passage is-a Corridor_Segment
is-a
is-a
Point_of_Interest
Path
is-a
Vertical_Passage is-a
Space
is-a
is-a
Exit
is-a
Description
Passage
is-a
is-a
is-a
is-a Door
Obstacle
Semantic Enrichment of Location-Based Services
Navigation Service This service can be defined as the interface between the system and its users. It accepts navigation requests and responds with the optimal path, if any. Path optimality depends on several factors, such as suitability for current user context and length of the selected paths.
the optimal path, in general. Thus, we compute k shortest paths, since path length is always important, although it may not always be the primary selection criterion. More details on the adopted algorithm can be found in Yen (1971).
Indoor Positioning System
The INO instances are created by a geometric representation of the indoor topology. Such geometric data may reside in a GIS as building blueprints and are transformed to actual spatial ontology instances. In our framework, it is assumed that such instances are available, (i.e., we do not deal with such data transformation issues).
OntoNav symbolically locates the users in the navigation space according to the spatial model described by INO. The positioning infrastructure may vary from infrared/ultrasound beacons to WLAN triangulation or ded-reckoning techniques. It is important to note that since the positioning accuracy may be more fine-grained than the location granularity, an approximation error may be introduced in the estimated location of the user. However, this error does not significantly affect the quality of navigation.
Routing Algorithm
Description of System Workflow
This algorithm is a central element of the framework and, in combination with the path-selection rules, is responsible for the determination of the optimal path between two given endpoints. The algorithm used in our system was a k-shortest paths searching algorithm. Similarly to the approach in Wu and Hartley (2004), we believe that such algorithm facilitates more flexible path selection by enabling us to incorporate additional path finding restrictions, imposed by the user profile. The main idea is that the shortest path may not be
The end-to-end functionality of the system is depicted by means of flowcharts in the following figures. Figure 3 depicts the system initialization process, in which the spatial ontology instances are loaded. Figure 4 illustrates the workflow that takes place upon a navigation request from a user. Initially, the user registers her profile to the system, and the destination she wishes to reach. Her current location is determined through the indoor positioning system. If the user has not registered to the system again, then her profile
Indoor Spatial Model
Figure 3. System initialization
Begin
Loading INO ontology
Loading UNO ontology
Initialization of Rules Engine
End
Figure 4. System workflow after a user navigation request has been received YES
Begin
End
Input User Profile
Input Navigation Origin and Destination
Output Shortest and Most Appropriate Path
Unknown User?
NO
Creation of User Instances
Retrieval of User Profile and relevant Data
Path Ranking
Instantiation of Navigation Service
Execution of k-shortest paths Algorithm
Computation of Shortest Paths’ Quality Score
859
S
Semantic Enrichment of Location-Based Services
Figure 5. The workflow for the creation of a new navigation service instance Begin
Exit
Input User Instances
Application of Navigation Rules to INO and UNO
Reasoning over User Profile
Persistence of Graph in User Profile
Marking of INO instances
Construction of Topology Graph from INO
is instantiated in the UNO ontology and a new Navigation Service instance is created. Such an instance mainly consists of the topology-graph associated with the specific user. It should be reminded that this graph is created from the INO instances, after the application of the physical navigation rules. Hence, the graph contains all the path elements that are accessible by that user. Such graph is the basic input to the next task, as illustrated in Figure 4, where the k-shortest paths between the user and the destination location are computed. Subsequently, for each path, a total quality score is calculated, denoting to which extend a path satisfies the perceptual navigation and preference rules. The path with the highest score is the final selected path that is proposed to the user. The task that refers to the creation of a new navigation service instance is very important, since it involves the most algorithmic and computational parts of the whole system functionality. In Figure 5, one can see that the UNO instances (i.e., user profile) are the main inputs to that task. Those instances pass through a reasoning engine, which results in the user classification with respect to the UNO classes. Subsequently, all types of rules are applied to the INO and UNO knowledge bases. Specifically, the physical navigation rules “mark” the INO instances for further exclusion from the process, whilst the perceptual navigation rules and navigation preferences reward or penalize certain INO instances pertaining to the specific user. The unmarked elements are used for the creation of the user-accessible topology graph, which is also stored in the user profile for future use.
ontologies). In our case, INO and UNO ontologies should have been standardized in some way, so that LBS providers could rely on their specifications in order to develop interoperable semantic navigation services. Moreover, ontological engineering is a very difficult and time-consuming process. Towards simplifying it, there is a great deal of past and current research in developing methodologies and tools for creating, managing, merging, and, updating ontologies (Gomez-Perez, Fernandez-Lopez, & Corcho, 2004). In addition, more and more research projects have commenced to design ontologies for specific application domains. Their attempts, if coordinated accordingly, can result in the creation of an extensible “ontology repository” usable by anyone. The proposed approach of designing navigation services directly involves human factors. Hence, another issue is the human evaluation of such systems, apart from their performance evaluation. Since semantic LBS are (according to their definition in this article) human-centered, user acceptance and quality assessment are considered as prerequisites before launching them in real-world systems. The parameters that affect the path-selection process should be adjusted carefully by real users prior to system deployment. In the future, we expect to see more research in user evaluation of innovative personalized mobile services. Pervasive computing research has already paved the way for such evaluation frameworks (Scholtz & Consolvo, 2004), but still much progress has to be achieved.
FUTURE TRENDS
This article discusses spatial modeling and processing issues regarding a human-centered navigation service. Such service is mainly targeted to people with navigational limitations, and pursues the vision of context-aware services for ubiquitous computing environments. The main goal of this work is to creatively integrate semantic knowledge engineering technologies with traditional location-based services. In our view, such integration is considered a key enabler for next-generation mobile services that focus on providing advanced user experience.
The introduction of knowledge engineering technologies in traditional services and applications is expected in the following years. This “paradigm shift” in system design and implementation is merely “pushed” by initiatives like inclusive design and universal access (Stephanidis & Savidis, 2001). However, there are still a lot of open issues before such approach can be massively adopted in commercial systems. One of the greatest challenges is the common definition and adoption of semantic application domain models (i.e., 860
CONCLUSION
Semantic Enrichment of Location-Based Services
ACKNOWLEDGMENTS This work has been partially funded by the Greek General Secretariat for Research and Technology (GSRT) under grant PENED2003 (No. 03ED173).
REFERENCES Abowd, G. D., Atkeson, C. G., Hong, J., Long, S., Kooper, R., & Pinkerton, M. (1997). Cyberguide: A mobile contextaware tour guide. Baltzer/ACM Wireless Networks, 3(5), 421-433. Berners-Lee, T., Hendler, J., & Lassila, O., (2001, May). The semantic Web. Scientific American, 284(5), 28-37.
Leonhardt, U. (1998). Supporting location-awareness in open distributed systems (PhD Thesis). London: Department of Computing, Imperial College. McGuinness, D. L., & Harmelen, F. (2004). OWL Web ontology language overview. World Wide Web Consortium Recommendation. Retrieved from http://www.w3.org/TR/ owl-features/ Schiller, J., & Voisard, A. (2004). Location-based services. San Francisco: Morgan Kauffman. Scholtz, J., & Consolvo, S. (2004). Toward a framework for evaluating ubiquitous computing applications. IEEE Pervasive Computing, 3(2), 82-88.
Brachman, R., & Levesque, H. (2004). Knowledge representation and reasoning. San Francisco: Morgan Kaufmann.
Stephanidis, C., & Savidis, A. (2001). Universal access in the information society: Methods, tools, and interaction technologies. Universal Access in the Information Society, 1(1), 40-55.
Duckham, M., & Kulik, L. (2003) Simplest paths: Automated route selection for navigation. In The Proceedings of COSIT 2003 (LNCS 2825, pp.169-185).
Studer, R., Benjamins, V. R., & Fensel, D. (1998). Knowledge engineering: Principles and methods. IEEE transactions on data and knowledge engineering, 25(1-2), 161-197.
European Institute for Design and Disability. (n.d.). Retrieved from http://www.design-for-all.org/.
Wu, Q., & Hartley, J. (2004). Using k-shortest paths algorithms to accommodate user preferences in the optimization of public transport travel. In Proceeding of UKSIM 2004 (pp. 113-117).
Gartner, G., Frank, A., & Retscher G. (2004). Pedestrian navigation system in mixed indoor/outdoor environment: The NAVIO project. CORP 2004 Geomultimedia04, Vienna, Austria. Gomez-Perez, A., Fernandez-Lopez, M., & Corcho, M. (2004). Ontological engineering: With examples from the areas of knowledge management, e-commerce and the semantic Web. London: Springer-Verlag. Grum, E. (2005). Danger of getting lost: Optimize a path to minimize risk. Tenth International Conference on Urban Planning & Regional Development in the Information Society (CORP), Vienna, Austria. Hightower, J., & Borriello, G. (2001). Location systems for ubiquitous computing, IEEE Computer, 34(8), 57-66. Horrocks, I., Patel-Schneider, P., Harold, B., Tabet, S., Grosof, B., & Dean, M. (2004). SWRL: A Semantic Web Rule Language combining OWL and RuleML. World Wide Web Consortium Member Submission. Retrieved from http://www.w3.org/Submission/SWRL/ Hu, H., & Lee, D.L. (2004). Semantic location modeling for location navigation in mobile environment. IEEE Mobile Data Management, 52-61.
Yen, J. (1971). Finding the k shortest loop-less paths in a network. Management Science, 17, 712-716.
KEY TERMS Design for All: A design approach that aims at constructing products and services in a way that no user is excluded from using them, independently of her capabilities or limitations. Indoor Positioning: The determination of an (moving) object’s location in an indoor environment. k Shortest Paths Problem: The identification of a set of paths {p1, …, pk} between two endpoints s and t (origin and destination, respectively) that satisfy the following criterion: length(pn-1) ≤ length(pn), for every n ≤ k, and p1 is the shortest path between s and t, as computed by a traditional shortest path search algorithm (e.g., Dijkstra). Navigation: Q service that finds a path between two locations (origin and destination) and gives appropriate instructions in order for the user to successfully follow it and reach the desired destination.
861
S
Semantic Enrichment of Location-Based Services
Ontology: A model representing the main entities and their relationships within a domain of discourse. Although similar to other modeling formalisms, ontologies (and especially those based on subsets of logic) can be more expressive and can represent complex restrictions and axioms that govern the domain entities. Reasoning: The computational procedure that infers new knowledge from explicitly asserted knowledge (expressed through statements and/or rules). Such new knowledge may
862
be of the form of new statement assertions or may just give extra information about the consistency and validity of the existing statements. Semantic Web: the evolution of the current WWW in a way that it is also machine-understandable in addition to being human-understandable. This evolution is based on the annotation of data with explicit semantics (i.e., metadata), which can describe purpose and attributes and classify it according to some knowledge models (i.e., ontologies).
Category: Location and Context Awareness 863
Sensor Data Fusion for Location Awareness Odysseas Sekkas University of Athens, Greece Stathes Hadjiefthymiades University of Athens, Greece Evangelos Zervas Tei-Athens, Greece
INTRODUCTION
BACKROUND
In pervasive computing environments, location is essential information as it is an important part of the user’s context. Applications can exploit this information for adapting their behavior. Such applications are termed location-aware applications (e.g., friend-finder, asset tracking). The location of a user is derived by various positioning methods. Especially for indoor positioning, different approaches have been proposed. The majority of indoor positioning systems rely on different technologies, usually of the same kind, like wireless LAN signal strength measurements (Bahl & Padmanabhan, 2000), IR beacons (Sonnenblick, 1998), or ultrasonic signals. At this point we will quote the definitions of accuracy and precision, the most important characteristics of a positioning system:
Indoor positioning systems have been an active research area since the Active Badge Project (Want, Hopper, Falcao, & Gibbons, 1992). Since then, several indoor location systems have been proposed. A large number of them use IEEE 802.11 (Wi-Fi) access points to estimate location. RADAR (Bahl & Padmanabhan, 2000) is a radio-frequency-based system for locating users inside buildings. It operates by recording and processing received signal strength (RSS) information. The RSS method is used also by the commercial system Ekahau (Ekahau Positioning Engine). The Cricket Location Support System (Nissanka, Priyantha, & Balakrishnan, 2000) and Active Bat location system (Harter, Hopper, Steggles, Ward, & Webster, 1999) are two systems that use the ultrasonic technology. Such systems use an ultrasound time-of-flight measurement technique to provide location information. They provide accurate location information, but also have several drawbacks like poor scaling and a high installation and maintenance cost. For these reasons they are rather inaccessible to the majority of users. Another category of location systems uses multiple sensor readings (Wi-Fi access points, RFIDs) and sensor fusion techniques to estimate the location of a user. Location Stack (Graumann, Lara, Hightower, & Borriello, 2003) employs such techniques to fuse readings from multiple sensors. Another similar approach is described in King, Kopf, and Effelsberg (2005). The drawback of these systems is their inability of supporting mobile devices with limited capabilities (CPU, memory) as the location estimation is performed at the client side; hence devices incur the cost of complex computations. The location estimation system described in this article relies on data from sensors to determine the location of a user. Our work differs from previous approaches in various aspects. Firstly, we use dynamic Bayesian networks for location inference. By using DBNs, we obtain better location estimation results. Along with heterogeneous sensor data that are processed in real time, we can also “fuse”
• •
Accuracy denotes the distance within which the system has the ability to locate a user, for example, 1-10 meters. Precision denotes the percentage of time the system provides a specific accuracy, for example, 80% of the time the system provides accuracy 1-5 meters (or else accuracy less than 5 meters).
The accuracy and precision are tradable, and it is clear that if we need less accuracy, the precision that the system provides increases. During the past few years, several location systems have been proposed that use multiple technologies simultaneously in order to locate a user. One such system is described in this article. It relies on multiple sensor readings from Wi-Fi access points, IR beacons, RFID tags, and so forth to estimate the location of a user. This technique is known better as sensor information fusion, which aims to improve accuracy and precision by integrating heterogeneous sensor observations. The proposed location system uses a fusion engine that is based on dynamic Bayesian networks (DBNs), thus substantially improving the accuracy and precision.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Sensor Data Fusion for Location Awareness
past information about the user. Secondly, our system can support a variety of mobile devices (PDAs, palmtops) with low computing power. Location estimation takes place in a server residing in the fixed network infrastructure. Mobile devices are just transmitting observations from sensors to this server and receive the location estimations. Finally, the adopted system architecture has the advantage of easy management and scalability (e.g., the installation of a new access point is completely transparent to users).
sonic signals are called transducers and are commonly used for distance measuring. In general, they integrate a sensor that can receive or transmit an ultrasonic signal and another RF transmitter/receiver which is used for synchronization. All the previously mentioned devices (elements) of different technologies (access points, beacons, tags, etc.) can be found in indoor environments either deployed in the building or attached to mobile devices. Some of them emit information, and others detect (read) information. According to their position and functionality, the elements can be categorized as follows:
POSITIONING TECHNOLOGIES
•
In this section we present the principal technologies that are used for indoor positioning and describe their characteristics. We also discuss a categorization of the devices. The most important wireless LAN standard today is the IEEE 802.11 (Wi-Fi) that operates in the 2.4 GHz ISM band or 5GHz band. This technology is used by several positioning systems that measure the signal strength from access points (RSS) to locate a user. Radio frequency identification (RFID) is the technology used for security tags in shops, ID cards, and so forth. Tags are powered by the magnetic field generated by a reader and transmit their ID or other information. Such tags do not require any battery and can be deployed in a building to detect object and person passing or proximity. Infrared (IR) beacons are programmable devices that periodically emit their unique ID in the IR spectrum. Usually the range of these beacons is approximately 10-20 meters, and the infrared receiver should have line of sight with the beacon in order to receive its ID. Ultrasonic signals are vibrations at a frequency greater than 20 kHz. The devices used to receive and transmit ultra-
Portable elements are those carried by users or attached to their mobile devices (RFID tags, Wi-Fi adapters) Infrastructure elements are those attached to the building (Wi-Fi access points, IR beacons, RFID tag readers). Active elements (sensors) are those which detect a phenomenon or take measurements (RFID tag readers, Wi-Fi adapters). Passive elements are those that emit information which is detected by active elements. Wi-Fi access points, IR beacons, and so forth fall into this category.
• • •
SYSTEM ARCHITECTURE The architecture of the proposed location estimation system is organized into three layers: the sensor layer, the collection layer, and the fusion layer. Figure 1 illustrates the generic architecture of the proposed system. In the same figure are also depicted location-aware applications which exploit the location information and databases where the personal profile of users and historical data about their behavior are stored.
Sensing layer
Collection layer
Fusion layer
Locationaware application
Figure 1. Architecture of the indoor location estimation system
864
AP 1
...
AP2
APK
Historical data
Fusion engine
... ...
Collectors
Sensors
Sensor Data Fusion for Location Awareness
The layered approach aims to facilitate effortless inclusion of new elements in order to improve the accuracy and the precision offered by the system. In the following paragraphs we provide a more detailed presentation of each layer.
Sensing Layer This is the lowest layer of the architecture, comprising sensors of different technologies. Sensors are attached either to the user’s mobile device (portable active elements) or to the building (infrastructure active elements). Below, we briefly discuss these two categories.
Portable Sensors A Wi-Fi adapter can measure the received signal strength (RSS) from a Wi-Fi access point (passive infrastructure element). Similarly, the IR port of a handheld device or a laptop is used as a reader for infrared transmissions from IR beacons that are wall-mounted.
Infrastructure Sensors RFID tag readers belong in this category. Such readers detect an RFID tag and read its ID when the latter is in proximity. Users can carry RFID tags which have unique IDs. Furthermore, ultrasonic devices, which estimate the distance of a user from a known point, also belong in this category.
Collection Layer This layer consists of software components called collectors. The role of a collector is to interact with the appropriate sensor and collect measurements or events. Sensors may produce raw data in a variety of formats according to their type. Hence, the output of a Wi-Fi adapter is a stream consisting of RSS measurements from access points; IR and RFID readers generate a stream of proximity events. When such raw data arrive at the collection layer, a preprocessing procedure is performed as described next.
An IR beacon collector, during this preprocessing procedure, operates differently. The two possible states of an IR beacon are: Visible and Not_Visible. Assume that an IR receiver (portable active element) is in the range of the IR beacon with ID IRB3 (infrastructure passive element). This situation will cause a proximity event which will be detected, and thus the collector sets the IRB3 to the value “Visible.” An RFID tag reader collector’s functionality is similar to the IR beacon collector’s, as RFID tag readers detect proximity events, too.
Tuple Forming After the preprocessing of raw data from the sensing layer, each collector forms a tuple of the type: (user_ID, IE_ID, value) where user_ID is the unique identifier of a user, IE_ID is the unique identifier of an infrastructure element, and value is a measurement or an event. A Wi-Fi collector may form the following tuple: (userA, AP1, S1) which denotes that the Wi-Fi adapter (portable active element) of the mobile device of userA measures the RSS from access point AP1 (infrastructure passive element), and the (quantized) RSS has value S1. A possible tuple generated by a RFID tag reader collector would be: (userB, RFR1, Visible) which denotes that an RFID tag (portable passive element) worn by userB (or attached to his/her mobile device) is in proximity of RFID tag reader with ID RFR1. These tuples are then forwarded to the upper layer where a location estimation procedure is invoked for each user. In the next section, we will show how such values are exploited for location estimations.
Preprocessing of Raw Data
Fusion Layer
Assume that a new RSS measurement arrives from a Wi-Fi adapter. Then, the appropriate collector (Wi-Fi collector) quantizes this on N discrete levels (values): S1, S2…SN. If, for example, the value from the access point with ID AP2 is between -70 dBm and -60 dBm, the value “S1” is assigned to this infrastructure passive element. It should be noted that the number N of quantization levels depends mainly on the thresholds (lower and higher) of the access point’s transmitted power and the environmental conditions (noise, etc.).
As mentioned in the introduction, the fusion engine is based on a dynamic Bayesian network, which is used for location inference. Below, we briefly discuss the basic concepts of Bayesian and dynamic Bayesian networks and discuss the adoption of DBNs in the proposed system. We assume that the reader is familiar with the theory of Bayesian and dynamic Bayesian networks. For a more complete introduction, the reader is referred to Jensen (1996) and Mihajlovic and Petkovic (2001).
865
S
Sensor Data Fusion for Location Awareness
Figure 2. (a) A Bayesian network (BN) showing four random variables and their dependencies. (b) A dynamic Bayesian network (DBN) showing dependencies between variables in different time-slots.
A
B
t-1
t
t+1
A
A
A
B
C
C
B
D
D
C
B
C
D
(a)
D
(b)
Figure 3. DBN for location estimation representing the dependencies between random variables at different time slots
IE1
t-1
t
t+1
L
L
L
IE2
...
IEN
IE1
IE2
...
IEN
IE1
IE2
...
IEN
Bayesian and Dynamic Bayesian Networks
DBN Integration in the Location System
Bayesian networks (BNs) present a statistical tool that has become popular in the areas of machine learning. They are well suited for inference because of their ability to model causal influence (cause-effect) between random variables. A BN (Figure 2(a)) consists of two parts. The first part is a directed acyclic graph (DAG), representing random variables as nodes and relationships between variables as arcs between the nodes. If there is an arc from a node A to a node B, it is considered that B is directly affected by A (A is the parent of B). Each node is conditionally independent from any other node given its parents. The second part of a BN is a probability distribution associated with each graph node. This describes the probability of all possible outcomes of the variable given all possible values of its parents. The parameters of this probability distribution would be estimated using observed data (Heckerman, 1995). The DAG and probability distributions together define the joint probability distribution. A DBN extends the static BN by modeling changes of stochastic variables over time. Random variables in a DBN are also affected by variables from previous time slots (see Figure 2(b)). For simplicity, it is assumed that the parents of a node are in the same or in the previous time slot (First Order Markov Chain).
The DBN that is used in our location estimation system is depicted in Figure 3. The random variables are:
866
• •
the location L of the user, which may take values from a set of K locations {L1, L2,…LK}; the N infrastructure elements IE1, IE2,…IEN. The range of values of those random variables depends on the type of element. Hence, an access point may take a value from the set {S1, S2,…} and an RFID tag reader from the set {Not_Visible, Visible}.as
The random variable L at time t, L(t), is directly affected by the random variable L at time t-1, L(t–1), so L(t–1) is the cause and L(t) is the effect. This is a reasonable assumption as the location of a user is dependent on his/her previous location. Also, the infrastructure elements at time t are affected by location at time t, L(t); the location of the user affects the value of an infrastructure element (e.g., the signal strength measured from a Wi-Fi access point depends on the location of the user). The probability distributions that are associated with each node of the DBN are estimated with Bayesian Network learning techniques. In particular, for every infrastructure
Sensor Data Fusion for Location Awareness
element (IE1, IE2,…IEN), we estimate the probability distribution P(IEi | L). This can be achieved by taking into account the fixed positions of infrastructure elements, the indoor propagation models of RF and IR signals, the time of flight of ultrasonic signals, and so forth. A simpler technique of learning that that we have adopted for our system is the method of sampling (signal, events) at every location for determining the values of infrastructure elements and the frequency of appearance of these values. According to this frequency we are able to form the probability distributions. In Table 1 we present a probability distribution of a passive infrastructure element (Wi-Fi access point) with ID AP1. Furthermore, the probability distributions P(Lt | Lt–1) for location transition can be generated according to the structure of the building, the distance between two locations, and the time required by a mobile user to cover this distance (see Table 2). It is important to note here that the determination of probability distributions takes place once, at system initialization (training phase).
Location Inference Queries After having structured the DBN of the fusion engine, we can use it for location estimations. A location inference query might be: “Where is user X given his/her previous location and given the values (observations) of infrastructure elements associated with this user?” To answer this, we calculate for each of the K locations {L1, L2,…LK } the following conditional probability: (t )
( t −1)
P( L | L
(1)
(t )
,O )
{IE1(t ) , IE 2(t ) ,...IE N(t ) } = O (t )
(2)
Equation (1) can be converted to the following equation: P( L(t ) | L(t −1) , O (t ) ) =
P( L( t ) , L( t −1) , O (t ) ) P( L( t −1) , O (t ) )
(3)
Taking into consideration that each node of our DBN is conditionally independent from any other node given its parents, we can compute the joint probability that appears in the numerator of (3). Also, as the denominator of (3) does not depend on the random variable L(t), it can be treated as a normalizing constant. Hence, the following equation is derived: P( L(t ) | L(t −1) , O (t ) ) =
P( L(t ) | L(t −1) ) * P(O (t ) | L(t ) ) K
∑ P( L
(t ) i
i =1
( t −1)
|L
(t )
(t ) i
) * P(O | L )
.
(4)
The probability distributions P(IEi | L) and P(L(t) | L(t–1)) are known from the training phase, so we can now compute the probabilities for each location {L1, L2,…LK}. Whenever a probability on the numerator of (4) is equal to zero, it is unnecessary to compute the final probability, as it is equal to zero, too. Thus, the calculations are pruned and the overall computation is optimized. The problem of location estimation is to find the location Li that maximizes the probability. max{P( L(i t ) | L( t −1) , O (t ) )}
(5)
which is the mathematical representation of the location inference query and denotes the probability of being at location L(t) at time t (the requested location) given the already known value of the previous location L(t–1) and given the values of the N infrastructure elements at time t, O(t). For simplicity reasons we write:
The location with maximum probability is stored in the database and the profile of the user is updated. Moreover, the location information is forwarded to location-aware applications for the provision of LBS services. After that the system proceeds to the next location estimation in light of the current observations and the previous estimated location of the user.
Table 1. A possible probability distribution for access point with identifier AP1. It can be shown that the probability P(AP1=S2 | L=L1 ) = 0.3.
Table 2. A possible probability distribution for location transition. The probability of transition from L1 to L2 is P(Lt=L2 | Lt-1=L1)=0.1.
L1
L2
…
S1
0.5
0.0
…
S2
0.3
0.8
…
…
…
…
…
L1
L2
…
L1
0.5
0.0
…
L2
0.1
0.8
…
…
…
…
…
867
S
Sensor Data Fusion for Location Awareness
FUTURE TRENDS Currently we are working on two issues which will have a direct impact on a system’s performance and scalability. The first issue is the adoption of a distributed architecture for the system. In this distributed architecture, the building is divided into regions. For each region there is one server responsible for location estimations. Servers of adjacent regions are interconnected in order to interchange information about the users (handovers between regions, etc). The distributed approach of the system will enhance its performance, improve its scalability, and make it more robust in case of server failures. The second issue that we are working on is the use of “dead reckoning” techniques to improve the precision and accuracy that the system provides. A user’s mobile device, which is equipped with an electronic compass and an accelerometer, could provide information about the direction and speed of its owner. Taking also into account the last known position of the user and the time elapsed since then, we can predict the current position and make more accurate estimations.
CONCLUSION In this article we presented a layered fusion system architecture that exploits information from sensors of different technologies to estimate the location of a user. A key difference from similar systems is the use of dynamic Bayesian networks for location inference. The use of DBNs improves our estimations. Along with sensor information, we take into consideration the previous location of the user thus improving the performance. Additionally, the system supports a variety of mobile devices, including those with restricted computational capabilities (PDAs, etc.) as they do not incur the burden of complex location calculations.
ACKNOWLEDGMENTS This work was performed in the context of the “PENED” Program, co-funded by the European Union and the Hellenic Ministry of Development, General Secretariat for Research and Technology (Research Grant 03ED173).
REFERENCES Bahl, P., & Padmanabhan, V. (2000). RADAR: An in-building RF-based user location and tracking system. Proceedings of IEEE INFOCOM (pp. 775-784). Tel-Aviv, Israel.
868
Ekahau. (n.d.). Ekahau positioning engine. Retrieved from http://www.ekahau.com/ Graumann, D., Lara, W., Hightower, J., & Borriello, G. (2003). Real-world implementation of the location stack: The universal location framework. Proceedings of the 5th IEEE Workshop on Mobile Computing Systems & Applications (WMCSA 2003) (pp. 122-128). Harter, A., Hopper, A., Steggles, P., Ward, A., & Webster, P. (1999). The anatomy of a context-aware application. Proceedings of the 5th Annual ACM/IEEE International Conference on Mobile Computing and Networking (Mobicom ’99). Heckerman, D. (1995). A tutorial on learning with Bayesian networks. Technical Report MSR-TR-95-06, Microsoft Research, USA. Jensen, F. (1996). An introduction to Bayesian networks. New York: Springer-Verlag. King, T., Kopf, S., & Effelsberg, W. (2005). A location system based on sensor fusion: Research areas and software architecture. Proceedings of the 2nd GI/ITG KuVS Fachgespräch “Ortsbezogene Anwendungen und Dienste,” Stuttgart, Germany. Mihajlovic, V., & Petkovic, M. (2001). Dynamic Bayesian networks: A state of the art. Technical Report, Center for Telematics and Information Technology, University of Twente, The Netherlands. Nissanka, B., Priyantha, A., & Balakrishnan, H. (2000). The cricket location-support system. Proceedings of MOBICOM 2000 (pp. 32-43). Boston: ACM Press. Sonnenblick, Y. (1998). An indoor navigation system for blind individuals. In CSUN Center On Disabilities (Ed.), Proceedings of the CSUN 1998 Conference, Los Angeles, CA. Want, R., Hopper, A., Falcao, V., & Gibbons, J. (1992). The Active Badge location system. ACM Transactions on Information Systems, 10, 91-102.
KEY TERMS Active Element (Sensor): One of the elements that detect a phenomenon or take measurements, like RFID tag readers and Wi-Fi adapters. Bayesian Network (BN): A directed acyclic graph where nodes represent random (stochastic) variables, and arcs represent dependence relations among these variables. Data Fusion: The combination of data derived from heterogeneous sources such that the resulting information is better than it would be if these sources were used individually.
Sensor Data Fusion for Location Awareness
Dynamic Bayesian Network (DBN): The extension of static Bayesian network by modeling changes of stochastic variables over time.
Pervasive Computing: The integration of computation into the environment in order to offer a broad range of services to users.
Infrastructure Element: One of the elements that are attached to the building, like Wi-Fi access points, IR beacons, and so forth.
Positioning: The capability to detect the location of a wireless device carried by a user (e.g., cell phone).
869
S
870 Category: Service Computing
Service Delivery Platforms in Mobile Convergence Christopher J. Pavlovski IBM Corporation, Australia Laurence Plant IBM Corporation, Australia
INTRODUCTION The demand for enriched multimedia content and entertainment services in mobile networks is being largely driven by the emergence of mobile broadband. A key problem for institutions attempting to capitalize on these new channels for service delivery is a capability to deploy many multimedia services rapidly and cost effectively. Traditional approaches in deploying new services have largely focused on discrete systems for each new service, often termed point solutions or silos. Recent emerging standards coupled with implementation constraints have led to the development of a more strategic approach. Such an approach involves the creation of a service delivery platform (SDP), capable of delivering a broad range of content and services from a host of multimedia applications. Several initiatives are attempting to lay the foundations for the architecture and framework for SDP solutions that support the emerging multimedia and entertainment services for mobile devices. Recent initiatives include platforms based upon the IP multimedia system (IMS), Parlay X, and IT standards-based designs. A central characteristic of the SDP approach to mobile service delivery is the capability to supply numerous services to mobile users with observed reductions in elapsed effort to bring these services online; this also bestows cost reduction and speed to market. The benefits of a service delivery business model are applicable to the mobile operator, mobile customers, and external third-party developers. Customers are offered more services quickly, while third-party developers are able to focus on core capabilities of their intended service, collectively offering benefits in terms of time to market and reduced cost. The business benefits illustrate why this service delivery approach is recently gaining increased attention by mobile operators globally. In this article we outline the fundamental principles of a service delivery platform and the business model to be addressed in mobile convergence. The emerging standards and reference architectures are presented, and their shortcomings are discussed. We outline the key requirements that a service delivery platform is expected to address from an operator perspective and summarize the key technol-
ogy design points. We also outline several future trends in how this emerging mobile technology is being deployed in various application scenarios. This involves straightforward mobile news services, through gaming, and complex interactive multimedia scenarios for the mobile device. These new services make further convergent demands upon three technology domains: the mobile network, IT systems, and the content/media sources.
BACKGROUND Traditional approaches to deploying new multimedia services involve development of a discrete system to deliver one or a related set of services. This approach involves the development of several common functions required by the multimedia service. A recent trend by many operators globally is deployment of one common service delivery platform that supports multiple applications. The SDP is intended to contain all the common functions and services that a wide range of applications may require in order to deliver its service or function. Figure 1 illustrates the change in design philosophy (Pavlovski & Staes-Polet, 2005), where traditional deployment involves development of common delivery functions for each (or a related set of) multimedia service(s). The SDP approach transforms this by combining the common functions used by multimedia services into one platform that may be exploited by a range of multimedia applications. Generally, the term convergence, when used in the context of mobile networks, is used to denote the rationalization of internetworking technologies and protocols. An additional form of convergence is related to the convergence of several operational and business domainsmore specifically, the convergence of information technology systems, the networks, and media/content applications. Such integration imposes additional complexity, and the service delivery platform is ideally suited to address this type of convergence. This notion of convergence and applicability of the service delivery platform is recently gaining widespread attention, with several aspects of these emerging service delivery
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Service Delivery Platforms in Mobile Convergence
Figure 1. Transformation to service delivery platform
Unique logic of multimedia service Common services Portal Services of multimedia application Infrastructure support functions Network & IT Systems Integration
Single system deployed for one, or several related multimedia
platforms actively studied within research and industry (Hanrahan, 2006; Deckers, 2006; Kimbler, Stromberg, & Dyst Appium, 2006). Mobile devices such as cellular phones, portable digital assistants, and tablets are becoming increasingly adorned with new services and media format. The fundamental business problem is to successfully integrate the network, information technology, and content applications in a unified manner that ameliorates costs for mobile operators, while supporting rapid deployment of new services in a cost-competitive manner.
SERVICE DELIVERY PLATFORMS The method used to construct service delivery platforms is based on either a network-centric or an information technology centric (IT-centric) view of the problem domain. Platforms based on the Parlay X or IMS standards and frameworks may be categorized as network-centric; there are several notable examples (Pailer, Stadler, & Miladinovic, 2003; Akkawi, Schaller, Wellnitz, & Wolf, 2004, Magedanz, Witaszek, & Knuttel, 2005; Hanrahan, 2006). In contrast, several broader attempts have been discussed in the literature that apply an IT-centric design (Pavlovski & Staes-Polet, 2005; Hwang, Park, & Jung, 2004; Pavlovski, 2002b). These platform design styles largely reflect the heritage of the originating body. Moreover, IMS and Parlay X have emerged from network standards bodies, while the other works reflect practical IT experiences in building service delivery platforms for mobile operators. Regardless of the approach taken, the underlying business model remains the same. In this section we first outline this business case, and then present the two strategies in addressing the service delivery model, outlining the benefits and shortcomings of each design viewpoint.
S Multimedia Applications Web Services Portal Services Infrastructure Support functions Network & IT Systems Integration
SDP platform supporting multiple multimedia
The Service Delivery Business Model The principle advantage of an SDP for mobile operators is the ability to reduce the cost for deploying new services, while increasing the potential for generating revenue (Deckers, 2006). It is suggested that around one-third of the effort required to introduce a new service is attributed to developing the unique business logic of the service (Pavlovski & StaesPolet, 2005). This means significant capability is common to a range of services and may be placed into a consolidated platform (see Figure 1). By combining this common functionality within the service delivery platform, the cost for building the platform is amortized since this may now be leveraged by several applications. Since the SDP is deployed by a mobile operator, external third parties are then able to develop applications that deliver mobile content. The third-party developer may now focus on building the unique business logic associated with the intended multimedia service or content, and is able to reduce the costs by leveraging a set of common capabilities within the service delivery platform. The service delivery business model for mobile applications has its origins with the iMode service (Pavlovski, 2002a), where the platform enables third-party application providers to deliver their content or service, via an iMode service platform, to mobile users. While the technologies have changed considerably over time, the service delivery business model has largely remained constant. The elementary model contains three entities: the mobile network operator, mobile customers, and external third-party developers. The mobile operator (alternatively this may be an MVNO) hosts the SDP and provides a set of common services to the external thirdparty developer from which to build multimedia applications. The mobile operator owns and maintains the relationship with mobile customers and is able to bring this consumer market to the external third-party developer. A key benefit 871
Service Delivery Platforms in Mobile Convergence
Figure 2. IT-centric service delivery platform 3rd Party Applications
Network Channels
Internet
Channel Gateways HTTP Server WiFi Gateway
Wireless network
Access Layer
Master Authentication Server
WAP Gateway
Consumer Devices
Parlay Messaging
Firewall AAA Store
Revenue Collection
Service Catalog
Billing Systems
Payment Gateways
Settlement Systems
Streaming, Gaming, SIP CRM
Location
of this relationship is that once the platform is constructed, the mobile operator is able to rely on external third parties to build, at their expense, new services. The principle motivation for the external developer, which may be an enterprise, is marketing access to the large customer base owned by the mobile operator. The advantages for the mobile customer is ease of access, a greater range of multimedia and content services, with the payment mechanisms for such access generally viewed as secure (Devine, 2001).
IT-Centric Design Viewpoint From an IT perspective the key principle is to abstract a range of network and IT services in a way that enables external third-party applications to be developed in a straightforward manner (Pavlovski, 2002a). Such an environment is sometimes referred to as a service creation environment (Schulke et al., 2005). Figure 2 illustrates the topology of an IT-centric SDP. A layered architecture is a key theme of this design, with each layer abstracting the services provided on the preceding layer. This allows the technology to alter without impact to all components within the platform. At the outer most level, the customer and third-party applications reside, external to the platform; note that additional internal applications may also be deployed by the mobile operator. Customers gain access to services provided by multimedia applications via the SDP. The network channel layer defines the various networks used to deliver content and services to consumer devices. The remaining components of an SDP are deployed within the channel gateway, access layer, and core services layer. These layers are now described in more detail. 872
Legacy IT
Web Service Gateway
SDP Portals
Telephony Server 3G mobile network
Core Services
Repositories
Channel gateways include those nodes responsible for integration and accessing various mobile and fixed networks. This includes the parlay gateway, messaging gateways (i.e., SMSc and MMSc), WAP gateways, and location and telephony servers. This underpins the IT-centric view of an SDP by viewing access to the network as an abstract service that the SDP builds upon. These network services are made available to application developers via the Web service gateway. Common services include the ability to send and receive messages, call control, initiating multimedia content downloads, and managing delivery receipt notifications for content delivery. The access layer hosts the master authentication server. As the central security node, this component conducts authentication of customers accessing the SDP platform and also authenticates access by external third-party applications that make use of the published Web services. In order to provide customers with trusted access to third-party applications, a federated identity management scheme is typically employed, for instance as defined by OASIS. This means that a customer need only login once to the SDP and may then access multiple third-party applications without the need to re-authenticate. Authentication exchange occurs on the user’s behalf between the SDP and external third-party application, for example using the Security Assertion Markup Language (SAML). Core services include streaming and gaming servers, digital rights management, a service catalog that provides a menu on the mobile device for each multimedia service available, several portals, and revenue collections facilities such as billing and payment. A key node is the Web service gateway which typically offers a range of network and IT-
Service Delivery Platforms in Mobile Convergence
Figure 3. Network-centric service delivery platform User Plane
Access Network
Transport Plane MRFP
Radio Access Network
S
Service Plane
Control Plane HSS
IN Apps
MRFC
SGSN
CSCF PDF
SIP Apps
Internet GGSN
PSTN
MGCF
MGW
Devices
BGCF Media Gateway Control
Service Delivery J2EE Apps Mgr
Applications
Support Planes (Service Management and Service Creation)
related services to third-party developers. This may include sending messages, prepaid, postpaid charging, payment, and additional authorization services. The portals furnish the visual interface accessed by customers, third-party developers, and internal administrators of the platform. Hence, there are generally several portals available to cater to the needs of each user community. The customer portal provides services to conduct registration, subscribe to services, and manage accounts. A service relationship management (SRM) portal is a vital component allowing third-party developers to register new applications, monitor usage, review revenues due, and provide an environment that facilitates development of new services; this may include testing and development tools (i.e., service creation environment). While clearer separation between the network and SDP environment is beneficial, particularly as network technologies evolve, there are several drawbacks. The design relies upon a mature network with standardized mechanisms for network access and control. Consequently, deployment under this assumption may be problematic where network access is inconsistent. IT-based SDPs generate billable events but do not themselves manage the balance or prevent access to services based on insufficient funds. Rather, they rely upon external systems for balance management. Furthermore, other than security and Web service definition, the supporting standards and literature are limited and emerging.
Network-Centric Design Viewpoint Service delivery frameworks that have emerged from standards bodies, such as IMS, Parlay X, and multimedia domain (MMD), largely reflect a network-centric design. In Magedanz et al. (2004, 2005), a service delivery platform
that extends the IP multimedia system has been developed. The authors outline a test-bed architecture that integrates the session initiation protocol (SIP) and Parlay with the telecommunications network to deliver multimedia services. A further platform for delivery of mobile games over the IMS has also been described (Akkawi et al., 2004). While there appears to be more standards work on the networkcentric design, the notion of IT-based systems is observed (Magedanz & Sher, 2006). In broad terms, IMS specifies an architecture that supports IP telephony and multimedia services such as instant messaging, videoconferencing, voice mail, and multiparty gaming (3GPP, 2001). Communications require session establishment between user devices and application servers, and the signaling protocol selected by IMS to establish these sessions is SIP. The architecture defines a layered model for the telecommunications network referred to as planes, similar in concept to the layers of the IT-centric design. These IMS planes are depicted in Figure 3 and include: access networks, being the physical network the end user connects to; transport plane, which is the common IP backbone network of the telecommunications service provider; control plane, which provides key functions of authentication, session establishment, and quality of service; and service plane, which houses application servers and provides an abstracted network interface. The access network connects the cellular, or fixed, network from the physical internetworking equipment to the customer premise or mobile device. For a device to connect to an IMS network, it needs to support an IMS client, providing various functions including the visual and audio presentation of the service to the user. Packets of data which encapsulate the service, such as VOIP from a PC, IPTV to a set top box, 873
Service Delivery Platforms in Mobile Convergence
or instant messaging to a mobile device, are all carried on a common transport plane. This plane contains the network elements, the gateway GPRS support node (GGSN) and serving GPRS support node (SGSN), media servers (media resource function processor (MRFP)) for playing announcements, and media gateways (MGW) for interconnecting IP traffic with other networks such as the PSTN. Control plane authentication of the user’s device is achieved by lookup to the home subscriber server (HSS), verifying that the device is enrolled and able to connect to the network and use the requested service. Session establishment is managed by the call session control function (CSCF), which connects the user to the correct application server in the service plane. The policy decision function (PDF) assigns resources to manage quality of service to ensure there is sufficient bandwidth available to deliver the requested service. Several nodes manage the media gateway (MGW) and internetworking between IP and SS7; this includes the media gateway control function (MGCF) and the breakout gateway control function (BGCF), which selects the required (local or foreign) network. As the broker between the MRFP and media applications, the media resource function controller (MRFC) is intended to perform conference, roaming, and media control; however, these capabilities are often developed directly within the applications of the service plane. The aim of the service plane is to provide an environment for executing various common IP services such as presence, location, messaging, videoconferencing, and multimedia. This is largely where the multimedia and telephony applications reside and is the least defined aspect of the architecture. IMS has to date largely focused upon network convergence, the establishment of IP sessions between users on different networks, how to ensure quality of service, and implementation of heritage features from the mobile phone network such as roaming between networks. As technology migrates to VOIP services, quality of service, mobility, and roaming across different access networks are seen as key features of IMS which will enable the telecommunications service provider to grow the business. Integral to the network-centric model is near-real-time balance management for the prepaid market, where there needs to be sufficient funds available within the customer account before the control plane grants access to the service being requested. This model becomes complicated when the customer may not select the desired service until the control plane has already allowed access to the service plane, particularly as the customer has access to the shopping portal resident in the service plane. In this instance, like the IT-centric model, the SDP needs to advise the balance manager what service is being requested before the service is initiated. A drawback of the IMS-based approach is lack of detail for SDP structures such as third-party relationship management, service or device management, and how to deliver 874
services in the service plane. This reflects the heritage of the authoring body, with specifications emerging from network engineering standards bodies. The focus is upon network application service delivery, omitting detail required for broader IT-based multimedia service delivery. Such capability would be added by introducing a service delivery manager (see Figure 3), which contains much of the detail of an IT-centric design.
FUTURE TRENDS Mobile devices continue to evolve in capability and computational strength. Together with the increasing reach of mobile networks, access to multimedia content and services will become ubiquitous. These factors contribute to trends in offering an increasing number of multimedia services to mobile customers. This includes gaming, podcasting, telephony services, and broadcast and video on demand. Digital rights management is as yet not aggressively implemented. However, as more content moves to this mobile digital environment, effective protection of digital content will become increasingly important. Perhaps the key consideration in the evolution of service delivery is the convergence of network IT systems and applications. As mobile networks move towards IP-based telephony, the distinction between the network and IT boundary grows dim, placing greater emphasis on well-defined standards and architectures. For instance, scenarios where users are able to view video content on a mobile device while roaming, pausing this when arriving home, with the ability to continue viewing using the home television, exemplify further the prospective convergence between mobile and fixed networks in the delivery of multimedia services. Services-oriented architecture (SOA) is emerging as the prevalent cross-industry IT integration regime and is likely to gain favor as a means of integrating components of an SDP. Given that integration of network IT and applications is central to an SDP, SOA may provide a means for IT and network-centric SDP models to converge.
CONCLUSION The standards for service delivery are still emerging, with specifications providing more detail regarding the network aspects of service delivery. An IT-centric SDP design lends more attention to the detail of integration with IT systems and third-party developers. This is beneficial where many implementations underestimate the effort and complexity of integrating with existing legacy systems. While there is a general lack of SDP standards for an IT-centric approach, Web service standards specified by OASIS in identify management and security exchange may be effectively
Service Delivery Platforms in Mobile Convergence
applied. A network-centric design is well defined at the network level, offering a greater range of network services. However, these implementations still require the additional consideration for supporting business processes, particularly for third-party service relationship management and application provisioning. Either approach largely reflects the heritage of the engineering discipline. However, as both the network and IT systems continue to converge, these boundaries will become less visible. Hence the need for clear architectural separation, as layers, becomes more important to ensure technology independence is maintained. Notwithstanding the design style adopted, there are several key considerations in an SDP to ensure the underlying business model is addressed. This includes the provision of an environment for rapid creation of multimedia applications, an intuitive user portal experience as an interface to a full range of content and services, and a comprehensive portfolio of Web services to enable the development of rich multimedia applications and services by external third parties.
REFERENCES Akkawi, A., Schaller, S., Wellnitz, O., & Wolf, L. (2004). A mobile gaming platform for the IMS. Proceedings of the 3rd International Workshop on Network and System Support for Games (Netgames 2004), Portland, OR. Anegg, H., Dangl, T., Jank, M. et al. (2004). Multimodal interfaces in mobile devicesthe MONA Project. Proceedings of the Workshop on Emerging Applications for Mobile and Wireless Access (WWW2004), New York. Deckers, G. (2006). Cost down, revenues up: SDP business case. Proceedings of the 10th International Conference on Intelligence in Service Delivery Networks (ICIN), Bordeaux, France. Devine, A. (2001). Mobile Internet content providers and their business models. Masters thesis, Department of Electrical Engineering and Management, The Royal Institute of Technology, Sweden. Hanrahan, H. (2006). Towards a standards based service delivery platform using service oriented reference points. Proceedings of the 10th International Conference on Intelligence in Service Delivery Networks (ICIN), Bordeaux, France. Hwang, T., Park, H., & Jung, J. W. (2004). The architecture of the digital home services delivery with OSGi. Proceedings of the IASTED Conference on Communication Systems and Applications (CSA 2004). Banff, Canada.
Kimbler, K., Stromberg, A., & Dyst Appium, J. (2006). The role of convergent service delivery platform in service migration to IMS. Proceedings of the 10th International Conference on Intelligence in Service Delivery Networks (ICIN), Bordeaux, France. Magedanz, T., & Sher, M. (2006). IT-based open service delivery platforms for mobile networks: From CAMEL to the IP multimedia system. In P. Bellavista & A. Corradi (Eds.), The handbook of mobile middleware. Chapman & Hall/CRC Press. Magedanz, T., Witaszek, D., & Knuttel, K. (2004). Service delivery platform options for next generation networks within the national German 3G beyond testbed. Proceedings of the South African Telecommunication Networks Architectures Conference (SATNAC04), Stellenbosch, South Africa. Magedanz, T., Witaszek, D., & Knuettel, K. (2005). The IMS Playground @ Fokusan open testbed for next generation network multimedia services. Proceedings of the 1st International IEEE Computer Society Conference on Testbeds and Research Infrastructures for the Development of Networks and Communities (TRIDENTCOM’05) (pp. 2-11). Pailer, R., Stadler, J., & Miladinovic I. (2003). Using Parlay APIs over a SIP system in a distributed service platform for carrier grade multimedia services. Wireless Networks, 9(4), 353-363. Pavlovski, C.J. (2002a). Reference architecture for mobile Internet service platform. Proceedings of the 2nd Asian International Mobile Computing Conference (AMOC 2002), Langkawi, Malaysia. Pavlovski, C.J. (2002b). Software architecture for mobile Internet service platform. Proceedings of the Workshop on Pervasive Computing, Going Beyond the Internet for Small Screens (OOPSLA 2002), Seattle, WA. Pavlovski, C.J., & Staes-Polet, Q. (2005). Digital media and entertainment service delivery platform. Proceedings of the 1st ACM International Workshop on Multimedia Service Composition (MSC ’05) (pp. 47-54). Singapore. Schulke, A., Kovacs, E., Stuttgen, H., Akkawi, A., Kuhnen, M., Riu, A., & Winkler, F. (2005). Creating new communication services efficiently. NEC Journal of Advanced Technology, 2(2), 170-178. Retrieved from http://www.nec. co.jp/techrep/en/r_and_d/a05/a05-no2/a0502p170.html 3GPP (3rd Generation Partnership Project). (2001). Technical specification group services and system aspects. Service Requirements for the IP Multimedia Core Network Subsystem (Stage 1) (Release 5) 3G TS 22.228 V5.0.0 (2001-01).
875
S
Service Delivery Platforms in Mobile Convergence
KEY TERMS Federated Identity Management: A standard that enables a user to use one set of credentials to sign on and access the networks of several enterprises in order to conduct transactions. Multimedia Domain (MMD): Specified by the 3GPP2 group, a set of specifications for the CDMA network based on IMS and Parlay X. Mobile Network Virtual Operator (MVNO): Being able to sell branded mobile network services without owning a mobile network, through an established relationship with a mobile network operator. Organization for the Advancement of Structured Information Standards (OASIS): A not-for-profit, international consortium that drives the development, convergence, and adoption of e-business standards.
876
OSE (Open Mobile Alliance) Service Environment: A reference architecture for how services may be combined to form new services (Schulke et al., 2005). The standard is intended to promote common architectural principles and broadly defines the functional entities, within a telecommunications service environment, for application developers and service providers. Parlay X: Web service extension to the Parlay APIs that allow developers to build applications using the Parlay API. The Parlay API is an open interface to the mobile and fixed telephone network. Security Assertion Markup Language (SAML): An XML-based framework that defines an interoperable standard for exchanging security information. 3rd Generation Partnership Program (3GPP): A collaboration between several telecommunications standards bodies to develop 3G GSM technical specifications.
Category: Service Computing 877
Service Provision for Pervasive Computing Environments Emerson Loureiro Federal University of Campina Grande, Brazil Frederico Bublitz Federal University of Campina Grande, Brazil Loreno Oliveira Federal University of Campina Grande, Brazil Nadia Barbosa Federal University of Campina Grande, Brazil Angelo Perkusich Federal University of Campina Grande, Brazil Hyggo Almeida Federal University of Campina Grande, Brazil Glauber Ferreira Federal University of Campina Grande, Brazil
INTRODUCTION The fast development on microelectronics has promoted the increase on the computational power of hardware components. On the other hand, we are facing a significant improvement on energy consumption as well as the reduction of the physical size of such components. These improvements and the emergence of wireless networking technologies are enabling the development of small and powered mobile devices. Due to this scenario, the so-called pervasive computing paradigm, introduced by Mark Weiser in 1991 (Weiser, 1991) is becoming a reality. Such a paradigm envisions a world where environments are inhabited by computing devices, all of them seamlessly integrated into peoples’ lives, and effectively helping to carry on their daily tasks. Among others, one major characteristic of Weiser’s vision is that each device in an environment becomes a potential client or provider of resources. Not surprisingly, pervasive computing environments are becoming dynamic repositories of computational resources, all of them available to mobile users from the palm of their hands. However, devices can unpredictably join and leave such environments. Thus, resources can be dynamically made available or unavailable. Such a scenario has a great impact on the way that resources are found and used. In the case of static environments, such as the Web, it is reasonable to look up and access resources,
such as Web pages, knowing the address of their providers beforehand. On the other hand, for dynamic environments, such as the pervasive computing ones, this is not a reasonable approach. This is due to the fact that one cannot guarantee that the provider of a resource will be available at any moment, because it may have left the environment or simply turned off. A better approach would be to discover these resources based on their descriptions, or any other feature that does not require the client to know the specific address of their providers. To this end, some of the current pervasive computing solutions, like Wings (Loureiro, Bublitz, Oliveira, Barbosa, Perkusich, Almeida, & Ferreira, 2006), Green (Sivaharan, Blair, & Coulson, 2005), RUNES (Costa, Coulson, Mascolo, Picco, & Zachariadis, 2005), and Scooby (Robinson, Wakeman, & Owen, 2004), are making use of a novel approach from the branch of distributed applications, the service-oriented computing paradigm (Papazoglou, 2003; Huhns & Singh, 2005). This is due to the fact that such a paradigm provides a crucial element for pervasive computing systems, the ability for dynamically binding to remote resources (Bellur & Narenda, 2005), which enables mobile devices to find needed services on demand. However, pervasive environments may be structured in different ways. They can range from wired networks to completely wireless ones, where communication among the
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Service Provision for Pervasive Computing Environments
Figure 1. General view of a service-oriented architecture
devices is performed in an ad hoc way. Such a characteristic indicates that the way services are provisioned in a pervasive computing environment should fit in its organization, in order to enhance the access to the services available. Considering the above discussion, in this article we provide a review on service provision and its applicability in pervasive computing. More precisely, we will list the existing service provision approaches and discuss the characteristics and problems associated with each one, as well as their usage in pervasive computing environments. We start by providing introductory concepts of service-oriented and pervasive computing, respectively in the service-oriented computing and pervasive computing sections. Next, we present the service provision techniques available and how they can be applied for pervasive computing environments. The main current solutions within this scope will be introduced in the service oriented technologies section. Some of the future trends associated with research for service provision in pervasive computing environments will be presented in the future research trends section. Finally, in the conclusions section we present the conclusions of this article.
SERVICE-ORIENTED COMPUTING The service-oriented computing (SOC) paradigm has been considered as the next step in distributed computing (Papazoglou, 2003). In a general way, this paradigm can be viewed as the development of applications through the runtime integration of software pieces named of services (McGovern, Tyagi, Stevens, & Mathew, 2003). In this process, three elements are involved: defining what is known as a service-oriented architecture (SOA), a service client, a service provider, and a service registry. The former is the one who whishes to use a service. Conversely, service providers are those which offer services for potential clients. 878
Finally, the service registry is where providers advertise or announce their services (through service advertisements), enabling clients to dynamically discover them. By dynamic, we mean that clients are capable of discovering services at runtime, thus providing a high degree of flexibility for applications. Once clients have discovered a service, they are able to bind to it; that is, to create a link with the service, in order to use it (through a proxy to the real service). This process of advertising, discovering, binding, and using a service is illustrated in Figure 1. In open environments, the dynamic discovery of services implies that they can be used by heterogeneous clients. Within this scope, heterogeneity is concerned with features like the operating system running in each client and the hardware platform it has been built on. As a consequence of such heterogeneity, for enabling an application to flexibly integrate services, they should present the following features (Papazoglou, 2003): • • •
Loose Coupling: A service must not require from the clients any knowledge about its internal implementation. Implementation Neutrality: The usage of services must not rely on any specific programming language, operating system, or hardware platform. Dynamically Discoverable: Services should be discovered at runtime.
PERVASIVE COMPUTING The field of pervasive computing has its origins at the Xerox Palo Alto Research Center. The pioneer work that has been led there has culminated in the novel article of Mark Weiser in 1991 (Weiser, 1991), where he describes the fist ideas of pervasive computing. Weiser’s vision is at the same time revolutionary and simple: a world where computing is embedded in everyday objects, like cars, televisions, and air conditionings, all seamlessly integrated into our lives and performing tasks for us (Turban, Rainer, & Potter, 2005). When Weiser talked about seamless integration, he meant that applications running in these objects should act proactively on our behalf. They should, for example, present us with relevant information, based on what we want/need and the resources (e.g., a printer) available in the environment we are immersed.
SERVICE PROVISION APPROACHES IN PERVASIVE COMPUTING When provisioning services in a pervasive environment, one aspect to be considered is the way it is organized; that
Service Provision for Pervasive Computing Environments
Figure 2. Example of a push-based service provision
is, whether the environment is based on a wired network infrastructure, whether it is formed in an ad hoc way, or both. This is necessary for dealing with the particularities of each environment, and within this scope, we can say that there are two major ways of performing service provision (Nickull, 2005): the push-based and the pull-based approach. In the next sections we will outline each of these approaches as well as describe how they fit into pervasive computing environments.
Push-Based Service Provision In this approach, a provider advertises its services directly to potential clients. In other words, it sends service advertisements to all network hosts, either they are interested on the service or not. Such advertisements, when received by a host, will be kept in a local service registry. Therefore, once a client wants to discover some service, it will then look up in such a registry for the services that match its needs. An example of the push-based service provision is illustrated in Figure 2. Note that the provider, host A, sends the advertisement of a service directly to hosts B and C, as illustrated in Step 1. When this advertisement is received, it will be stored on a local service registry at each host (Step 2). Then, once a host, in our example host C, wants to discover a service, it then inquires this registry, as illustrated in Step 3. Finally, considering that a relevant service has been found, it is possible to ask its provider (host A) for a proxy to the service (Step 4), enabling host C to use it.
S
Using this approach, one major problem to be pointed out is about the validity of the service advertisements. It is concerned with the fact that a service provider can leave the network, but the advertisements associated with its services can still be available. One approach which could be used for solving this problem is to require the provider to explicitly notify the network hosts about its leaving, and consequently its services will be no longer available. The problem is that it is not always possible to do that. For example, if the provider is a mobile device, it may be suddenly run out of energy. To deal with this, providers could be aware of the energy level of the device, in order to notify the network hosts that within some minutes it may not be accessible anymore. However, other factors can be involved in the leaving of a provider from the network. It can be just turned off by its user, or leave the network coverage area. Keeping track of all these factors is a task that certainly overloads the provider. A more reasonable solution would be requiring both providers and clients to cooperate for renewing the service advertisement. Therefore, as long as the advertisements are renewed, it is possible, but not guaranteed, that the service is available. On the other hand, when the advertisement has not been renewed within a time interval, then the service is probable, but also not guaranteed, to be unavailable, either because its provider has left the network or because the service has been unadvertised. One interesting point to notice is that the push-based service provision does not require any physical infrastructure. This means that such an approach is well suited in decen-
879
Service Provision for Pervasive Computing Environments
tralized and/or infrastructure-less pervasive environments. One problem, in the scope of pervasive computing, is that the advertisement task can consume a lot of bandwidth if many devices are provisioning services in the environment. Therefore, in environments with very limited wireless links this is certainly a major problem. On the other hand, as services are searched locally, the discovery process does not involve costs of communication.
Pull-Based Service Provision In the pull-based approach, in order to discover services clients must inquiry remote registries for the needed services. This can be performed in two ways; either using centralized or distributed registries.
Centralized Provision The centralized service provision consists in scattering service registries in specific servers (i.e., registry servers) of the network. Therefore, for advertising a service, the provider must initially find which of these servers are available in the network. After that, it has to determine in which of them the service will be advertised (instead, the provider could advertise the service in all the available servers). When a client wants to discover services, it must also find the registry servers available in the network, and then
Figure 3. Example of a centralized service provision
880
discover the services advertised in them. It is important to notice that, once the registry servers are found, unless they become unreachable, clients and providers do not need to discover them anymore. Jini and Web Services are some of the current technologies that support centralized service provision. In Figure 3 we illustrate an example of such an approach. In such a figure, services are advertised by hosts A and B (Step 1), in a single registry server, host C (we are considering that the clients have already found the registry server). After that, each advertisement is stored in the service registry maintained by host C (Step 2). Also considering that host D has already found the registry server, it is then able to inquiry such server for the advertised services (Step 3). When a relevant service is found (Step 4), host D can interact with its provider, in this case host A, to retrieve the proxy for the service (Step 5). A problem with this approach is that, if all these servers are off the network, services can not be discovered, even when their providers are available. In these cases a possible solution could be the election of new servers from the moment that is detected that the others are no longer available. Furthermore, the centralized service provision raises the same problem of the pull-based one concerned with the validity of the service advertisements. In this way, the same, or at least similar, solutions can be applied here. As the centralized service provision requires the existence of registry servers, it is not well suited for highly dynamic
Service Provision for Pervasive Computing Environments
Figure 4. Example of a distributed service provision
pervasive environments. In these cases, as nodes join and leave the environment all the time, there would be too many changes in the current registry servers, which would in turn degrade the provisioning of services. On the other hand, this approach is very useful for environments equipped with wired network. In such environments, services can be deployed in the wired network and thus be accessed by mobile clients through wireless links. In environments populated with lots of mobile clients, this is certainly a good choice, as the bandwidth available in the wired network could support a great number of accesses.
Distributed Provision In the distributed service provision, services are advertised in registries located in each host. Undoubtedly, in this approach the advertising task is easier to be performed than in the other ones, as it does not involve sending advertisements to central servers or directly to the other hosts. However, service discovery is more complicated, as it must be performed by inquiring each available host for the needed services. As no centralizer hosts are necessary for advertising services, discovery is possible whenever a client and a provider are present in the network. An example of the distributed service provision approach is illustrated in Figure 4. Initially, each host advertises its services (Step 1). Once a client needs to perform service discovery, in our example host A, it asks every host in the network for the needed service (Step 2). It
S
is important to note the possibility of redirecting the service discovery request to hosts that are not reachable from the client. In Figure 4 this is performed by host B when it redirects the request of host A to C. When a host has a service matching the client’s needs, it sends a notification (Step 3). From this notification, the client can then retrieve a proxy to the service (Step 4). The major problem with this approach is associated with the protocols for performing service discovery. As any node in the network is a potential service provider, the discovery protocol must be well designed, in order to cover all hosts of the network. If any host is missing in the discovery process, it is possible that a relevant service may not be found. In ad hoc pervasive environments, one possible solution to this problem is first to discover the hosts in the vicinity, and then to retrieve the relevant services they provide. However this solution only performs well in small networks, where the available hosts can be discovered from any other one. Another point to be considered is about the network congestion that such service discovery protocols can cause. As the search should include all hosts in the network, the protocol must apply techniques for avoiding flooding the network. Obviously, this is not a serious problem in wired networks, but considering the wireless ones, it must be strictly taken into account. A good usage scenario of the distributed provision approach is a decentralized pervasive environment where the edge of the network is formed by mobile clients and its core is populated by service providers connected through a wired network. 881
Service Provision for Pervasive Computing Environments
SERVICE ORIENTED TECHNOLOGIES In this section we present the main technologies related to the provision of services in pervasive computing environments.
Universal Plug and Play (UPnP) The Universal Plug and Play (Richard, 2000) is an open architecture, which uses standard Internet protocols for pervasive peer-to-peer network connectivity (http://www.upnp. org). The UPnP protocol defines a set of steps, addressing, discovery, description, control, eventing, and presenting, which enables the automatic network configuration and service provision. Through the addressing, devices get their network address, which is performed by a dynamic host configuration protocol (DHCP). The discovery step consists of notifying the other hosts, through the push-based approach, about the services, and embedded devices, that a joining host provides. The discovery can also be performed in a distributed pull-based fashion. The next step, description, is about the description of a device, stored as an XML document. In such a document, it is kept, among other information, the description of the services that the device provides. Therefore, by discovering a device, it is possible to retrieve the services it provides, and then, through the control step, invoke their actions. Changes in a service can be notified to interested devices through the eventing step. Finally, through the presenting step, it is possible to load a specific URL, specified in the device description, in order to display a user interface for controlling a device.
Jini Jini is a service-oriented Java technology based on a centralized pull-based approach (Waldo, 1999). Therefore, service
Figure 5. Service discovery in Bluetooth
882
advertisements are stored in central servers, which are named lookup servers. Jini uses the RMI (http://java.sun.com/products/jdk/rmi - Remote Method Invocation) protocol for all interactions involved in the advertisement, discovery, and invocation of services. When a client discovers and binds to a service, it is incorporated to the client by downloading the code of a proxy to the required service, named remote control object. The Jini platform uses the concept of lease for controlling the access to the services. A lease is a sort of warrant that a client has for using a service during a specific period of time. When the lease expires the client needs to renew it with the provider if it wishes to continue using the service.
Bluetooth Bluetooth is a standard for wireless communication among small devices within short distances (Johansson, Kazantzidis, Kapoor, & Gerla, 2001), defining higher-level protocols for both host and service discovery (http://www.bluetooth. org). The discovery of services in the Bluetooth standard is defined by the service discovery protocol (SDP), which enables to enumerate the devices in the vicinity and retrieve the services they provide. Bluetooth uses a distributed pull-based approach for service advertising and discovery. To this end, each device maintains its own service discovery database (SDDB), which is a registry where its services are advertised. Therefore, a Bluetooth device performs service discovery by querying the SDDBs of the devices around. These advertising and discovery processes are illustrated in Figure 5. Notice that, initially, all devices advertise their services on their respective SDDBs (1). Next, a client searches for all the Bluetooth devices on the range of its wireless interface (2). For each device found, the client sends a query about the availability of an interested service (3). The devices answer these queries
Service Provision for Pervasive Computing Environments
by informing whether they offer the needed service or not (4). Once localized a device providing the desired service, the client can connect directly to such device and finally use the service (5).
FUTURE RESEARCH TRENDS Although the first service-oriented pervasive computing solutions have been developed, much work has to be done yet. For example, the matching of the user’s needs and the services’ functionalities should be enhanced to improve the interaction of the user with the pervasive application. Still, problems remain in the context of service continuity. Solutions to this problem would enable the user to use a service continuously, as he/she walks through different environments. This is important because, sometimes, the service a client was using is not available in the new environment, and thus, some mechanism should allow it to use a similar one without, or at least with a minimum, of interruption. These, and possibly other problems related to the provision of services in pervasive environments, must certainly be completely solved so that we can enjoy the full potential of merging service-oriented and pervasive computing.
CONCLUSION The service-oriented paradigm has proved to be an important element in pervasive computing systems, in order to provide anytime and anywhere access to services. Its dynamic binding feature enables to build applications powered with on-demand extensibility and adaptability, two important elements of any pervasive system. Given this trend, in this chapter we have tried to present an overview of service provision in pervasive computing environments. More precisely, we have showed an introduction to the main characteristics, challenges, and solutions concerning the way that services are advertised, discovered, and used in pervasive environments. Although we presented concepts at an introductory level, we believe they may serve as a good source of knowledge, helping both students and researchers involved with these fields.
REFERENCES Bellur, U., & Narendra, N. C. (2005). Towards service orientation in pervasive computing systems. In International Conference on Information Technology: Coding and Computing (Vol. II, pp. 289-295). Las Vegas, NV.
middleware: A reconfigurable component-based approach to networked embedded systems. In Proceedings of the 16th IEEE International Symposium on Personal Indoor and Mobile Radio Communications. Berlin, Germany: IEEE Communications Society. Huhns, M. N., & Singh, M. P. (2005). Service oriented computing: Key concepts and principles. IEEE Internet Computing, 9(1), 75-81. Johansson, P., Kazantzidis, M., Kapoor, R., & Gerla, M. (2001). Bluetooth: An enabler for personal area networking. IEEE Network, 15(5), 28-37. Loureiro, E., Bublitz, F., Oliveira, L., Perkusich, A., Almeida, H., & Ferreira, G. (2006). A flexible middleware for service provision over heterogeneous pervasive networks. In Proceedings of the 4th International Workshop on Middleware for Mobile and Distributed Computing. Niagara Falls, NY: IEEE Computer Society. McGovern, J., Tyagi, S., Stevens, M., & Mathew, S. (2003). Service oriented architecture. In J. McGovern, S. Tyagi, M. Stevens, & S. Mathew (Eds.), Java Web services architecture (pp. 35-63). San Francisco: Morgan Kaufmann. Nickull, D. (2005). Service oriented architecture (White Paper). San Jose, CA, USA: Adobe Systems Incorporated. Retrieved March 28, 2006, from: http://www.adobe.com/enterprise/pdfs/Services_Oriented_ Architecture_from_Adobe.pdf Papazoglou, M. P. (2003). Service-oriented computing: Concepts, characteristics, and directions. In Proceedings of the Fourth International Conference on Web Information Systems Engineering (pp. 3-12). Rome: IEEE Computer Society Robinson, J., Wakeman, I., & Owen, T. (2004). Scooby: Middleware for service composition in pervasive computing. In Proceedings of the 2nd Workshop on Middleware for Pervasive and Ad hoc Computing. Toronto, Canada. Richard, G.G. (2000). Service advertisement and discovery: Enabling universal devicecooperation. IEEE Internet Computing, 4(5), 18-26. Satyanarayanan, M. (2001). Pervasive computing: Vision and challenges. IEEE Personal Communication, 8(4), 10-17. Sivaharan, T., Blair, G., & Coulson, G. (2005, October). GREEN: A configurable and reconfigurable publish-subscribe middleware for pervasive computing. In Proceedings of the International Symposium on Distributed Objects and Applications (Vol. 3760, pp. 732-749). Agia Napa, Cyprus: Springer Verlag.
Costa, P., Coulson, G., Mascolo, C., Picco, G. P., & Zachariadis, S. (2005). The RUNES 883
S
Service Provision for Pervasive Computing Environments
Turban, E., Rainer, R. K., & Potter, R. (2005). Mobile, wireless, and pervasive computing. Information Technology for Management: Transforming Organizations in the Digital Economy (pp. 167-206). New York: John Wiley & Sons. Waldo, J. (1999). The Jini architecture for network-centric computing. Communications of the ACM, 42(7), 76-82. Weiser, M. (1991). The computer for the 21st century. Scientific American, 265(3), 94-104.
KEY TERMS Pervasive Computing: The vision conceived by Mark Weiser which consists of world where computing will be embedded in every day objects.
884
Service: a software entity that can be integrated to a remote distributed application. Service-Oriented Computing: The newest paradigm for distributed computing, where applications should be built by dynamically integrating services. Service Advertisement: The element used for publishing and discovering a service. Service Client: The one wishing to use a service. Service Provider: The one that offers services. Service Registry: The place where services are published.
Category: M-Entertainment 885
Short Message Service (SMS) as an Advertising Medium Shintaro Okazaki Autonomous University of Madrid, Spain
INTRODUCTION The proliferation of the Internet-enabled mobile device has extended into many parts of the world. Collectively, the mobile-network operators paid more than $100 billion for licenses to operate “third-generation” (3G) networks, which were among “the largest bet in business history on the introduction of a new technology” (Economist, 2005). This drastic move has been most illustrated by the use of short message service (SMS) and multimedia messaging service (MMS) by mobile users. For example, a recent survey indicates that SMS in the Asia-Pacific region will increase to up to 75% of mobile subscribers in 2006 (IDC Asia/Pacific, 2003). As a result, marketers and agencies are increasingly interested in taking advantage of this growth, by incorporating SMS advertising as part of an integrated marketing communications (IMC) strategy. However, there has been little academic research on mobile advertising, perhaps because its growth is still in an early stage and the technological infrastructure varies across markets. The study has two objectives: (1) to identify the factors influencing MNCs’ managerial intention to adopt SMS advertising, and (2) to test a statistical relationship between these factors and managerial intention to use SMS advertising. To this end, we conducted telephone interviews of senior executives of MNCs operating in European markets.
CONCEPTUAL FRAMEWORK AND HYPOTHESES Branding Technique In an environment where building the brand is a fundamental goal for many managers, the need to build brand equity is likely to be at the center of many marketing decisions. Firms using SMS-based campaigns can attract consumer attention and produce consumer responses to a much greater degree than via other direct marketing channels, because SMS has been claimed to be an effective tool in building and testing customer loyalty by developing demographic databases (Mylonopoulos & Doukidis, 2003). From an industry perspective, McDonald’s conducted a text-messaging campaign in conjunction with a popular TV song contest in the UK,
offering concert tickets and backstage passes, while entry in the Coca-Cola Grand Sweepstakes Competition was offered to U.S. college students who sent a text message to a number printed on a Diet Coke can (Dano, 2002).
Facilitating Conditions Lu, Yu, Liu, and Yao (2003) suggest that facilitating conditions are one of the most important determinants, along with the ease of using wireless Internet. In this light, the integration of competing standards and fragmented systems across countries, cross-network support for SMS, and higher connection speeds are all necessary conditions for a wider transmission of mobile advertising. In addition, the availability of Web-enabled mobile handsets with 2.5G or 3G functionality would significantly affect the adoption of MMS-based (multimedia message services) campaigns. In this light, a wider selection of handsets must be available, to enable consumers to choose their preferred combination of necessary functions and diverse features.
Location-Based Services The satellite-based global positioning system (GPS) offers the ability to tailor services and promotional offers to individual consumers’ needs, by locating their position (Sadeh, 2002). Mobile handset makers and content providers are increasingly attracted by the commercial feasibility of applying GPS to their service. For example, on an extended menu of i-mode, “i-area” includes a diverse range of location-based services: weather news, restaurant guide, local hotel information, zoomable maps with an address finder function, and traffic updates and estimation of travel times. This facility would give MNCs strategic leverage in mobile marketing, because individuals’ behavior and receptiveness to advertising is likely to be influenced by their location and time, and marketers can thus induce impulse buying by providing the right information for the right place (Barnes, 2002).
Connection Costs Another important factor is the concept of connection costs. For example, to send or receive one megabyte of data on 2.5G i-mode costs 32 euros (0.3 yen) per packet. At a rate
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Short Message Service (SMS) as an Advertising Medium
of 19 euro cents per 160-character SMS message, European consumers would have to pay 1,356.98 euros to send one megabyte of data by SMS, or approximately 62 times as much as the Japanese pay (Scuka, 2003). In addition, European mobile operators have passed on to consumers the additional costs incurred in obtaining 3G spectrum licenses, and this has made any dramatic price reduction impossible (Baldi & Thaung, 2002). Such cost factors adversely affect mobile players’ revenues.
Public Regulation The idea behind mobile advertising is very similar to e-mail on the wired Internet, but with one big difference: it is “optin.” This function is essential to give users total control over what they receive, because consumers’ demand for highly personalized messages has to be reconciled with their desire for privacy (Sadeh, 2002). The Mobile Marketing Association (MMA) has attempted to establish industry guidelines for mobile marketers, as follows: (1) MMA members should not send mobile advertising without confirmed opt-in, and (2) such opt-in subscriber permission is not transferable to third parties without explicit permission from the subscriber (Petty, 2003).
Lifestyle and Habits In general, European consumers habitually commute by car, and this provides fewer incentives to access the mobile Internet (Baldi & Thaung, 2002). In addition, a systematic “word-of-mouth” helped the rapid diffusion of i-mode in Japan, especially given the “normative beliefs attributed to significant others (friends, colleagues, or family members) with respect to adopting or continuing to use the technology” (Barnes & Huff, 2003). This may partially explain a high subscription rate (almost 75%) to e-mail newsletters among i-mode users, and this makes acceptance of mobile advertising much easier. However, this factor is unlikely to be present in many European countries, which are characterized as more individualist than Asian countries. On the basis of the preceding discussions, the following hypotheses were formulated to test the principal thesis of the research: • • •
886
H1: MNCs’ intention to adopt SMS-based advertising is directly and positively associated with branding technique. H2: MNCs’ intention to adopt SMS-based advertising is directly and positively associated with facilitating conditions. H3: MNCs’ intention to adopt SMS-based advertising is directly and positively associated with location-based services.
• • •
H4: MNCs’ intention to adopt SMS-based advertising is directly and negatively associated with connection costs. H5: MNCs’ intention to adopt SMS-based advertising is directly and negatively associated with public regulation. H6: MNCs’ intention to adopt SMS-based advertising is directly and negatively associated with lifestyle and habits.
METHODOLOGY Questionnaire Items and Measures A structured questionnaire was prepared, drawing on prior literature. A majority of the items were originally developed for this study, because of the scarcity of empirical research on mobile advertising. Each item was measured on a Likert-type five-point scale. A five-point scale was preferred to a sevenpoint scale, because telephone interviews were used, rather than a mail or other form of paper-and-pencil survey. This method was considered more appropriate because mobile advertising is still in its infancy, and company executives may not be able to make fine distinctions regarding their attitudes on this topic. During the telephone interview, interviewers followed a script. However, respondents were free to ask questions whenever they encountered definitional problems.
Multinational Corporations With regard to Japanese firms, the selection was based on the Multinational Companies Database. The database was created by the Research Institute for Economics and Business Administration at Kobe University (2003) and includes Japanese companies listed in the first section of the Tokyo Stock Exchange with foreign direct investment in more than five countries (Kobe University, 2003). American firms were chosen from The Forbes 500 (Forbes, 2003a). Finally, European firms were singled out from The Forbes International 500 (Forbes, 2003b), because this list indicates the nationality of each firm. Regardless of nationality, however, companies associated with aerospace and defense, food and drug retail chains, forestry and fishery, general public utilities, health care providers, heavy machines, industrial goods, local banking and insurance, metals and mining, and oil and gas extraction were excluded. Next, firms operating in Spain were identified. As a result, 43 Japanese, 47 American, and 31 European firms’ Spanish subsidiaries were identified.
Short Message Service (SMS) as an Advertising Medium
Table 1. Regression analysis and hypotheses testing
S
Hypotheses
Independent Variables
Standardized β
Results
H1
Branding technique
.553
**
Supported
H2
Facilitating conditions
.325
**
Supported
H3
Location-based services
.023
H4
Connection costs
-.397
H5
Public regulation
.125
Rejected
H6
Lifestyle and habits
-.115
Rejected
R2
.598
∆R
.013
∆F
1.509
2
Rejected **
Supported
**
Note: Dependent variable = MNCs’ intention to use mobile advertising ** Significant at p < .001 * Significant at p < .05
Telephone Interview Telephone interviewing was considered appropriate because of the novelty of the research subject. It was expected that interviewers would be able to clarify doubts or answer any questions that interviewees might have regarding mobile communications. To this end, four bilingual assistants were employed (two Spanish and two Japanese, all fluent in English). During the second and third weeks of February 2004, intensive training was provided so that the assistants could gain sufficient skills and knowledge to conduct the telephone interview. The actual interviewing was carried out during March 2004, under the supervision of the researcher. It was established that when the target executives were absent or unavailable for interview, assistants had to ask: (1) for an appointment for the next phone call, or (2) about the availability of the person next in seniority in the marketing department to the target executive. As a result, a total of 53 interviews was conducted, with 27, 16, and 10 respondents from Japanese, American, and European firms, respectively. The response rate was 43.8%.
FINDINGS First, an exploratory factor analysis with Equamax rotation with Kaiser Normalization was carried out. The rotation, converged in 12 iterations, produced a clear-cut six-factor solution with a cut-off value of .50. Only factors with eigenvalue greater than 1 were retained. It should be noted that the proposed construct “connection costs” was merged into a mixed construct “security and costs.” However, because
of the exploratory nature of the study, it was considered acceptable to use this six-factor solution for the subsequent analysis. The extracted factors explain 68.6% of the total variance, and the level of loading is consistently high across the six factors. Factor scores were retained as variables with the Anderson-Rubin method to minimize the level of multicollinearity, for the use of regression analysis. The reliability was calculated with Chronbach’s alpha for each construct. The scores range from .60 to .85, exceeding the cut-off point of .60 suggested by Hair, Anderson, Tatham, and Black (1998). Next, the hypotheses were tested by performing regression analysis with a step-wise method. Each of six independent variables (i.e., factor scores) was regressed on the dependent variable, “MNCs’ intention to use mobile advertising,” in order of their expected contributions. The results of regression analysis are shown in Table 1.
DISCUSSION This study aims to identify MNCs’ principal perceptions of SMS-based push-type mobile advertising and their intention to use it. On the basis of the data obtained from 53 MNCs, our principal propositions were tested by multiple regression analysis. The results were mixed: only half of the six hypotheses gained empirical support. The regression analysis identified branding technique, facilitating conditions, and connection costs as the three primary predictors influencing MNCs’ intention to use mobile advertising. The contribution of branding technique in particular is substantial, indicating that MNCs are likely to perceive mobile advertising as an effective branding tool to increase brand awareness and im887
Short Message Service (SMS) as an Advertising Medium
age. Also, technological infrastructure and the availability of sophisticated mobile handsets are prerequisites for mobile marketing. As expected, unfavorable mobile Internet pricing negatively affects the MNCs’ intention to use mobile advertising. On the other hand, the contributions of location-based services, public regulation, and lifestyle and habits are not only statistically insignificant, but also trivial in terms of the coefficient magnitude. One reason why location-based services were not identified as a significant factor is that the GPS system is not as widespread in Europe as it is in Japan. In addition, many Scandinavian firms, leaders of sophisticated mobile Internet service practitioners, were not included in the study. Admitting the danger of simple generalization, the findings of this study may imply that MNCs are concerned to a lesser extent with regulatory and cultural impediments to adopting mobile advertising.
Lu, F., Yu, C. S., Liu, C., & Yao, F. E. (2003). Technology acceptance model for wireless Internet. Internet Research, 13(3), 206-222. Mylonopoulos, N. A., & Doukidis, G. I. (2003). Introduction to the special issue: Mobile business: Technological pluralism, social assimilation, and growth. International Journal of Electronic Commerce, 8(1), 5-22. Petty, R. D. (2003). Wireless advertising messaging: Legal analysis directly and public policy issues. Journal of Public Policy and Marketing, 22(1), 71-82. Robinson, J. P., Shaver, P. R., & Wrightsman, L.S. (1991). Criteria for scale selection and evaluation. In J. P Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 1-16). San Diego: Academic Press.
REFERENCES
Sadeh, N. (2002). M-commerce: Technologies, services, and business models. New York: John Wiley & Sons.
Baldi, S., & Thaung, H. P. P. (2002). The entertaining way to m-commerce: Japan’s approach to the mobile InternetA model for Europe? Electronic Markets, 12(1), 6-13.
Scuka, D. (2003). How Europe really differs from Japan. Retrieved February 11, 2004, from http://www.mobiliser. org/article?id=68
Barnes, S. J. (2003). Wireless digital advertising: Nature and implications. International Journal of Advertising, 21, 399-420.
KEY TERMS
Barnes, S. J., & Huff, S. L. (2003). Rising sun: i-mode and the wireless Internet. Communications of the ACM, 46(11), 79-84. Barwise, P., & Strong, C. (2002). Permission-based mobile marketing. Journal of Interactive Marketing, 16(1), 14-24. Dano, M. (2002). Coke, Toyota, McDonald’s test mobile advertising. RCR Wireless News, 21(46), 8. Forbes. (2003a). The Forbes 500s. Retrieved November 3, 2003, from http://www.forbes.com/2003/03/26/500sland. html Forbes. (2003b). The Forbes International 500. Retrieved November 15, 2003, from http://www.forbes. com/2003/07/07/internationaland.html Hair, J. F. Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall. Kobe University. (2003). Multinational Companies Database. Available by permission of Research Institute for Economics and Business Administration of Kobe University. Retrieved January 11, 2004, from http://www.rieb.kobe-u.ac. jp/liaison/cdal/takokuseki/dbenterprises.html
888
Barcode Mobile Coupon: Mobile barcoding can be used in the form of a picture SMS which is delivered to a mobile phone. Recipients save the image, arrive at the destination, and present their barcode SMS to be scanned. i-mode: A broad range of Internet services for a monthly fee of approximately three euro, including e-mail, transaction services (e.g., banking, trading, shopping, ticket reservations, etc.), infotainment services (e.g., news, weather, sports, games, music download, karaoke, etc.), and directory services (e.g., telephone directory, restaurant guide, city information, etc.), which offers more than 3,000 official sites accessible through the i-mode menu. Push Messaging Service: Various forms of messaging services are generally offered in mobile Internet. For example, SMS and WAP Push messaging generally allow users to send 100-160 characters, while mobile e-mail in Japanese i-mode allows up to 1,000 characters. SPAM: Unsolicited or undesired bulk electronic messages. Because of the development of anti-SPAM programs, they are often deleted without being opened. SPIM: A variation of SPAM through instant messaging systems.
Category: Mobile Multimedia 889
Shot Boundary Detection Techniques for Video Sequences H. Koumaras N.C.S.R., Demokritos, Greece G. Xilouris N.C.S.R., Demokritos, Greece E. Pallis Technological Educational Institute of Crete, Greece G. Gardikis University of the Aegean, Greece A. Kourtis N.C.S.R., Demokritos, Greece
INTRODUCTION The advances in digital video encoding and compression techniques that achieve high compression ratios by exploiting both spatial and temporal redundancy in video sequences have made possible the storage, transmission, and provision of very high-volume video data over communication networks. Today, a typical end user of a multimedia system is usually overwhelmed with video collections, facing the problem of organizing them in a browsing-friendly way. Thus, in order to allow an efficient exploitation and browsing of these video-anthologies, it is necessary to design techniques and methods for content-based search and access. Therefore, the issue of analyzing and categorizing the video content by retrieving highly representative optical information has been raised in the research community. Thus, the current trend has led to the development of sophisticated technologies for representing, indexing, and retrieving multimedia data. A common first step towards this is the segmentation of a video sequence into elementary shots, each comprising a sequence of consecutive frames that record a video event or scene continuously in the spatial and temporal domain. Moreover, these elementary shots appear as they have been captured by a single camera action. Two adjacent elementary streams are divided by a shot boundary or shot transition, also known as scene cut, when the change of video content occurs over a single frame, or gradual shot boundary, when the changes occur gradually over a short sequence of frames (e.g., dissolve, fade in/out, etc.) (Lu & Tan, 2005). In general, gradual transitions are more demanding in detection than abrupt scene cuts, because they must be
distinguished from regular camera operations that cause similar temporal variances and usually trigger false detections. Especially for video content with high spatial and temporal activity level, the detection of gradual scene changes becomes even more challenging (Hampapur, Jain, & Weymouth, 1995). Hence, the goal of this temporal video segmentation is to divide the video stream into a set of meaningful and manageable segments that are used as basic elements for indexing. Further analysis may be performed, such as representation of the video content and event identification. In future multimedia systems, the offered video services will be provided in the form of MPEG-21 digital items, which integrate a typical encoded media clip along with its XML-based metadata descriptors, enabling in this way advanced search and retrieve abilities. Also future multimedia implementations will adapt MPEG-21 schema, which means that upcoming media recorders must be able to automatically create video content indexing. This chapter will outline the various existing methods of boundary shot and scene change detection.
BACKGROUND A primitive typical approach to indexing video data was the manual creation of textual annotations along with time headers in the metadata of a media file. However, such a human-based method is time consuming and practically not applicable. Moreover, such methods suffer from the subjectivity of the human operator during the textual description. Therefore, it is necessary to develop an integrated framework for automatic extraction of the most character-
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Shot Boundary Detection Techniques for Video Sequences
istic frames of a video sequence, which will finally enable the efficient indexing and description of a video sequence. More specifically, by developing methods that enable the automatic build of a scene-access menu for a video clip, the viewer may use this index for quick access at a specific scene or for performing scene searches. Several approaches have been proposed in the literature for automatic video indexing, which can be basically categorized as methods for temporal segmentation in an uncompressed or compressed video domain (Koprinska & Carrato, 2001; Lienhart, 1999; Dailianas, Allen, & England, 1995). Thus, the various temporal video segmentation methods for each class (i.e., uncompressed/compressed) will be discussed in the following sections.
SHOT BOUNDARY DETECTION IN UNCOMPRESSED DOMAIN Video segmentation in an uncompressed domain includes all the boundary shot detection methods that perform using metrics and mathematical models on the uncompressed/spatial video signal. Most existing methods detect shot boundaries of video based on some change of the video content on the visual domain between consecutive frame pairs. If the measured change is above a predetermined threshold, then a shot boundary is assumed and reported. Based on the metrics nature that is used to detect the differences between successive frames, the algorithms can be generally classified into the following classes: pixel-based, block-based, and histogram-based (Zhang, Low, Gong, & Smoliar, 1994, 1995).
Pixel-Based Methods Pixel-based methods evaluate the differences in luminance or color domain between pixel values of successive frames (Kikukawa & Kawafuchi, 1992). Hence, a per pixel comparison is performed between frame pairs. Depending on the measured difference from the pixel-based comparison, a scene cut is detected and reported if the calculated difference is above a pre-defined threshold value. Otherwise no scene change is reported. The sensitivity and the efficiency of the pixel-based methods are strongly related to the selection of the reference threshold.
Block-Based Methods In contrast to the aforementioned pixel-based methods, where the whole frame of a video movie is taken under consideration for the scene change detection and the corresponding measured difference in the pixel values, either in color or 890
luminance domain, in block-based methods each frame is divided into blocks that in turn are compared to their corresponding blocks in the successive frame (Kasturi & Jain, 1991; Shahraray, 1995). More specifically, in contrast to the aforementioned pixel-based techniques, where the critical unit is the number of pixels whose difference is above a threshold value, these methods report a scene change, when the number of changed blocks is greater than a predefined threshold.
Histograms Comparisons The aforementioned categories exploit pixel comparison in order to derive a decision. On the contrary, histogram-based methods exploit the fact that a set of frames that belong in the same scene retain, in general unchanged, their luminance- or color-level histograms. A luminance- or color-level histogram of a frame depicts the density of the number of pixels that have specific luminance or color value. As has been described, the majority of the aforementioned methods are implemented based on metrics of the uncompressed video domain, utilizing a common framework: a similarity measurement between successive frames.
SHOT BOUNDARY DETECTION IN COMPRESSED DOMAIN Multimedia applications that distribute audiovisual content over communication networks (such as video-on-demand (VOD) and real-time entertainment streaming services) are based on digital encoding techniques (e.g., MPEG-1/2/4 and H.261/2/3 standards) that achieve high compression ratios by exploiting the spatial and temporal redundancy in video sequences. Most of the standards are based on motion estimation and compensation, using the block-based discrete cosine transformation (DCT). The use of transformation facilitates the exploitation in the compression technique of the various psychovisual redundancies by transforming the picture to a domain where different frequency ranges with dissimilar sensitivities at the human visual system (HVS) can be accessed independently. The DCT operates on a X block of N Χ N image samples or residual values after prediction and creates Y, an N Χ N block of coefficients. The action of the DCT can be described in terms of a transform matrix A. The forward DCT is given by: Y=AXAT
Shot Boundary Detection Techniques for Video Sequences
where X is a matrix of samples, Y is a matrix of coefficients, and A is an N Χ N transform matrix. The elements of A are: Aij = Ci cos
(2 j + 1)i 2N ,
•
where Ci =
1/ N , i= 0 2 / N , i> 0
(1)
Therefore the DCT can be written as: N −1 N −1
Yxy = Cx C y ∑∑ X ij cos i =0 j =0
(2 j + 1) y (2i + 1) x cos 2N 2N
(2)
Afterwards in the encoding chain, quantization of the aforementioned DCT coefficients is performed, which is the main reason for the quality degradation and the appearance of artifacts, like the ‘blockiness’ effect. Several methods for shot boundary detection in the compressed domain have been developed. According to Koprinska and Carrato (2001), they can be classified into the following categories, depending on the used metric: •
•
•
DCT Coefficients: The temporal video segmentation methods based on DCT coefficients apply a comparison technique to the DCT coefficients of the corresponding successive video frames. The difference metric is somewhat similar to the aforementioned pixel-based metric, where a scene change is detected and reported when the measured difference exceeds a specific threshold value (Zhang et al., 1994, 1995). It must be noted that these methods can be applied only on intra-coded frames of a DCT-based coded signal, because only they are fully encoded with DCT coefficients. Thus, the processing requirements may be low, but the temporal accuracy of the detected frame drops dramatically, and it is highly dependent on the intra-frame periodicity. DC Terms: The DC term is a scaled version of the average value for each block and thus the DC terms are directly related to the pixel domain. So, in a similar way to the uncompressed domain methods, the DC terms-based metrics measure the DC terms differences between successive frames. Again a frame is reported as shot boundary, if the aforementioned measurement is higher than a pre-defined threshold (Yeo & Liu, 1995). DC Terms, Macroblock (MB) Coding Mode: This is a hybrid method in which, except from the aforementioned DC terms, the type of the macroblock (MB)
•
coding is taken under consideration as well. When a scene change takes place, then some macroblocks of an inter-coded frame may be intra-coded due to limited reference options, demonstrating where scene change occurs (Meng, Juan, & Chang, 1995). MB Coding Mode and Motion Vectors (MVs): Similarly to the previously described method, a hybrid model is exploited, where it takes under consideration both the MB coding mode and the MV information of the encoded video sequence. MB Coding Mode and Bit Rate Information: Finally this method uses both bit rate information and motion-predicted MB types in order to derive accurate estimation of the scene changes. The forced intra-coding of some MBs over a scene change increases the deduced bit rate due to the inefficiency of the intracoding method.
Similarly to the shot change detection methods of the uncompressed domain, this section has shown that also the methods of the compressed domain exploit analogous frameworks at their implementation, which is based on the comparison of the calculated metric between successive frame pairs.
FUTURE TRENDS All the aforementioned methods use a threshold parameter in order to distinguish shot boundaries and changes. Thus, a common problem in shot boundary detection lies in the selection of an appropriate threshold for identifying whether a change is sufficiently large to signify a shot boundary or not (Lu & Tan, 2005). If a global threshold is used for the detection of shot boundaries over the whole video, then successful detection rate may vary up to 20% even for the same video content (O’Toole, Smeaton, Murphy, & Marlow, 1999). To improve the efficiency and eliminate this performance variation, some later works propose to use an adaptive threshold which can be dynamically determined based on video content (Lienhart, 1999; Dailianas et al., 1995). But even these methods require a lot of computational power in order to estimate successfully the appropriate threshold parameter, making their implementation a challenging issue, especially for real-time applications. Thus, the research community faces the challenge of developing new techniques and methods for detecting scene changes over a video signal by eliminating the necessity of threshold parameters in the decision process. Moreover, in order to allow a more efficient exploitation and browsing of video-anthologies, it is necessary to integrate these boundary shot detection techniques within content-based search and access methods, where the categorization of the video content is occurred by retrieving 891
S
Shot Boundary Detection Techniques for Video Sequences
highly representative optical and semantic information. In this respect, the combination of frame extraction techniques and semantics will help towards the evolution of the current Web to the Semantic Web, where the browsing and searching of information will be based on semantic information.
CONCLUSION This article outlines the various methods for detecting and extracting the scene changes from a video sequence. Depending on the metric that is exploited for the detection procedure, the methods that have been proposed are classified into two broad categories: those based on the uncompressed domain and those that exploit the metric of the compressed domain. Both the categories share the common drawback that they use threshold values for their decisions. Thus, the research community faces the challenge to develop new techniques that eliminate the use of threshold values, eliminating in this way the complexity and the computational requirements of the proposed methods.
ACKNOWLEDGMENTS
Lu, H., & Tan, Y-P. (2005). An effective post-refinement method for shot boundary detection. IEEE Transactions on Circuits and Systems for Video Technology, 15(11), 1407-1421. Meng, J., Juan, Y., & Chang, S.-F. (1995). Scene change detection in a MPEG compressed video sequence. Proceedings of the SPIE International Symposium on Electronic Imaging (Vol. 2417, pp. 14-25). San Jose, CA. O’Toole, C., Smeaton, A., Murphy, N., & Marlow, S. (1999). Evaluation of automatic shot boundary detection on a large video suite. Proceedings of the 2nd UK Conference on Image Retrieval: The Challenge of Image Retrieval (pp. 1-13). Newcastle, UK. Shahraray, B. (1995). Scene change detection and contentbased sampling of video sequences. Proceedings of SPIE (pp. 2-13). Tam, W. J., Stelmach, L., Wang, L., Lauzon, D., & Gray, P. (1995, February 6-8). Visual masking at video scene cuts. Proceedings of SPIE, 2411, (pp. 111-119). San Jose, CA. Yeo, B., & Liu, B. (1995). Rapid scene analysis on compressed video. IEEE Transactions on Circuits and Systems for Video Technology, 5(6), 533-544.
This article is carried out within the “PYTHAGORAS II” research framework, jointly funded by the European Union and the Hellenic Ministry of Education.
Zhang, H. J., Low, C. Y., Gong, Y. H., & Smoliar, S. W. (1994). Video parsing using compressed data. Proceedings of the SPIE Conference of Image and Video Processing II (pp.142-149).
REFERENCES
Zhang, H. J., Low, C. Y., Gong, Y. H., & Smoliar, S. W. (1995). Video parsing and browsing using compressed data. Multimedia Tools and Applications, 1, 89-111.
Dailianas, A., Allen, R. B., & England, P. (1995). Comparison of automatic video segmentation algorithms. Proceedings of SPIE (Vol. 2615, pp.2-16). Hampapur, A., Jain, R., & Weymouth, T. E. (1995). Production model based digital video segmentation. Multimedia Tools and Applications, 1(1), 9-46. Kasturi, R., & Jain, R. (1991). Dynamic vision (pp. 469-480). IEEE Computer Society Press. Kikukawa, T., & Kawafuchi, S. (1992). Development of an automatic summary editing system for the audio visual resources. Transactions on Electronics and Information, 75-A(2), 204-212. Koprinska, I., & Carrato, S. (2001). Temporal video segmentation: A survey. Signal Processing: Image Communication, 16, 477-500. Lienhart, R. (1999). Comparison of automatic shot boundary detection algorithms. Proceedings of SPIE (Vol. 3656, pp. 290-301). 892
KEY TERMS Bit Rate: A data rate expressed in bits per second. In video encoding the bit rate can be constant, which means that it retains a specific value for the whole encoding process, or variable, which means that it fluctuates around a specific value according to the content of the video signal. Frame: One of the many still images which as a sequence composes a video signal. Histogram: A luminance- or color-level histogram of a frame depicts the density of the number of pixels that have specific luminance or color value. Multimedia: The several different media types (e.g., text, audio, graphics, animation, video). Pixel: Considered the smallest sample of a digital image or video.
Shot Boundary Detection Techniques for Video Sequences
Shot: An unbroken sequence of frames taken continuously from a single camera.
S
Video Codec: The device or software that enables the compression/decompression of digital video. Video Coding: The process of compressing and decompressing a raw digital video sequence.
893
894 Category: Mobile Phone
Smartphone Acceptance among Sales Drivers Jengchung V. Chen National Cheng Kung University, Taiwan
INTRODUCTION The objective of this research is to find out the acceptance of sales drivers in logistic industry to use the smartphone in their work. This research uses two methods to collect data: survey and experiment. This research integrates technology acceptance model (TAM) (Davis, 1989), self-efficacy (Bandura, 1982, 1986), and innovation diffusion theory (IDT) (Rogers, 1962) into the research model to find out the factors of sales drivers in logistic industry accepting the smartphone. This experiment is focused on three user groups: the employees of a business that implements smartphone usage, the employees of businesses that do not use smartphone, and the students that are currently studying at a department of transportation and communication management. The results will help us understand whether this technology should be integrated into the traditional logistics system, and get to know the pros and cons of this idea.
BACKGROUND Freight businesses in the logistics industry perhaps have few examples of utilizing mobile devices thoroughly in their daily operations. Sales drivers who are responsible for distributing goods on time need to not only interactively exchange information with the headquarters, but also need to use spare time visiting their customers. Because of the needs to better serve their customers and other operational purposes, logistics businesses have their employees equipped with all sorts of devices like hand held terminal (HHT), bar code reader, GPS, and on board unit (OBU) to keep track of the goods. In addition, those drivers who for a long time have been considered as shippers now have another important role—salesperson.
phone. The cellular technology has evolved so drastically that phones with high-resolution digital cameras, voice recording and digital assistant are very familiar to most people. And to satisfy business users, powerful handheld devices that run the smartphone operating system have been developed. All these gadgets are here so that they can make peoples lives easier. Smartphones support many features that are really helpful in the business sector, for example, this cellular can be associated with OBU (on board unit) to create a more efficient delivery system for logistics companies. This technology has greatly benefited logistics companies since the implication of this technology took place. And more and more functions are being included for example portable bar code scanners and real-time communications systems that can update detailed information amongst the driver and the office at all times. Smartphones not only provide useful business oriented functions; they also provide useful functions such as calendars, task planning and even high-speed mobile Internet over the 3G network (Valletti & Cave, 2002; Ralph, 2002; Funk, 1998).
Different Categories of the Cellular Phones Cellular phones are separated into three categories; the categories are sorted by the limitations of their functions shown as follows: •
SMARTPHONES Since the invention of the telephone in 1876, peoples’ lifestyles have been changed drastically as time passes. Then, Martin Cooper introduced a whole new level of communication by using the concept of wireless technology, called the cellular phone. Since then, the cellular phone has been part of many people’s lives, and nowadays almost everyone owns a cellular
•
PDA Phones: These kinds of phones support all the functions that a PDA can do, it is actually a whole PDA integrated into the phone. PalmOS, Symbian and Windows CE are examples of operating systems that are used in PDA phones. More and more software developers are developing operating systems that are becoming more powerful (e.g., Linux). Most of these PDA phones are even able to read and edit Word, Excel and even PowerPoint files, which is really convenient for business users who don’t like to carry their computers around. PIM Phones: Known as the Personal Information Manager, it could also support features such as Outlook synchronization with a personal computer, but the functions are more limited compared to the PDA phones. PIM phones use a closed operating system; they do not support as many applications like the PDA phones.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Smartphone Acceptance among Sales Drivers
Figure 1. 2000-2007 shipment of smartphones (IDC) 90, 000
S
Shipment (K)
80, 000 70, 000
80607.06
60, 000
2002-2007 CAGR 85.9%
50, 000
62627.84
40, 000
45290.45
30, 000 20, 000 10, 000
3629.51 109.51 1398.19
0 2000
•
2001
2002
27595.42 13081.66 2003
Cellular Phones: This is the most basic form of wireless phones, which usually offer simple address book functions, messaging, GPRS, WAP, MMS, and video calling. Some of these phones might support Java applets, but the functions are still very limited.
Operating System There are currently many operating systems for mobile phones on the market, namely Linux, Windows Mobile, Symbian and Palm. Microsoft Windows Mobile and Symbian are most commonly used; this is because of the ease of use and the high compatibility of applications. Some of the operating systems’ source codes for mobile phones are open for software developers to use, this can allow more software applications to be developed and allow a higher usage of mobile Smartphones (Gruber & Verboven, 2000; Harrison & Holley, 2001). Differences between the different mobile phone operating systems are shown as follows: •
•
•
2004
Windows Mobile is separated into three different categories, namely Pocket PC, Pocket PC Phone edition and Smartphone; these are all developed by Microsoft. All of the three operating systems are very powerful and can support vast amounts of applications just like the PDA and even some PC applications. Symbian operating system is very commonly used because of the ease of use of the system. It also has integration of other software-developing companies that are continuously developing new applications and thinking out new ideas to improve it. Linux is well-known for being free; this makes no exception for Linux on mobile phones. This operating system’s source code is open to the public; it allows any
•
2005
2006
2007
kind of alteration to the system itself and allows any developer to develop applications, all for free. Linux also turns out to be one of the most stable operating systems on the mobile phone market and the PC market too. But the problem with this operating system is the low support for applications. Palm OS has very high usage in the PDA market; they have successfully integrated the PDA technology with mobile phone technology. They are famous for their highly efficient input method by using the touch screen, and the high support of applications. Most applications that can be used on a Palm OS PDA can be used on a mobile phone running Palm OS for mobile phones.
Because of the rapid growth of smartphones’ technology, more and more people are switching from a PIM or a normal cellular phone to using a smartphone, because they realize that it really can make a difference in their busy life. Smartphone-based technology has also been integrated into many logistic systems. From the data gathered from the IDC (International Data Collection), it shows that the number of Smartphones shipped is growing annually, and it also has been forecasted that this trend will continue to grow.
LOGISTICS In the old days the word logistics had to do with the military’s operations, it mainly dealt with procurement, distribution, maintenance and replacement of material and personnel. Nowadays it mainly has to do with the flow of material from one place to another, used mainly in the transportation industry. Logistics operation can include many other functions such as warehousing, packaging and other information based functions.
895
Smartphone Acceptance among Sales Drivers
Figure 2. The composition of logistics Logistics activities:
Logistics composition:
Warehousing, packaging, transportation related functions, information related function
Manufacturers, Ports, transportation industry, market including consumers
Logistics
Logistics can be separated into four main categories: •
Manufacturing Logistics: Logistics activities starting from the manufacturer’s location to the location of the market. Sales Logistics: Logistics activities from the market to the consumer. Procurement Activities: Logistics activities based on purchasing of raw materials or other products. Recycling Logistics: Logistics activities based on returned merchandise.
• • •
Examples of motor vehicle logistics: • • • • • •
Door to door service Post office service 3rd party logistics service Specialized transportation service Transportation of dangerous or abnormal goods Transportation rental service
As seen from the above, motor vehicles still play an important role in logistics, motor vehicles can access almost any hard to reach area, and it does not need any kind of facility (e.g., railroad, airport, port etc.) to enable the usage. It also can carry a tremendous amount of cargo.
logistics company to track your parcel status. GPS also can navigate the driver to the shortest route available with realtime data; this can save a lot of time because drivers are not always familiar with the area that they are at.
SETTINGS FOR THE EXPERIMENT This experiment is based on statistical analysis of data that is gathered through a questionnaire. The questionnaire is based on data from logistics companies that have already integrated mobile smartphone technology into their systems. The study is to investigate a case of a logistics company in the freight business; it is one of the major logistics companies in Taiwan that provides overnight delivery service of parcels and mail. They have already integrated smartphone and GPS technology into their delivery system. Bar codes are scanned at every point in the delivery process, and the data scanned is sent through the GSM network to the main control center; therefore a customer’s area allows one to track the status of the objects that are sent. GPS is used to navigate the driver, and to avoid delays caused by heavy traffic zones. GPS can also get the driver to the right location without getting lost.
ANALYSIS OF RESULTS SMARTPHONE AND LOGISTICS
The study collected 30 samples for this analysis:
To achieve maximum efficiency in logistics, it is recommended that smartphone technology be integrated. The use of modern technology can be really helpful in logistics service: bar code usage, GPS tracking and GPS navigation facilities have been of great help to the logistics industry in the recent years. These technologies allow customers to use logistics services with more confidence, because they know every detail that they need to know about the parcel. This is achieved by combining bar code technology with real-time GPS tracking. Drivers no longer need to keep calling the
•
896
• •
10 employees from logistics companies that use smartphone 9 employees from logistics companies that do not use smartphone 11 students from a transportation and communications management major
The entire groups’ ages range between 19 and 50 with an average of 31. Level of education is from high school or higher.
Smartphone Acceptance among Sales Drivers
T-Type Analysis Statistic t-test is used to analyze the data collected and it is found that only two hypotheses had positive results. After the experiment of function comparison, the VOIP function is more important compared to GPS and barcode reading; barcode has been shown to be very important too. Five hypotheses are proposed, and only one is rejected. The only one that is rejected is the GPS function; they do not think that this is a very important function. But the result shows us that there is a very high level of acceptance of students, and statistics shows that logistics companies might be implementing more modern technology in the future.
FUTUTE TRENDS The demand for smartphone technology will keep increasing in the future as the technology matures. The acceptance of smartphones in the transportation industry is growing, mainly because it can greatly increase the efficiency of delivery and customer relationships. Later models of smartphones bind many functions together so that you don’t need to carry a lot of different equipment for different functions; for example handhelds units have GPS, GPRS and some even have bar code reading integrated in just one handheld PDA device. And the prices of these products are dropping gradually making it more affordable and increasing the will for a company to use these products.
CONCLUSION From the results of the experiment, it shows that: 1. 2. 3. 4.
VOIP function shows more acceptance than other functions such as calendar functions. GPS function is more important than the presentation function; navigation can allow drivers to avoid a lot of problems on the road. VOIP is better than the messaging function; voice communication is direct and instantly responsive unlike messages. GPS is more important than the barcode reading function. The barcode reading system has to be integrated with other technology to be able to function, which could sometimes be inconvenient. GPS shows more importance.
It is also identified the different levels of importance between different user groups:
1. 2. 3.
GPS is considered useful for logistics companies; with this technology anyone is able to deliver products even if they are unfamiliar with the area. Barcode is very important to companies that are currently using it. A majority of companies that are not using it yet show that they are willing to try it out. VOIP has shown to be important to everyone, because this function can allow companies to save a tremendous amount of money on phone bills.
The utilizing of smartphones can benefit all parties by an increase the efficiency, which is one of the most important factors in the transportation industry. As smartphone technology grows, users outside the industry can also make use of smartphone devices in their daily lives, for example, GPS can be installed in all kinds of vehicles even on some bicycles. One reason for the rapid growth of technology is the high demand for new ideas, there is new technology designed for everyone, and making good use of it can improve our quality of life. More academic theories should be surveyed to investigate the sales drivers’ intention to use such the new technology. Take TAM for example, it excludes Hartwick and Barki’s (1994) findings, which says subjective regulation impacts to the new technology acceptance. Other researchers proposed many more factors to be included in the TAM model. Future studies should also take these factors in to account: social influence processes and cognitive instrumental processes (Venkatesh & Davis, 2000), trust (Gefen, 2000), task-technology fit model (Dishaw & Strong, 1999), perceived characteristics of innovation (Plouffe et al., 2001), social influence and behavior control (Taylor & Todd, 1995).
REFERENCES Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147. Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice Hall. Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319-340. Dishaw, M. T., & Strong, D. M. (1999). Extending the technology acceptance model with task-technology fit constructs. Information & Management, 36(1), 9-21. Funk, J. L. (1998). Competition between regional standards and the success and failure of firms in the world-wide mobile communication market. Telecommunication Policy, 22(4), 419-441.
897
S
Smartphone Acceptance among Sales Drivers
Gefen, D. (2000). E-commerce: The role of familiarity and trust. Omega: The International Journal of Management Science, 28(6), 725-737. Gruber, H., & Verboven, F. (2000). The diffusion of mobile telecommunications service in the European Union. European Economic Review, 45, 557-558. Hartwick, J., & Barki, H. (1994). Explaining the role of user participation in information system use. Management Science, 40(4), 440-465. Harrison, F., & Holley, K. A. (2001). The development of mobile is critically development on standards. BT Technology Journal, 19(1), 32-37. Plouffe, C. R., Hulland, J. S., & Vandenbosch, M. (2001). Research report: Richness versus parsimony in modeling technology adoption decisions - understanding merchant adoption of a smart card-based payment system. Information Systems Research, 12(2), 208-222.
KEY TERMS Barcode: A series of vertical bars of varying widths, in which each of the digits zero through nine are represented by a different pattern of bars that can be read by a laser scanner. Global Positioning System (GPS): A system of satellites, computers, and receivers that is able to determine the latitude and longitude of a receiver on Earth by calculating the time difference for signals from different satellites to reach the receiver. Logistics: Operations that deal with the procurement, distribution, maintenance, and replacement of material and personnel. On Board Unit (OBU): Portable electronic device similar to a PDA.
Ralph, D. (2002). 3G and beyond—The applications generation. BT Technology Journal, 20(1), 22-28.
Operating System: Software designed to control the hardware of a specific data-processing system in order to allow users and application programs to make use of it.
Rogers, E.M. (1962). Diffusion of innovations. NY: Free Press.
Smartphones: Mobile phone that had PDA functions integrated into it.
Taylor, S., & Todd, P. (1995). Assessing IT usage the role of prior experience. MIS Quarterly, 19(4), 561-570.
3G: Wireless technology that provides high-speed data transfer and portable video phone call service.
Valletti, T. M., & Cave, M. (2002). Competitive in UK mobile communications. Telecommunication Policy, 22(2), 109-131.
Voiceover Internet Protocol (VoIP): Voice communication that is done through Internet: usually can reduce costs of international calling.
Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186-204.
898
Category: M-Learning 899
SMS-Based Mobile Learning
S
Krassie Petrova Auckland University of Technology, New Zealand
INTRODUCTION Students today combine study and work and expect significant cost and time savings from the use of information and communication technologies, including mobile communication. A strong interest in implementing mobile technologies in learning has emerged. Experiments with one of the most popular technologies—text messaging—have been reported in the literature (e.g., Stone & Briggs, 2002; Finn, 2004) with some including the development of blended learning models (Stone, Briggs, & Smith, 2002). An early definition of mobile learning (m-learning) as “learning through mobile computational devices” can be found in Quinn (2000). Later, frameworks and research models were developed, providing guidelines for implementing suitable pedagogical approaches and for building services and applications relevant to a variety of mobile platforms and contextual settings (Garner, Francis, & Wales, 2002; Seng & Lin, 2004; Berth; 2005; Brown, 2005). SMS (short message service, or text messaging) is an extremely popular and still growing 2G mobile data service, especially with young adults (Finn, 2004; Prensky, 2005; MMA, 2006; Grinter & Eldridge, 2003), which makes it suitable as a learning technology. This short article presents and illustrates the concepts of SMS-enabled m-learning, describing a series of SMS learning scenarios derived from the literature. The defining features of the scenarios are identified and discussed, including future trends.
BACKGROUND Mobile learning is often referred to as a type of e-learning (Vavoula & Sharples, 2002; Leung & Chan, 2003; Seng & Lin, 2004). For the purposes of this article mobile learning is defined as a form of e-learning, which can take place anytime and anywhere through the use of a wireless and mobile communication device and the related network technology (Brown, 2005; Kukulska-Hulme, Evans, & Traxler, 2005; Wagner, 2005; Petrova, 2007). An SMS scenario can be defined as a self-contained learning experience focused on a group of participants who act in a specific context and perform specific tasks to achieve knowledge acquisition oriented goals using the SMS mobile technology (Petrova & Sutedjo, 2004; Evans & Taylor, 2004). As text messaging is enabled on all types
Figure 1. An SMS-based learning scenario framework Design & Pedagogy
Participants
Architecture
of 2G and 3G mobile phones, an SMS-learning scenario will be accessible to virtually any mobile phone user. The framework in Figure 1 captures the main aspects from which researchers have described and evaluated m-learning, including SMS-based learning (Trifonova, 2003; Attewell, 2005; Riordan & Traxler, 2005; Silander & Rytkonen, 2005; Chinnery, 2006).
Architecture Architecture deals with the specific mobile platform or platforms developed and used within a mobile learning scenario. The basic architecture would include access to the SMS provider network, an SMS-enabled cell phone and an SMS server or gateway (Petrova, 2007; Capuano, Gaeta, Miranda, & Pappacena, 2004). It might also include a number of auxiliary servers, such as a Web site, used to send bulk SMS messages and/or to provide instructions for participants. In some cases, additional infrastructure is needed to support a scenario where SMS is used in conjunction with another mobile technology (e.g., WAP), or a scenario where SMS learning is integrated with another e-learning approach to become part of a blended learning model.
Design and Pedagogy Design and pedagogy describes the context for which the scenario was designed and developed, including its activities and expected learning outcomes. The framework in Figure 2 presents the general design contexts and the pedagogical aspects of m-learning.
Participants The participants might be learners (university students, adult learners, or the general public) and teachers (faculty,
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
SMS-Based Mobile Learning
Figure 2. SMS learning scenarios: Design contexts and pedagogical aspects SMS Scenarios: Design Contexts (Roibas, 2002; Bollen, Eimler, & Hoppe, 2004; Kadirire, 2005; Pincas, 2004; Colley & Stead, 2003; McMillan & Keough, 2005) •
Supporting both independent & collaborative learning
•
Supporting “just-in time learning”
•
Supporting content delivery in a condensed format
•
Supporting multiple learners’ learning styles
administrators). Participants’ background, perceptions, attitudes and priorities play a critical role in the successful adoption of a scenario where they have a stakeholder role (Barker, Krull, & Mallinson, 2005).
TYPES OF SMS-BASED SCENARIOS FOR MOBILE LEARNING Scenario descriptions were extracted from the literature on mobile learning (2002-2005). These include cases from Europe (UK, Ireland, Greece, Germany, Finland), Asia (Japan, Thailand, Malaysia), Africa (South Africa), Australia and New Zealand. The scenarios were classified based on their context, content, and orientation (Figure 3). Two major categories were identified, each comprising five scenario types: learning (delivery of new content, test and quizzes, learning for revision, simulation-based learning, collaborative learning), and learning support (student support, communication, teacher support, blogs, Q&A sessions). Examples that illustrate each type are presented as follows, providing details about the participants, the educational setting, and the required additional technology. Figure 3. Types of SMS learning scenarios
900
Learning
Learning support
Delivery of new content
Student support
Tests and quizzes
Communication
Learning for revision
Teacher support
Simulation based learning
Blogs
Collaborative learning
Q & A sessions
SMS scenarios: Pedagogical Aspects (Pincas, 2004; Singh, 2003) •
The urgency of user needs
•
The ownership of initiative
•
The mobility of setting
•
The interactivity of process
•
The situated-ness of needs
•
The integration of content
Delivery of New Content All examples in this category refer to learning a foreign language. The participants are learners—there is no interaction with teachers. The SMS server needs to be able to handle registrations and store content. The InLET project (Pincas, 2004) provided language support for international tourists attending the Olympic games in Athens. Tourists who subscribed to the service received SMS messages with short phrases in Greek. They could also request and receive a translation from or into Greek of a commonly used phrase. Another example is the structured short course in English delivered to working learners in Hong Kong (Song & Fox, 2005). New words and expressions were sent to learners on a regular basis, following a predefined sequence of learning tasks. The course material was also available on the Web. Similar scenarios were implemented in Japan (Thorton & Houser, 2005) and in Australia (Levy & Kennedy, 2005). Commercial mobile services for language support are also available (Chinnery, 2006). Munro (2005) describes a commercial application, which delivers a pair of learning objects to a paid customer: a text-based object using SMS, and a sound object via podcast to an MP3 player or a smartphone.
Tests and Quizzes Tests and quizzes are used for formal learning and for selfassessment. Tests are typically conducted in controlled conditions. Quizzes are used in class or as a supplementary homework activity. Receiving feedback (“the score”) is an important feature of a test or a quiz. Tretiakov and Kinshuk (2005) describe a scenario where students in class were given a quiz and then had to submit
SMS-Based Mobile Learning
their answer via SMS. The quiz question could be “multiple choice,” “fill in the blank” or “matching lists.” Students could work individually or share a mobile phone to send the answer as a group. Individual feedback was accessible on a Web site. Very similar is the setting described in Iliescu and Hines (2005), which also expanded the range of question types to include free text questions. The example found in Capuano et al. (2004) offers functionality similar to the scenarios above. However there is no need for an additional platform (the Web), as all communication is based on SMS: the question text is sent to the student who responds with the answer and receives feedback. In the example the SMS service is part of a larger, multi-channel, integrated platform. Finally, a formal testing experiment is reported in Whattananarong (2004). Students sent test answers via SMS after being shown or read the questions in the classroom. There is immediate feedback.
Learning for Revision Mellow (2005) describes the study platform StudyTXT, which was implemented to facilitate a “flash card” scenario at a New Zealand university: students studying sports medicine could download content relevant to a topic of their choice. They could see the list of available topics on a Web site. A similar use of the same platform is proposed in Petrova (2007). A different approach is adopted in the scenario described in Riordan and Traxler (2003): rather than being given the option to request a revision snippet, students identified as being at risk of failing a test were sent revision tips and directions for further study by their lecturers.
Simulation-Based Learning Cheung (2004) describes a series of classroom experiments involving simulation gaming in a postgraduate microeconomics class. Students were given the game plan and had to submit responses related to the game via SMS. All responses were transmitted from the mobile network over the Internet to the teacher at their workstation, who then used specialised software to generate individualised return messages. These were broadcast back to students, simulating direct interaction among students. Another participatory simulation game is described in Lonsdale, Baber, and Sharples (2004). Participants were involved in a role play (as “water droplets”) and were sent “entry” and “exit” messages (commands) directing them to perform certain actions in the physical classroom environment. As the simulation game was run under Java, the text messages were sent from a mobile phone connected to the PC running the game.
Collaborative Learning Bollen, Eimler, and Hoppe (2004) describe a scenario for a constructive discussion implementing a decision support system (“Cool Modes”) in a literature class. Students were assigned specific roles and sent text messages related to the role from a PDA (mimicking SMS). The messages were delivered via the internal wireless LAN to a MySQL database on the teacher’s PC. The teacher could query the database and display the discussion threads to the class on an electronic whiteboard. A similar scenario is presented in Kadirire (2005): while listening to a presenter, students used mobile phones to text their comments which were stored in a database and then subsequently organised and formatted to be displayed to the class. The presenter could give feedback based on the comments.
Student Support In most instances of this scenario (Mohhamad & Norhayati, 2003; Stone, 2004a; 2004b; Nonyongo, Mabusela, & Monene, 2005), assistance is offered to a class or a cohort of students informing them about changes in schedules, or reminding them about deadlines and other events. A Web accessible database might be used to register users for the service.
Communication Seppala and Alamaki (2002) describe an integrated scenario where voice communication, SMS messaging and a WAP gateway were used to create a media-rich environment for communications between instructors and trainee teachers. Text messages were used to transmit general information among all participants.
Teacher Support Silander and Rytkonen (2005) provide an example for meeting teachers’ pedagogical needs. SMS was used to assist lecturers as part of an intelligent tutoring system (Alykko). Text messages were stored in a Web accessible format and were used to record learning logs (similar to blogs), to send instructions to students and to respond to questions (similar to Q&A sessions).
Blogs The blog scenario presented in Divitini, Haugalokken, and Morken (2005) included SMS as part of a learning management system (LMS). It enabled students doing their teaching practice to upload entries to a shared community blog and
901
S
SMS-Based Mobile Learning
to individual student blogs. The blogs were accessible for viewing on a Web server.
•
Questions and Answers (Q&A) This scenario involves a general tool available to all students on and off-campus, which collects questions and displays them on a Web site (anonimized). Answers from staff are also displayed (Ng’ambi, 2005).
SMS SCENARIOS: SUMMARY Applying the design contexts framework (Figure 2, left-hand column) it can be concluded that: •
• •
•
SMS allows for the development of both independent and collaborative learning models. In a significant number of examples a mixed model is implemented—where all participants can benefit from individual experiences Multiple learning styles might be supported in cases where SMS is integrated into a larger LMS or VLE (virtual learning environment). As a technology, SMS is especially well suited for “just-in-time learning;” however, if integrated with other technologies, technology access issues might interfere with the process (e.g., learners on the move might have access to a mobile phone but not to the Web). SMS content always needs to be condensed due to the limitations of the technology (160 characters in one text message).
Analysing the pedagogical aspects of SMS learning (Figure 2, right-hand column) it can be seen that: •
•
• •
902
Learner support scenarios are especially well tailored to satisfy urgent needs as most are based on a pull approach (the learner owns the initiative). Some of the learning scenarios are also pull-based (e.g., the flash card scenario) and rely on the learner to realise their need and engage in an m-learning activity. While the initiative in some cases belongs entirely to the teacher (e.g., sending tips to “weak” students), a group of scenarios supports both push and pull approaches (e.g., scenarios which augment classroom activities.) All scenarios except one use mobile and wireless networks. Most scenarios are highly interactive, often using auxiliary infrastructure (e.g., for displaying results, or for providing detailed instructions).
•
The enhancement of SMS through adding other technologies to the interactions interferes with the mobility of the learner (e.g., a mobile learner might not have mobile access to all technologies such as the Web, involved in an integrated scenario). The cost of integration needs to be balanced with the benefits derived from the use of other communication channels. In most scenarios the level of integration between content and situation is very high (e.g., discussions, Q&A sessions, simulations).
FUTURE TRENDS SMS has already evolved into enhanced services, which are capable of transferring images and animation (EMS, MMS). Multimedia messaging offers an integrated platform and thus eliminates the need of embedding SMS into environments based on other technologies. MMS and EMS allow the learner to be truly mobile, as users with smartphones can have access to the mobile Internet and to the Web. Some of the SMS scenarios may migrate to MMS (one example is the study of a foreign language). New types of personal learning are expected to evolve around these technologies, and pedagogical models will need to be developed and evaluated alongside with the development of specially organised and formatted content. However, SMS scenarios will continue to be attractive in environments where cost of using more sophisticated or advanced technologies might be prohibitive.
CONCLUSION This article extensively reviews the literature on mobile learning models using SMS. Based on the results of the review, it classifies SMS-learning scenarios into two major categories and provides examples to illustrate the types identified under each category. A framework of design contexts and pedagogical aspects is used to summarise the findings. Future trends including multimedia and MMS scenarios are also discussed.
REFRENCES Attewell, J. (2005, October). From research and development to mobile learning: Tools for education and training providers and their learners. In Proceedings of the 4th World Conference on Mobile Learning (MLEARN05). Retrieved March 23, 2006, from www.mlearn.org.za/CD/papers/Attewell.pdf
SMS-Based Mobile Learning
Barker, A., Krull, G., & Mallinson, B. (2005). A proposed theoretical model for m-learning adoption in developing countries. In Proceedings of the 4th World Conference on mLearning (Paper 14). Berth, M. (2005). Adaptive ethnography: Methodologies for the study of mobile learning in youth culture. Paper presented at Seeing, Learning, Understanding the Mobile Age, Budapest, Hungary. Retrieved January 20, 2006, from http://www.fil.hu/mobil/2005/Berth.pdf Bollen, L., Eimler, S., & Hoppe, U. (2004). SMS-based discussions- technology enhanced collaboration for a literature course. In J. Roschelle, T.-W. Chan, Kinshuk, & S. J. H. Yang (Eds.), Proceedings of the Second International Workshop on Wireless and Mobile Technologies in Education (pp. 209-210). Brown, T. (2005). Towards a model for m-learning in Africa. International Journal on E-Learning, 4(3), 299-315. Capuano, N., Gaetta, M., Miranda, S., & Pappacena , L. (2004). A system for adaptive platform independent mobile learning. In J. Attewell & C. Smith (Eds.), Mobile Learning Anytime Everywhere: A Book of Papers from MLEARN 2004 (pp. 53-56). Cheung, S. (2004). Fun and games with mobile phones: SMS messaging in microeconomic experiments. In R. Atkinson, C. McBeath, D. Jonas-Dwyer, & R. Phillips (Eds.), Beyond the Comfort Zone: Proceedings of the 21st Australasian Society for Computers in Learning in Tertiary Education (pp. 180-183). Chinnery, G. (2006). Going to the MALL: Mobile assisted language learning. Learning Language & Technology, 10(1), 9-16 Colley, J., & Stead, G. (2003). Take a bite: Producing learning materials for mobile devices. In J. Attewell & C. Smith (Eds.), Learning with Mobile Devices Research and Development: A Book of Papers from the 2nd World Conference on Mobile Learning (MLEARN 2003) (pp. 43-46). Divitini, M., Haugalokken, O., & Morken, E.M. (2005). Blog to support learning in the field: Lessons learned from a fiasco. In Proceedings of the Fifth International Conference on Advanced Learning Technologies (pp. 219-221). Evans, D., & Taylor, J. (2004). The role of user scenarios as the central piece of the development jigsaw puzzle. In J. Attewell & C. Smith (Eds.), Mobile Learning Anytime Everywhere: A Book of Papers from the 3rd World Conference on Mobile Learning (MLEARN 2004) (pp. 63-66). Finn, M. (2004). The handheld classroom: Educational implications of mobile computing. Australian Journal of Emerging Technologies and Society, 2(10), 21-35.
Garner, I., Francis, J., & Wales, K. (2002). An evaluation of the implementation of a short messaging system (SMS) to support undergraduate students. In Proceedings of the European Workshop on Mobile and Contextual Learning (pp. 15-18). Birmingham, UK. Grinter, R. E., & Eldridge, M. (2003). Wan2tlk?: Everyday text messaging. In Proceedings of the Conference on Human Factors in Computing Systems (pp. 441-448). Iliescu, D., & Hines, E. (2005). WES: The SMS based student feedback, voting and notification system. Interactions 59(1), Article 2. Retrieved January 3, 2006, from http://www2. warwick.ac.uk/services/cap/resources/interactions/archive/ issue25/iliescu/ Kadirire, J. (2005). The short message service (SMS) for schools/conferences. Recent Research Developments in Learning Technologies, 2, 856-859 Kukulska-Hulme, A., Evans, D., & Traxler, J. (2005). Landscape study and mobile learning in the post16 sector: Summary report. Retrieved January 23, 2006, from http://www.jisc.ac.uk/uploaded_documents/ SUMMARY%20FINAL%202005.doc Leung, C.-H., & Chan, Y.-Y. (2003). Mobile learning: A new paradigm in electronic learning. In Proceedings of the 3rd International Conference on Advanced Learning Technologies (pp. 76-80). Levy, M., & Kennedy, C. (2005). Learning Italian via mobile SMS. In A. Kukulska -Hulme & J. Traxler (Eds.), Mobile Learning: A Handbook for Educators and Trainers. London: Taylor & Francis. Lonsdale, P., Baber, C., & Sharples, M. (2004). Engaging learners with everyday technology: A participatory simulation using mobile phones. In S. Brewster & M. D. Dunlop (Eds.), Proceedings of the 6th International Mobile Human-Computer Interaction Symposium (pp. 461-465). McMillan, J., & Keough, M. (2005). Seven reasons why mLearning doesn’t work. In Proceedings of the 4th World Conference on mLearning (Paper 44). Mellow, P. (2005). The media generation: Maximise learning by getting mobile. In Proceedings of the 2005 Conference of the Australasian Association for Computers in Learning in Tertiary Education (pp. 469-476). MMA. (2006). Portio research: Mobile messaging futures 2005-2010. Mobile Marketing Association. Retrieved January 23, 2006, from http://mmaglobal.com/modules/wfsection/article.php?articleid=71 Mohhamad, M. A., & Norhayati, A. (2003). A short message service for campus-wide information delivery. In Proceedings 903
S
SMS-Based Mobile Learning
of the Fourth National Conference on Telecommunication Technology (pp. 216-221). Malaysia. Munro, A. (2005). 5th digit language application. In Proceedings of the 4th World Conference on mLearning (Paper 49). Ng’ambi, D. (2005). Mobile dynamic frequently asked questions (m-DFAQ) for student and learning support. In Proceedings of the 4th World Conference on mLearning (Paper 51). Nonyongo, E., Mabusela, K., & Monene, V. (2005). Effectiveness of SMS communication between university and students. In Proceedings of the 4th World Conference on mLearning (Paper 53). Petrova, K. (2007). Mobile learning as a mobile business application. International Journal of Innovation and Learning, 4(1), 1-13. Petrova, K., & Sutedjo, Y. (2004). Just-in-time learning: Ready for SMS? In Kinshuk, D. Samson, & P. Isaias (Eds.), Proceedings of the International Conference on Cognition and Exploratory Learning in the Digital Age (pp. 495498). Pincas, A. (2004). Approaches to just-in-time learning with mobile phones: A case study of support for tourists’ language needs. In J. Attewell & C. Smith (Eds.), Mobile Learning Anytime Everywhere: A Book of Papers from the 3rd World Conference on Mobile Learning (MLEARN 2004) (pp. 157-162). Prensky, M. (2005). What can you learn from a cell phone? Anything! Innovate, 1(5). Retrieved January 4, 2006, from http:// www.innovateonline.info/index.php?view=article&id=83 Quinn, C. (2000, Fall). mlearning: Mobile, wireless and in-your-pocket learning. LineZine Magazine. Retrieved January 23, 2006, from http://www.linezine.com/2.1/features/cqmmwiyp.htm Riordan, B., & Traxler, J. (2003). Supporting computing students at risk using blended technologies. Paper presented at 4th Annual Conference of the Learning and Teaching Support Network for Information and Computer Sciences, Galway, Ireland. Riordan, B., & Traxler, J. (2005). The use of targeted bulk SMS texting to enhance student support, inclusion and retention. In Proceedings of the 2005 International Workshop on Wireless and Mobile Technologies in Education (pp. 257-260). Roibas, A. C. (2002). Designing scenarios of m-learning. Paper presented at the Knowledge Management Workshop 2002, Multimedia University, Malaysia. 904
Seng, J.-L., & Lin, S. (2004). A mobility and knowledgecentric e-learning application design method. International Journal of Innovation and Learning, 1(3), 293-311. Seppala, P., & Alamaki, H. (2002). Mobile learning and mobility in teacher training. In M. Milrad, U. Hoppe, & Kinshuk (Eds.), Proceedings of the 2002 International Workshop on Mobile technologies in Education (pp. 130-135). Silander, P., & Rytkonen, A. (2005). An intelligent mobile tutoring tool enabling individualization of students’ learning processes. In Proceedings of the 4th World Conference on mLearning (Paper 59). Singh, H. (2003). Leveraging mobile and wireless Internet. Retrieved January 3, 2006, from http://www.learningcircuits. org/2003/sep2003/singh.htm Song, Y., & Fox, R. (2005). Integrating m-technology into Web-based ESL vocabulary learning for working adult learners. In Proceedings of the 2005 International Workshop on Wireless and Mobile Technologies in Education (pp. 154-163). Stone, A. (2004a). Mobile scaffolding: An experiment in using SMS text messaging to support first year university students. In K. Kinshuk et al. (Eds.), Proceedings of the Fourth International Conference on Advanced Learning Technologies (pp. 405-409). Stone, A. (2004b). Blended learning, mobility and retention: Supporting first-year university students with appropriate technology. In J. Attewell & C. Smith (Eds.), Mobile Learning Anytime Everywhere: A Book of Papers from the 3rd World Conference on Mobile Learning (MLEARN 2004) (pp. 183-185). Stone, A., & Briggs, J. (2002). ITZ GD 2 TXT: How to use SMS effectively in m-learning. Proceedings of the European Workshop on Mobile and Contextual Learning (pp. 11-14). Stone, A., Briggs, J., & Smith, C. (2002). SMS and interactivity: Some results from the field and its implication on effective use of mobile telephony for education. In M. Milrad, U. Hoppe, & Kinshuk (Eds.), Proceedings of the 2002 International Workshop on Mobile Technologies in Education (pp. 147-151). Thorton, P., & Houser, C. (2005). Using mobile phones in English education in Japan. Journal of Computer Assisted Learning, 21(3), 217-228. Tretiakov, A., & Kinshuk, K. (2005). Creating a pervasive testing environment by using SMS messaging. In Proceedings of the 2005 International Workshop on Wireless and Mobile technologies (pp. 62-66).
SMS-Based Mobile Learning
Trifonova, A. (2003). Mobile learning: Review of the literature. University of Trento. Retrieved December 15, 2005, from http://eprints.biblio.unitn.it/archive/00000359/01/009. pdf Vavoula, N., & Sharples, M. (2002). KLeOS: A personal, mobile, knowledge and learning organisation system. In M. Milrad, U. Hoppe, & Kinshuk (Eds.), Proceedings of the 2002 International Workshop on Mobile Technologies in Education (pp. 152-156). Wagner, E. D. (2005). Enabling mobile learning. Educause Review, 40(3), 40-53. Whattananarong, K. (2004, September). An experiment in the use of mobile phones for testing. Paper presented at the International Conference on Making Educational Reform happen: Learning from the Asian Experience and Comparative Perspectives, Bangkok, Thailand.
Blogging: Periodic publishing on the Web in a specially designed space for collaborative writing (Web log). E-Learning: Learning facilitated by information and communication technologies (“electronic learning”). EMS: Enhanced message service. Flash Card Learning: A method of learning by memorising, often used in studying disciplines such as medicine or law. A flash card has two parts—a question part and an answer part. Just-In-Time Learning: A learning model where learners acquire knowledge as the need arises LMS: Learning management system MMS: Multimedia message service Smartphone: A mobile phone, which can connect to the Web via a browser and support email.
KEY TERMS
SMS: Short message service
Blended Learning: A learning paradigm where multiple pedagogical approaches are used.
WAP: Wireless application protocol VLE: Virtual learning environment
905
S
906 Category: M-Business and M-Commerce
Snapshot Assessment of Asia Pacific BWA Business Scenario Chin Chin Wong British Telecommunications (Asian Research Center), Malaysia Chor Min Tan British Telecommunications (Asian Research Centre), Malaysia Pang Leang Hiew British Telecommunications (Asian Research Center), Malaysia
INTRODUCTION The world is moving forward at a soaring rate. Within this change wireless technology is rapidly evolving and is playing an increasing role in the lives of people throughout the world. The people of today demand hassle-free and compact products, which can be used at anytime and anywhere with “always-best-connected” network solutions (Wong, Tan, & Hiew, 2005b). Wireless technology is a possible solution to meeting the immediate needs of society in the case of high-speed data delivery. This article is devoted to assessing the deployment of wireless networks in the Asia-Pacific region, with special focus on the existing Wi-Fi and the emerging WiMAX solutions. Further, the overviews of wireless technologies as used in different business environments are given. The article is categorized into two sections: logistics, and retail and distribution. Each section discusses an example of a wireless solution adopted in the Asia-Pacific region. This article offers a compilation of wireless solutions in the Asia-Pacific in order to map possible future scenarios on the use of wireless technologies in this region. The article serves as a foundation for further studies concerning the use of wireless technologies to improve quality of life.
BACKGROUND The wireless local area network (WLAN) based on the IEEE 802.11 family of standards has demonstrated great efficiency and received positive responses for delivering broadband services. On the other hand, IEEE 802.16 (WiMAX), is a new wireless standard for broadband wireless access (BWA). WiMAX,1 an acronym that stands for Worldwide Interoperability for Microwave Access, is a certification mark for products that pass conformity and interoperability tests for the IEEE 802.16 standards (Marks, 2006). It further extends
the performance of IEEE 802.11 (Wi-Fi2) in terms of capacity, coverage range, quality of service (QoS), and mobility (with 802.11e-2005). Technological revolution in communications is taking place in logistics. The industry is advancing significantly, adopting wireless technologies to secure its assets and improve services. The benefits accrued using wireless solutions in logistics include lowered insurance premiums, instant notification of security breach, flexible and secure handling of high-security cargo by authenticated personnel, data access through a wide range of mobile devices, and so forth. In addition, wireless solutions have the fastest returns on investment in the back office and supply chain functions of retail environments. Wireless applications in retail and distribution make workers more productive, streamline operations, help goods flow faster, and provide access to realtime data and inventory. As a result, productivity increases with the reduction in errors, which ultimately improves the customer’s experience. Businesses in other sectors have embraced the information revolution to reduce costs and improve productivity (Frist & Clinton, 2004). They use information technologies not as an end, but as a means to innovate and improve.
Problem Description: Logistics Industry Today transportation companies are experiencing unprecedented upheaval (Baracoda, 2005). Amid growing customer demands and soaring costs, the logistics industry struggles to develop successful business models that can drive profitable results and achieve customer loyalty. Managing logistics business in the Asia-Pacific is very challenging and highly complex due to multiple countries, currencies, languages, and customs; varying technologies and logistics infrastructure; and multi-modal transportation (The Logistics Institute-Asia Pacific, 2002). Other problems faced by the logistics industry include security breaches, theft, high insurance premiums, and inability to track goods delivery in real time.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Snapshot Assessment of Asia Pacific BWA Business Scenario
Retail and Distribution The common problems faced in retail and distribution include the time taken in filling out and sending order forms, as well as printing of customers’ orders and receipts, failure to provide one-to-one effective marketing due to the inability to access customer and product data at all times, the hassle to retrieve an up-to-date product catalog, and so forth. The increasing general enthusiasm on mobile technologies such as Bluetooth and radio frequency identification (RFID) has a positive effect on the acceptance of new mobile applications and services in retail and distribution (Ondrus & Pigneur, 2004). This would explain the reasons why wireless point of sale (POS) solutions are adopted by retailers. Wireless payments and ticketing are becoming a new trend for quickservice-oriented industries such as toll booths (e.g., Smart Tag in Malaysia).
VIEWS OF ENVIRONMENT AND PROCESSES: LOGISTICS INDUSTRY The cargo transportation industry is advancing significantly adopting wireless technologies to secure its assets and improve services (Nithyasree, 2005). Figure 1 shows the business environment of wireless applications used in the industry.
Uses of Wireless Solutions in Logistics Industry •
Asset/ Cargo Tracking System: A satellite-based vehicle tracking system using global positioning system
Figure 1. Business environment for the logistics industry
Driver
Fleet Manager
Truck
Warehouse
Internal External
Workmen
Carrier Airport Insurance Company
Factory Ship
Example of Cargo Tracking System When the workmen pack goods to be delivered onto a truck, the fleet manager checks his personal digital assistant (PDA) for a list of guards on duty. He can see on his PDA the whereabouts of the security guards, and he makes sure that there is no sign of intrusion. Elsewhere, at a seaport, another fleet manager checks his PDA for information on each container, including its physical location based on GPS, parameters such as temperature and humidity, and whether there is any sign of intrusion. The information gathered can be connected to centralized databases. A service-oriented infrastructure allows the staffs to instantly share information. At the same time, a customer checks the location of his goods using his laptop at a hotspot (Wi-Fi). He is pleased that the goods will arrive on time. Once the goods are safely delivered to the customer, the driver enters details into his PDA to notify the fleet managers instantly. An example of a cargo tracking system deployed in the Asia-Pacific region is Kwikfleet (http://www.kwikfleet.com/). Kwikfleet is a Malaysian company offering products and services either fully or jointly developed in Malaysia. With mobile data terminals (MDT), ruggedized portable computers, and wireless modems in their vehicles, fleet managers and their drivers in the field can take advantage of two-way computer-aided dispatching to stay connected while maintaining optimized scheduling and lowest time to destinations through advanced matching and dispatch algorithms. On-the-fly route planning technology will allow dynamic route planning algorithms to be run remotely on the MDT or locally on the intelligent vehicle location system server to serve portable data terminals (Kwikfleet, 2005). By using a geographic information system (GIS) and GPS, fleet managers are able to track a vehicle’s location, speed, route traveled, as well as fuel level and so forth.
Business Processes
Supplier
Train
•
(GPS) with satellite communications, geofencing3, and cellular communication technologies allows fleet managers to remotely monitor, track, and communicate with their drivers in real time (Nithyasree, 2005). Electronic Seals and RFID: RFID technology is effectively utilized in the shipping and railroad industries alike. Electronic tracking tags and seals attached to a rail or ship create a WLAN that automatically informs the driver or a central control station of a broken seal (Nithyasree, 2005). These tags can also send vital information about the shipments such as the current status, whether tampered prior to destination, and so forth.
Plane
The logistics industry-related business processes involved in the cargo tracking system example are: 907
S
Snapshot Assessment of Asia Pacific BWA Business Scenario
Figure 2. Business environment for retail and distribution
Sales Assistant Store Manager
Warehouse
Internal External
$
Cashier Bank Insurance Company
Customer
Supplier
• • • •
security monitoring, asset/goods management, logistics personnel communications, and route selection.
Retail and Distribution Retailers are using employee-activated wireless handheld devices to update inventory processes and increase accuracy and efficiency in all areas of the supply chain. In the process, they are taking significant steps toward reducing human error, returning salespeople to the business of helping customers, and avoiding problems like out-of-stocks and overstocking in order to gain competitive advantage (Schwartz, 2002). Figure 2 shows the business environment of wireless applications used in retail and distribution today.
Uses of Wireless Solutions in Retail and Distribution Industry •
•
908
Wireless POS: Eliminating cables not only improves user convenience, productivity, and safety, but also cuts down on extra expenses over time. As cables age, they have to be replaced at a rate of two cables per terminal per year. Wireless POS provides better customer service by accommodating fluctuations in customer volumes and providing timely service. This can be achieved using Wi-Fi for WLAN connections, and subsequently using WiMAX or Mobile-Fi (based on IEEE 802.20) for connections between branches at different locations (Wu & Yallapragada, 2006). Inventory Management and Replenishment: A supplier typically would deliver a certain quantity of
•
items, scratch out an invoice for the retail store manager to file away, and deliver a copy of that invoice back to the supplier’s own accounts department for processing. Retailers frequently dispute bills submitted for payment because of pricing discrepancies, or charged back for unauthorized deliveries. Payments were slow and often incomplete. As a result of this tedious process, retailers suffered from inaccurate inventories. At the same time, suppliers were troubled by lengthy check-in times and high administrative costs, and struggled with remittance. By accessing information from inventory management by means of wireless access technologies, a store manager will be able to retrieve information about goods availability anywhere in real time and plan ahead to schedule purchases by notifying the suppliers. Suppliers will be able to respond almost instantly whether they are able to fulfill the order at a specific time and arrange for delivery. Price Management: If the retailer has a concern with pricing and price markdowns, a price management application with real-time access (mobile wireless) to the in-store computer has proven to improve item price accuracy (Pillar, 2003).
Example of Wireless Solutions for Retailers At a corner of a supermarket, a customer scans a number of products at a kiosk while cruising through the store to check for prices and additional information. Elsewhere, a store manager accesses stock details from his portable device. The inventory management system prompts him to make a number of purchase orders to replenish a number of goods. He immediately sends purchase orders to suppliers. One of the suppliers responded that the goods could not be delivered on time. The store manager searches for other suppliers in order to stock up the goods. Customers are queuing up to make payment at the counters. One of the customers who is in a rush makes purchases online via his PDA and is happy that the goods will be delivered to his house within an hour. In the warehouse, a number of workmen place goods into a truck according to the list shown on their portable devices. An example of wireless solutions for retail and distribution industry deployed in the Asia-Pacific region is the SkyWire Wireless Automatic Data Capture Solution (http://www. skywire.com.au/). The solution aims at making stocktaking, price checking. and other retail applications more “hassle free.” The improved functionality offered by real-time data flow between shop floor and back office systems helped a duty-free retail chain in Australia to deliver even better customer services and enjoy improved operational efficiencies (SkyWire, 2005).
Snapshot Assessment of Asia Pacific BWA Business Scenario
Figure 3. Technical environment for wireless solutions in logistics industry
S
Fleet manager remotely monitors, tracks and communicates with drivers in real-time RFID technology for electronic tracking tags and seals attached to cargo allows workmen to manage inventory efficiently
Mobile data terminal, wireless modem and satellite-based vehicle tracking system using GPS
Internet Service Provider
Business Processes The retail and distribution business processes involved in the example of wireless solutions for retailers are: • • •
inventory management and replenishment, retailers and distributors negotiation, and price management.
FUTURE TRENDS: LOGISTICS INDUSTRY Figure 3 proposes the devices required to deploy wireless solutions in the logistics industry. In this business scenario, PDAs can be used within the coverage of WLAN. Switches/ routers are connected (wired) to a central database where fleet managers and workmen access timesheets, inventory, and so forth. A firewall is required to protect the system from intrusion to ensure that confidential data is not tampered with. Access to the Internet is likely to be restricted by security policy. Drivers will be accessing the warehouse database using mobile WiMAX (802.16e-2005) or Mobile-Fi (802.20) which supports mobility. RFID is an automatic identification method, relying on storing and remotely retrieving data using devices called RFID tags or transponders (Want, 2004). An RFID system may consist of several components: tags, tag readers, edge servers, middleware, and application software. The purpose of an RFID system is to enable data to be transmitted by a mobile device, called a tag, which is read by an RFID reader and processed according to the needs of a particular applica-
tion. The data transmitted by the tag may provide identification or location information, or specifications about the product tagged, such as price, color, date of purchase, and so on. GPS is a satellite navigation system used for determining one’s precise location and providing a highly accurate time reference almost anywhere on Earth or in Earth orbit (Beisner, Rudd, & Benner, 1996). The GPS system is divided into three segments: space, control, and user. The space segment comprises the GPS satellite constellation, whereas the control segment comprises ground stations around the world that are responsible for monitoring the flight paths of the GPS satellites, synchronizing the satellites’ onboard atomic clocks, and uploading data for transmission by the satellites. The user segment consists of GPS receivers.
Retail and Distribution Figure 4 depicts the devices required to deploy wireless solutions in retail and distribution. In this business scenario, PDAs and laptops can be used within the coverage area of WLAN. Switches/routers are connected (wired) to a central database where store managers and sales assistants access inventory. A firewall is required to protect the system from intrusion to ensure that confidential data is not tampered with. Access to the Internet is likely to be restricted by inhouse security policy. Customers may also scan barcodes found on items in the store to retrieve more information, for example, the number of items available, date of availability, item description, and so forth. Although wireless networking and computing is widely used in store operations, relatively few retailers have leveraged these investments with complementary wireless printing applications (Retail Biz, 2005). 909
Snapshot Assessment of Asia Pacific BWA Business Scenario
Figure 4. Technical environment for wireless solutions in retail and distribution
Retailers can take advantage of wireless printing. Various wireless printing applications can lower total in-store printing expenses, provide total cost of ownership benefits compared with traditional printers, improve labor efficiency, reduce store operating expenses, improve safety measures, and increase customer satisfaction (Retail Biz, 2005). Eliminating cables not only improves user convenience, productivity and safety but also cuts down on extra expenses significantly adding to the printer’s cost of ownership over time (Retail Biz, 2005).
CONCLUSION This article focuses on addressing one of the most widespread issues facing executives: aligning IT with business. The accomplishment of any major IT project is measured by the extent to which it is linked to business requirements, and demonstrably supports and enables the enterprise to reach its business goals (Wong, Tan, & Hiew, 2005a). This article has presented business scenarios in logistics as well as retail and distribution sectors. These analyses shape important techniques that may be exploited at various stages of defining enterprise architecture in order to derive characteristics of the architecture directly from high-level requirements of the business. In this study, this is achieved by examining business and technical environments, as well as related processes to enable successful deployment of wireless solutions in both of these industries. The technique has been used to help identify and understand business requirements, and hence to derive business requirements that the architecture development and ultimately the IT has to address. This 910
helps to encourage the uptake of wireless technologies in the Asia-Pacific region.
REFERENCES Baracoda. (2005). When it comes to transportation Baracoda really delivers. Retrieved September 26, 2005, from http:// www.baracoda.com/baracoda/solutions/p_1.html Beisner, H. M., Rudd, J. G., & Benner, R. H. (1996). Realtime APL prototype of a GPS system. ACM SIGAPL APL Quote Quad, 26(4), 31-39. Frist, B., & Clinton, H. (2004). How to heal health care. Washington Post, (August 25), A17. Kwikfleet. (2005). About Kwikfleet. Retrieved September 29, 2005, from http://www.kwikfleet.com/kwikfleet/index2. htm Marks, R.B. (2006, January 15). The IEEE 802.16 Working Group on Broadband Wireless Access Standards. Retrieved January 18, 2006, from http://grouper.ieee.org/ groups/802/16/ Nithyasree, M.G. (2005). Wireless—Paramount in cargo security. Retrieved September 27, 2005, from http://logistics. about.com/library/weekly/uc120602a.htm Ondrus, J., & Pigneur, Y. (2004). Coupling mobile payments and CRM in the retail industry. Proceedings of the IADIS International E-Commerce Conference, Lisbon, Portugal.
Snapshot Assessment of Asia Pacific BWA Business Scenario
Pillar, M. (2003). Where is wireless in retail? Integrated Solutions Magazine. Retrieved October 3, 2005, from http:// www.ismretail.com/articles/ Retail Biz. (2005). High wired. Retail Biz. Retrieved Febuary 28, 2005, from http://www.ismretail.com/articles/ Schwartz, K. (2002). Retail goes wireless. Retrieved September 29, 2002, from http://www.kioskbusiness.com/JanFeb02/articles/article1.html SkyWire. (2005). Wireless mobile solution makes for “hasslefree” operations at downtown duty free stores throughout Australia. Retrieved September 29, 2005, from http://www. skywire.com.au/show_this_item.php?pageId=123&secId=8 &parentId=8&division=retail The Logistics Institute-Asia Pacific. (2002). The Logistics Institute-Asia Pacific launches centre of competence in optimization. Retrieved September 26, 2002. from http:// www.tliap.nus.edu.sg/tliap/Media_Events/E08Feb2002/ E08Feb2002.aspx Want, R. (2004). RFID: A key to automating everything. Scientific American, 290(1), 46-55. Wong, C. C., Tan, C. M., & Hiew, P. L. (2005a, December). Business scenarios assessment in healthcare and education for 21st century networks in Asia Pacific. Proceedings of the 8th International Conference on Enformatika, System Sciences and Engineering, Krakow, Poland, 175-180. Wong, C. C., Tan, C. M., & Hiew, P. L. (2005b, December). Early assessment of WLAN/ BWA exploitation opportunities in Asia Pacific. Proceedings of the IADIS International Conference on E-Commerce, Porto, Portugal, 434-438. Wu, G., & Yallapragada, R. (2006). IEEE 802.20 Mobile Broadband Wireless Access (MBWA). Retrieved January 18, 2006, from http://grouper.ieee.org/groups/802/20/
KEY TERMS Global Information System (GIS): Enables one to envision geographic aspects of a body of data. Basically, it allows query of a database and receives results in the form of map. A GIS can have many uses, for example, weather forecasting, sales analysis, population forecasting, land use planning, and so forth. Global Positioning System (GPS): A “constellation” of 24 well-spaced satellites that orbit the Earth and make it possible for people with ground receivers to pinpoint their geographic location. The location accuracy is anywhere from 100 to 10 meters for most equipment. Accuracy can be pinpointed to within one meter with special military-approved equipment. GPS equipment is widely used in science and
has now become sufficiently low cost so that almost anyone can own a GPS receiver. Quality of Service (QoS): On the Internet and in other networks, QoS is the idea that transmission rates, error rates, and other characteristics can be measured, improved, and, to some extent, guaranteed in advance. QoS is of particular concern for the continuous transmission of high-bandwidth video and multimedia information. Transmitting this kind of content dependably is difficult in public networks using ordinary “best effort” protocols. Radio Frequency Identification (RFID): A technology that incorporates the use of electromagnetic or electrostatic coupling in the radio frequency portion of the electromagnetic spectrum to uniquely identify an object, animal, or person. RFID is coming into increasing use in industry as an alternative to the bar code. The advantage of RFID is that it does not require direct contact or line-of-sight scanning. Wireless Fidelity (Wi-Fi): A term for certain types of WLAN that use specifications in the 802.11 family. The term Wi-Fi was created by an organization called the Wi-Fi Alliance, which oversees tests that certify product interoperability. Wi-Fi has gained acceptance in many businesses, agencies, schools, and homes as an alternative to a wired LAN. Many airports, hotels, and fast-food facilities offer public access to Wi-Fi networks. These locations are known as hotspots. Wireless Point of Sale (POS): A component of wireless telemetry, or machine-to-machine correspondence. Specifically, wireless POS is the ability to make a purchase using a credit or debit card in businesses that would use a wireless POS terminal, such as taxicabs, limos, home repair services (e.g., carpet cleaners or refrigerator repair), restaurants, or mobile merchants. Worldwide Interoperability for Microwave Access (WiMAX): A wireless industry coalition whose members organized to advance IEEE 802.16 standards for BWA networks. WiMAX 802.16 technology is expected to enable multimedia applications with wireless connection and, with a range of up to 30 miles, enable networks to have a wireless last-mile solution.
ENDNOTES 1
2
3
For further reading, access the WiMAX Forum at http://www.wimaxforum.org/ For further reading, access the Wi-Fi Alliance at http://www.wi-fi.org/ Restrict the movement of a vehicle or other object to within a specified area. The location of the vehicle is monitored by telemetry, and an alarm is raised if it goes outside that area. 911
S
912 Category: Mobile Software Engineering
Software Platforms for Mobile Programming Khoo Wei Ju Malaysia University of Science and Technology, Malaysia K. Daniel Wong Malaysia University of Science and Technology, Malaysia
INTRODUCTION Java 2 Micro Edition (J2ME), .NET Compact Framework (.NET CF), and Active Server Pages .NET (ASP.NET) Mobile Controls are commonly used alternatives in mobile programming. They provide an environment for applications to run on mobile devices. However, they are different in many ways, such as supported mobile devices, architecture, and development. Hence, it is important for mobile application developers to understand the differences between them in order to choose the one that meets their requirement. Therefore, in this article we will discuss the general architecture of J2ME, .NET CF and ASP.NET Mobile Controls and compare the three alternatives.
BACKGROUND AND INTRODUCTION Since the mid-1990s, the growth of wireless communications has led to the mushrooming of mobile devices in the market. Initially, the mobile devices were mainly cell phones with limited programmability. However, many analysts and company executives were worried that mobile phone sales would eventually slow down, prompting research and development into software suitable for cell phones (Grice & Charny, 2001). Hence, now, there is a rise of programmable mobile devices. Furthermore, programmable mobile devices these days include not just cell phones but smartphones, PDAs, and pocket PCs. There are three well-known alternatives in mobile programming for general-purpose applications: J2ME, .NET CF, and ASP.NET Mobile Controls. J2ME is a version of Java that provides an application environment running on consumer devices and embedded devices. It targets machines with as little as 128KB of RAM (Tauber, 2001). J2ME consists of Java virtual machines (JVMs) and a set of standard Java application program interfaces (APIs) defined through the Java community process (JCP). J2ME can be used with different configurations and profiles, which provide specific information to a group of related devices. Configurations support the Java core APIs. Profiles are built on top of configurations to support devicespecific features like networking and user interfaces. The J2ME is available in two main configurations: connected
limited device configuration (CLDC) and connected device configuration (CDC). Figure 1 shows the hierarchical structure of J2ME. .NET CF is a lightweight version of Microsoft’s .NET framework. It provides an environment for executing client-side code and eXtensible Markup Language (XML) Web services to smart devices. It is compatible with C# and Visual Basic.NET (VB.NET), and it supports (.NET Compact Framework Team, 2005): • • •
Windows mobile (2000, 2002, 2003)-based pocket PC, Windows mobile-based smartphones, and embedded systems running Windows CE .NET 4.1 and later.
.NET CF consists of two main components: the development environment and the runtime environment. The development environment, known as smart device extensions (SDEs), is a Visual Studio .NET (VS.NET) 2003 project type that allows .NET CF applications to be developed rapidly by simply dragging appropriate controls into the application. The runtime environment is the common language runtime (CLR). The size of the CLR and relevant class libraries is smaller than 2MB, which is suitable for mobile devices. The architecture of .NET CF is shown in Figure 2. Active server pages (ASPs) is Microsoft’s server-side scripting technology. An active server page has an .asp extension, and it mixes HyperText Markup Language (HTML) and scripting code that can be written in VBScript or JavaScript. ASP is distributed with Microsoft’s Internet information services (IIS) Web server, so most hosts using IIS will also offer ASP for dynamic Web programming. ASP.NET is the version of ASP that works with Microsoft’s .NET Framework. ASP.NET Mobile Controls was previously known as Microsoft mobile Internet toolkit (MMIT). It was renamed as ASP.NET Mobile Controls to reinforce the concept that it is a collection of ASP.NET controls designed for mobile applications. It extends the ASP.NET server-side technology to allow developers to develop applications for a variety of mobile devices. Executing on the IIS Web server, ASP.NET Mobile Controls allows Web applications to be accessed by
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Software Platforms for Mobile Programming
Figure 1. Hierarchical structure of J2ME
S
Services on Server
Local Code Internet Connection
CDC Optional Package
CLDC Optional Package
Personal Profile Personal Basis Profile
Mobile Information Device Profile (MIDP)
Java2ME
Foundation Profile
CLDC
CDC
Java Virtual Machine (JVM)
Java Virtual Machine (JVM)
Device Operating System Platform Hardware
Figure 2. .NET compact framework Sevices on Server
Local Code Internet connection
NET Compact Framework Class Library Common Language Runtime (CLR) Native Application
NET CF
Application Domain Host
Device Operating System Platform Hardware
almost any Internet-enabled mobile device. During runtime, it will automatically detect the device running the application. The application is then transformed into a form suitable for that device. This frees the developer to concentrate on the application logic and leaves the user interface rendering to the runtime (Lee, 2002a). Furthermore, it allows developers to visually drag and drop controls on forms aimed at mobile
devices using VS.NET. The rest of the work, such as writing the proper markup language (e.g., Wireless Markup Language (WML), wireless application protocol (WAP)), is handled by the toolkit. The application development environment for ASP.NET Mobile Controls should be familiar to most ASP. NET programmers. Figure 3 shows the architecture of ASP. NET Mobile Controls. 913
Software Platforms for Mobile Programming
Figure 3. ASP.NET mobile controls Web Server Remote Web Pages
Internet connection
ASP.NET Mobile Controls
Mobile Web Brower Device Operating System
Platform
Hardware
FEATURE COMPARISON J2ME, .NET CF, and ASP.NET Mobile Controls cannot be easily compared feature-to-feature because the analysis must include non-technology aspects such as market acceptance, development and testing tools, reach, standardization, and platform coherence. Besides, the final releases of J2ME’s mobile information device profile (MIDP), personal basis profile, and personal profile are still in production (Sun, 2005). On the other hand, .NET CF is in the final stages of its beta tests. Nevertheless, a feature comparison, although limited, should still be useful.
Flexibility of Machine Control and Scope of Applications Virtual Machines, Pointers, Native Features In the .NET Compact Framework, the common language runtime (CLR) environment executes .NET’s Microsoft Intermediate Language code. The CLR also offers support services, such as code verification, memory, and code security. The managed code is always translated into native machine code rather than interpreted. CLR supports interfaces and pointers. As for security policy, .NET CF grants full trust to all code (Microsoft, 2005b). The standard frameworks cover only a limited set of commonly used mobile device features. Other features are accessible via native methods. Besides, it is believed that .NET CF has better support for native methods than J2ME because Microsoft controls both .NET CF and the Windows operating system (Yuan, 2002). With J2ME, Java source code is compiled into machineindependent byte code. The byte code is then interpreted 914
by the Java virtual machine (JVM) during runtime. J2ME employs different versions of the JVM based on the needs of a particular situation. The configuration specifications define the characteristics of the J2ME virtual machines. In most cases, features of the JVM are removed to accommodate the needs of a configuration. The CDC runs on a C-virtual machine (CVM) that is fully compliant with the Java virtual machine specification. The CDC profile accommodates devices with as little as 512kB of memory, although it is really designed for platforms with about 2 MB of available memory (White & Hemphill, 2002). Sun provides a reference implementation of the CLDC specification that is based on the KVM, a small footprint of JVM that satisfies the CLDC requirements. However, products need not be based on KVM—any virtual machine that has the features required by the specification and can work within the resource restrictions of the CLDC environment can be used (Topley, 2002). Although JVM supports interfaces, it does not support pointers because it can result in unsafe code. The Java native interface (JNI), which allows access to native methods, can be used but only by CDC. For CLDC, the native features must be built into the runtime.
Consumer Applications, Multimedia, Gaming .NET CF supports direct draw on canvas, double buffering, and device button remapping through its rich Windows Forms User Interface library. It also supports multimedia playback by using the native methods from Windows Media Player on Pocket PC (Yuan, 2002). In J2ME, the mobile information device profile (MIDP) 2.0 for CLDC includes animation and game controls in the javax.microedition.lcdui.game package. Multimedia play-
Software Platforms for Mobile Programming
back is supported via the Java media framework (JMF) on the CDC or the multimedia optional package for the CLDC. Many game developers prefer J2ME, because it is supported by a wider range of mobile platforms.
Development Support Programming Languages ASP.NET supports any language supported by the .NET Framework, including C, C++, C#, Visual Basic, and even Java. However, .NET CF currently supports only two major .NET languages: C# and VB.NET (Microsoft, 2005a). C# and VB.NET are standardized by EMCA and ISO/IEC. Hence, Microsoft has long been criticized for tightly controlling its technologies. However, the support of multiple standardized languages allow developers flexibility in programming in .NET CF. J2ME only supports Java. Anyone can propose a Java specification request (JSR) to the Java community process (JCP) for a new platform extension. Unlike with the tightly controlled development of .NET compact framework and ASP.NET, it may appear that under a more free process like the JCP, developers have to spend much time understanding the features to make use of all extensions in the language. However, J2ME APIs undergo rigorous standardization processes to ensure wide industry support and minimum learning for developers.
Platforms .NET CF supports high-end PDAs such as Windows pocket PCs, Windows smartphones, and embedded devices running
on the Windows CE .NET platform (Microsoft, 2005a). Windows devices consist of only a small part of today’s mobile device population. With J2ME, most of the cell phone devices (Motorola iDEN, Nokia Symbian OS, and Qualcomm Brew platforms) and low-end PDAs (Palm OS and Real-Time OS platforms) have built-in Java support because Java allows developers to be productive across many mobile platforms. Because it is server-side based, placing minimal requirements on the client, ASP.NET is supported by the widest range of mobile devices, including all devices that support .NET CF, all devices that support J2ME, and more. However, each of these devices may present the output of the ASP. NET controls differently due to their different limitations and capabilities.
Development Tools Regarding .NET CF and ASP.NET, Visual Studio .NET provides similar design interfaces for both mobile and nonmobile applications. It supports Web services integration and relational database access, and VS.NET is tightly integrated with Visio Enterprise Network Tools edition, which can generate C# or VB.NET code from UML (Unified Modeling Language) diagrams. Furthermore, VS.NET supports debugging on both emulators and real devices. However, VS.NET is not free. Sun’s J2ME Wireless ToolKit is a widely used MIDP development tool. Furthermore, command-line tools and vendor-specific toolkits are readily available. All major Java integrated development environments (IDEs) have J2ME modules or plug-ins. A big challenge for all J2ME IDEs is vendor software development kit (SDK) integration. Every
Figure 4. Devices supported by .NET CF, J2ME, and ASP.NET mobile controls
Devices with Web browser
.NET CF
Devices on Windows CE .NET and Windows mobile platform
Some PDA, Smartphone
Java-enabled devices
J2ME
ASP .NET Mobile Controls
915
S
Software Platforms for Mobile Programming
device vendor provides SDKs for their device emulators and proprietary J2ME extensions. The unified emulator interface (UEI) is designed to standardize the interfaces between IDEs and device SDKs. However, the UEI is available only through a Sun licensing program.
Miscellaneous Specification Process When a new technology emerges, Microsoft has the veto power to make decisions and make it available on .NET CF and/or ASP.NET. This saves time and effort. On the other hand, this also means that developers have no say on the specification process. The Java community process (JCP) decides the new J2ME standard APIs, whereas Sun has veto power on only Java language specifications. The JCP develops all current J2ME configurations, profiles, and optional packages, so the specification process is arguably very lengthy and inefficient. However, some developers like this because more people are allowed to contribute and decide.
Gateways There are technical difficulties in using .NET CF in mobile gateways because it was not designed to run lightweight application servers required in mobile gateways. Although Microsoft mobile information server (MIS) is a powerful gateway, messaging, and synchronization server, .NET CF lacks built-in APIs to interact with Microsoft MIS (Yuan, 2002). For J2ME, the primary mobile service gateway product is from IBM. The Oracle9i wireless application server and Oracle J2ME SDKs provide gateway integration points for mobile devices to many other Oracle or third-party application servers (Yuan, 2002).
Additional Comparisons Between .NET CF and ASP.NET Mobile Controls Server-Client Side .NET CF uses client-side technology. Code is executed on the mobile device using just-in-time (JIT) compilation and native execution. ASP.NET uses server-side technology. Code is executed on the server, producing markup-language-based output such as HTML to be interpreted by a Web browser.
Web Server For .NET CF, a Web server is not needed because code is executed directly on the device. For ASP.NET, a Web server (such as Microsoft IIS 5.0 or 6.0) that supports ASP.NET is required. The ASP.NET HTTP runtime is used to handle and process requests via a set of ASP.NET server controls.
Device Support Only devices that have .NET CF runtime can execute programs written on .NET CF. On the other hand, since ASP.NET is server-side based, less processing is required on the client side compared to .NET CF. However, each of these devices may present the output of the ASP.NET controls differently due to their different limitations and capabilities.
Connectivity For .NET CF, standalone applications can be created and installed on portable devices such as the pocket PC, pocket PC phone, and smartphone. By having the applications downloaded into such devices, the devices can either be connected or not connected. The XML or SQL Server 2000 Windows CE editions are used for local storage when
Table 1. Comparison summary between .NET CF, J2ME (CDC and CLDC) and ASP.NET Mobile Controls
916
J2ME Connected Limited Device Configuration
.NET Compact Framework
J2ME Connected Device Configuration
ASP.NET Mobile Controls
Virtual Machine
Common Language Runtime (CLR)
Java Virtual Machine (JVM)
Common Language Runtime (CLR)
Portable Code
Intermediate Language (IL)
Byte code
Intermediate Language (IL)
Just-In-Time (JIT) Compiling
Yes
Yes
Yes
Garbage Collection
Yes
Yes
Yes
Portability
No
Yes
Yes
Software Platforms for Mobile Programming
Table 1. continued
S
Cross-Language Integration
Yes
No
Yes
Standardized
EMCA, ISO/IEC
Yes
EMCA, ISO/IEC
Server-Client Side
Client side
Client side
Server side
Web Server
Not needed
Not needed
Needed
Device Support
Pocket PC, smartphone, Windows CE
General-purpose Java phone, smartphone, and PDA
Device independent
Connectivity
Standalone
Standalone
Connected
Market Focus
Enterprise
Enterprise
Consumer and enterprise
Consumer and enterprise
Language Support
VB.NET, C#
Java
Java
VB.NET, C#, C++, C, Java
Platforms
Pocket PC, Windows CE
Major mobile platforms except Palm OS
All mobile platforms
All mobile platforms
API Compatibility
Subset of .NET
Subset of J2SE plus standard optional packages
Partial compatibility with CDC with additional standard optional packages
Subset of .NET
Native APIs
Platform Invoke
JNI; device and OS specific
-
-
Coding and Development Tools
Smart Device Programming (SDP), Microsoft Visual Studio .NET
Command line, vendor SDKs, CodeWarrior, and WebSphere
Command line, vendor SDKs, all major Java IDEs
Microsoft Visual Studio .NET
Specification Process
Single company
Community
Community
Single company
Service Gateway
-
Run gateways as OSGi servlets; run gateway clients via vendor-specific SDKs
Run gateway clients via vendor-specific SDKs
-
Security Model
Simplified.NET model
Full Java security manager
Limited Java 2 model supplemented by OTA specification
Simplified.NET model
Client Installation
ActiveSync, Internet Explorer download
Sync, download
Formal OTA specification
Lifecycle Management
-
OSGi for gateway apps, J2EE Client Provisioning Specification for generic clients
Included in OTA spec, works with J2EE Client Provisioning Specification
-
User Interface
Rich subset of Windows Forms
Rich subset of AWT (Abstract Windowing Toolkit), vendorspecific UI libraries
PDA Profile subset of AWT, vendor-specific UI libraries
Rich subset of Windows Forms
Mobile Database
SQL Server CE, Sybase iAnywhere Solutions(coming soon)
IBM DB2 Everyplace, iAnywhere Solutions, PointBase, Oracle9i Lite
Vendor-Specific relational implementation over RMS, Oracle SODA
SQL Server CE
Database Synchronization
Vendor specific
Vendor specific
Vendor specific
Vendor specific
XML API
Built into ADO.NET and other standard APIs
Third-party tools
Third-party tools
Built into ADO.NET and other standard APIs
917
Software Platforms for Mobile Programming
Table 1. continued E-Mail and PIM (Personal Information Manager)
Platform InvokeOutlook APIs
PDA optional packages
PDA optional packages
-
Short Message Service (SMS)/Multimedia Messaging (MMS)
Platform InvokeSMS/MMS
Wireless Messaging API (WMA)/WMA 2.0
Wireless Messaging API (WMA)/WMA 2.0
Simple Mail Transport Protocol (SMTP) and third party
Instant Messenger
Platform InvokeMicrosoft Network (MSN) and other IM client APIs
Third-party APIs for most IM clients including Jabber and Jxta
Third-party APIs for most IM clients including Jabber and Jxta
-
Enterprise Messaging
Platform InvokeMicrosoft Message Queuing (MSMQ)
Proprietary JMS (Java Message Service) APIs
JMS via thirdparty toolkits (e.g., WebSphere MQ Everyplace, iBus Mobile)
-
Cryptography
Third-party APIs
JCE (Java Cryptography Extension) and thirdparty libraries
Third-party libraries
Third-party APIs
Multimedia
Platform InvokeWindows Media Player APIs
Subset of Java Media Framework (JMF)
Built into MIDP plus J2ME multimedia APIs
-
Game
Included Windows Forms UI
Direct draw on Canvas
GameCanvas support in MIDP
-
Location API
APIs provided by carriers
Location API
Location API
-
working off-line. With ASP.NET Mobile Controls, an HTTP connection is required to request an ASP.NET page that uses the Mobile Controls.
FUTURE TRENDS In recent years, a strange trend can be seen in the design of mobile devices; they are getting bigger and bigger. They are gradually taking on more of the features of regular computers. We say this is a strange trend because mobile devices were originally meant to be stripped down, barebones devices with only the most useful features for mobile usage. Nevertheless, this trend is getting encouraging responses from users because the mobile devices are able to hold all the files they need to carry around. Due to the encouraging response from the users and the advances in technology, it is predicted that the trend will continue. Besides, the sales of traditional PDAs have declined in the past few years (ETForecasts, 2003). This is because the PDA market is gradually being taken over by the smartphone, also known as the PDA phone. Users favor phones with computer features, such as storage capacity and clearer display (Kewney, 2005).
918
Currently, non-Microsoft operating systems like Symbian dominate the smartphone operating system market. As the sales and variety of smartphones are increasing worldwide, assuming Windows mobile platform keeps the same percentage of the market, the usage of Windows mobile platform will grow as well. Besides, given the past success of Microsoft to expand into new related markets, it is quite likely that they could increase their market share over the next few years, and there is much room for them to grow. Since Windows mobile platform will support only .NET CF (Java is not included), the growth of Windows mobile platform will directly lead to the increase in the use of .NET CF in mobile programming. Furthermore, the current versions of .NET CF grant full trust to all code. However, the upcoming versions of .NET CF will offer a subset of the policy-driven, evidence-based code access security of the full .NET framework (Microsoft, 2005b). This is a good feature from the security point of view; however, it may be less convenient for developers compared to the current version. Hopefully in the future, both Microsoft and Java applications can coexist in the same mobile device.
Software Platforms for Mobile Programming
CONCLUSION
REFERENCES
.NET CF and J2ME are both excellent platforms for developing smart clients for mobile commerce applications. J2ME has already gained a lot of industry support as the most favorable platform for developing mobile applications, and there are over 500,000 skilled Java developers around the world (Wishart, 2002). J2ME implements a modular design and is portable across a variety of devices. The platform provides balanced support for both enterprise and consumer applications. J2ME vendors offer excellent selections of mobile databases and gateway application server products. On the other hand, .NET has the advantage over J2ME where it provides a single development platform and common coding practices based around Visual Studio. There are more than 1.5 million skilled developers worldwide (Wishart, 2002), and Microsoft has the largest tools and third-party developers program. The .NET CF platform focuses on enterprise applications with rich user interface, database, and XML Web services support. Hence, .NET CF is suitable for cash-rich customers with controlled mobile environments. However, .NET CF runs only on Windows-powered high-end PDAs. As a young platform, it currently lacks support for gateway servers and choices for mobile databases. For the near future, the choice between .NET CF and J2ME is not so much a question of the desired platform features (both are excellent in this respect) as the targeted devices. In the short run, J2ME is supported by more devices than .NET CF. In the long run, most experts expect both platforms to coexist in all market sectors. Developers must choose the right tools and make them all work in heterogeneous environments. For example, J2ME clients would need to work with .NET backend servers and vice versa. So it would ultimately not come down to a choice between J2ME or .NET CF. Both .NET CF and J2ME have advantages over ASP. NET Mobile Controls in supporting code that will execute on the device, and that can run in disconnected, connected, or occasionally connected modes. Therefore, for most enterprise mobile solutions, .NET CF and J2ME are more appropriate. On the other hand, if browser-based applications are required, Microsoft ASP.NET Mobile Controls can be used to develop mobile Web applications that adapt their page rendering for a range of devices, such as micro-browsers on PDAs, smartphones, and WAP phones. ASP.NET Mobile Controls allows the developers to target the users they need to target, without worrying about the device they are using.
ETForecasts. (2003, June 16). Smartphones have started to impact PDA sales. Retrieved December 13, 2005, from http://www.etforecasts.com/pr/pr0603.htm
ACKNOWLEDGMENTS
Faridi, M. (2003). Beginning compact framework. Retrieved from http://www.ilmservice.com/twincitiesnet/presentations/BeginningCF.NET.ppt Grice, C., & Charny, B. (2001, February 2). Wireless jungle still waiting for its king. Retrieved from http://news.com. com/2100-1033-252009.html Jagers, B. (2003). Comparing file transfer and encryption performance of Java and .NET. Retrieved from http://www. lore.ua.ac.be/Publications/pdf/Jagers2004.pdf Kewney, G. (2005, February 8). Landscape phones mark the resurgence of the PDA smartphone. Retrieved December 13, 2005, from http://www.newswireless.net/index. cfm/article/1918 Lee, W. (2002a, December 2). Developing mobile applications using the Microsoft Mobile Internet Toolkit. Retrieved from http://www.devx.com/wireless/Article/10148 Lee, W. (2002b, November 18). Announcing .NET Framework 1.1. Retrieved from http://www.ondotnet.com/pub/a/ dotnet/2002/11/18/everett.html Leghari, N. (2003, December 17). Tools and platforms: Choices for a mobile application developer. Retrieved from http://weblogs.asp.net/nleghari/articles/mobiledeveloper. aspx Microsoft. (2005a). .NET Compact Framework. Retrieved from http://msdn2.microsoft.com/en-us/library/ f44bbwa1(en-us,vs.80).aspx Microsoft. (2005b). Security in the .NET Compact Framework. Retrieved from http://msdn2.microsoft.com/en-us/ library/13s3wxyw.aspx Milroy, S. (2003, March 6). .NET Compact Framework overview. Retrieved from http://www.windowsitpro.com/Articles/Index.cfm?ArticleID=38314&DisplayTab=Article NET Compact Framework Team. (2005, January 6). .NET Compact Framework FAQ. Retrieved December 6, 2005, from http://msdn.microsoft.com/smartclient/community/cffaq/default.aspx Sun. (2005). Java 2 Platform Micro Edition (J2ME). Retrieved December 8, 2005, from http://java.sun.com/j2me/ index.jsp
The assistance of Lisa Tang in reviewing, and commenting on, a draft of this article is gratefully acknowledged. 919
S
Software Platforms for Mobile Programming
Tauber, D. A. (2001, August 3). What’s J2ME? Retrieved from http://www.onjava.com/pub/a/onjava/2001/03/08/ J2ME.html
Cell Phone (Mobile Phone): Electronic telecommunications device that is able to move over a wide area, connected using wireless radio wave transmission technology.
Topley, K. (2002). J2ME in a nutshell. O’Reilly & Associates.
Personal Digital Assistant (PDA): Mobile device that serves as personal organizer. The basic features of a PDA include phone book, address book, task list, memo pad, clock, and calculator software.
White, J.P., & Hemphill, D.A. (2002). Java 2 Micro Edition. Manning Publications. Wishart, A. (2002, April 29). Mobile development environments: .NET Contra J2ME. Retrieved from http://www. datalogforeningen.dk/fa/fa-20020221.html Yuan, M.J. (2002, February 21). Let the mobile games begin, Part 1. A comparison of the philosophies, approaches, and features of J2ME and the upcoming .NET CF. Retrieved from http://www.javaworld.com/javaworld/jw-02-2003/jw0221-wireless.html Yuan, M.J. (2003, May 16). Let the mobile games begin, part 2 J2ME and .NET Compact Framework in action. Retrieved from http://www.javaworld.com/javaworld/jw-05-2003/jw0516-wireless.html
KEY TERMS Active Server Page (ASP): Microsoft’s server-side technology that allows scripting language for dynamically generated Web.
920
Pocket PC: Operating platform for handheld devices introduced by Microsoft, based on the Windows CE operating system. Smartphone: Mobile device that integrates the functionality of a mobile phone and PDA by adding telephone functions to a PDA or including the PDA capabilities on a mobile phone. Web Service: Software system designed to support interoperability among machines over a network using a standardized interface. Windows CE: A simplified version of the Windows operating system designed to run on handheld-size computers. Windows Mobile: Operating system that replaced the Windows CE operating system on mobile devices. It includes a suite of basic applications for mobile devices based on the Microsoft Win32 API.
Category: Wireless Networking 921
Standard-Based Wireless Mesh Networks Mugen Peng Beijing University of Posts & Telecommunications, China Yingjie Wang Beijing University of Posts & Telecommunications, China Wenbo Wang Beijing University of Posts & Telecommunications, China
INTRODUCTION As various wireless networks evolve into the next-generation fixed broadband wireless access (BWA) systems, the wireless mesh network (WMN), expected as a promising technology, is still being standardized in IEEE 802.16 and commercialized in the World interoperability for Microwave Access (WiMAX) forum at present. In fixed BWA systems, the objective of applying mesh-typed topology is to build self-organized networks in the places where wired infrastructure is not pre-existing or not worthy to be deployed. The term “mesh-typed” here can also be described as “relay-based” or “multi-hopping,” which means that the connection from a particular mesh subscriber station (mesh SS) to the mesh base station (mesh BS) is via one or more successive wireless links. Multi-hop wireless networking has traditionally led to significant research in the context of ad hoc or peer-to-peer networks. However, the fundamental goal of relaying augmented networks like WMNs is to provide wide-bandwidth coverage and high-data-rate throughput, while the defining goal of conventional ad hoc networks is to accomplish communications without any pre-existing infrastructure in a short time. The mesh concept applying in WMAN systems has the relay-based and multi-hopping features. Since communications could take place through relay nodes, link distance could become much shorter, frequency and spatial reuse could become much more efficient, and interference could also become much lower. Thus WMNs could provide non-line-of-sight (NLOS) connectivity with high-data-rate capacity to extend the coverage range of existing pointto-multipoint (PMP) wireless networks, such as cellular mobile networks. Figure 1 depicts a possible scenario where WMNs can be deployed to provide broadband access to the IPv6 backbone network. In WMNs, each mesh SS operates as not only a host but also a wireless router, which forwards transferring traffic within the network as well as traffic that goes out to other networks. The network is dynamically self-organizing and self-configuring, with both mesh BS and mesh SS
automatically establishing and maintaining routes among themselves. All the nodes can use the distributed scheduling to ensure collision-free transmissions within their two-hop neighborhood, or use the centralized scheduling to complete functions in a more centralized manner through conveying much of the control work to the mesh BS; the combination of these two control mechanisms is termed as hybrid-controlling. Mesh BS is connected with WiMAX PMP BS through first tire wireless backhaul, and then WiMAX PMP BS is connected with the IPv6 backbone network through second tire wireless or wired backhaul. In the centralized scheduling mode of WMNs, all traffic is restricted to be either in the direction of the mesh BS or away from the mesh BS. However, in the distributed scheduling mode, the transmissions are communicated between arbitrary pairs of nodes. Hence, an ad hoc network, described in Figure 1, could be considered as a type of uncoordinated distributed WMN. Wireless sensor networks (WSNs) differ from the WMNs in that they contain hundreds or thousands of sensor nodes to allow for sensing over large geographical regions and these sensor nodes have much more limited computation capabilities, sensing capabilities, storage space, battery power, and transmission range. Even so, these sensors have the basic ability to communicate either among themselves or directly with an external BS, making WSN similar to both centralized and distributed WMN. This article introduces a functional architecture supporting the wireless mesh networks for the IEEE 802.16 standard. Three essential techniques—collision avoiding, packet scheduling, and wireless routing—are intensively presented. Based on the mesh extension of the IEEE 802.16 medium access control (MAC) layer protocol and the relaybased characteristic of WMNs, the algorithms concerning those three essential techniques are briefly reviewed. The suitable algorithms for collision avoiding and packet scheduling mechanisms are analyzed. Meanwhile, the wireless routing algorithm for the proposal architecture is discussed. The future research work is presented and the research problems are focused.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
S
Standard-Based Wireless Mesh Networks
Figure 1. WMN internetworking architecture Sensor Network (=Centralized and Distributed Mesh)
Ad Hoc Network (=Uncoordinated Distributed Mesh)
H
H
H
H
H
H
H
H
WiMAX PMP BS Wireless Mesh Network IPv6 Network
Mesh BS
SS Wireless Connection for Centralized Scheduling Wireless Connection for Distributed Scheduling
Figure 2. Frame structure for IEEE 802.16 mesh mode Super frame #n-1
#n
#n+1
Control subframe
Data subframe
Network control subframe MSH-NENT Opportunity
MSH-NCFG opportunity
Schedule control subframe MSH-CSCH&MSH-CSCF MSH-DSCH coordinated distributed Central schedule control minilslots schedule opporturnity
SYSTEM BLOCK IEEE 802.16 mesh mode is an optional extension of the IEEE 802.16 MAC layer as an alternative or a complement to the conventional PMP architecture in the fixed BWA systems. The physical layer of it supports OFDM modulation, particularly as presented in the IEEE 802.16 standard with both the licensed frequencies and the license-exempt frequencies operating below 11G Hz. In this physical environment with long wavelength, requirements of line-of-sight (LOS) are not necessary and impacts of multi-path may be significant. IEEE 802.16 mesh mode provides a novel method for new nodes to enter the network with the help of a full functionality member (sponsor node) in the network and has three kinds of scheduling modes including centralized scheduling, coordinated distributed scheduling, and unco922
ordinated distributed scheduling for efficient transmission of the data packets as well as control messages. The mesh frame structure defined in IEEE 802.16 mesh mode, which has no difference between uplink and downlink, is demonstrated in Figure 2. We refer to the frame containing the network control subframe as the “network frame” and similarly term the frame including the schedule control subframe as the “schedule frame” for short. One network frame and several schedule frames constitute a super frame. In the network control subframe, the first opportunity with seven OFDM symbols is for NENT (network entry) message transmission in which a new node sends an entry request or entry acknowledgement. The symbols remaining are for NCFG (network configuration) message transmission in which the network configurations are advertised. In the schedule subframe every seven symbols
Standard-Based Wireless Mesh Networks
Figure 3. Functional architecture for IEEE 802.16 standard-based WMN Wireless Mesh Routing
S
Mesh Link Link Establishment
Quality Scheduling Tree Updates
Mesh Packet Scheduling Centralized Scheduling
Forwarding Tree Updates
Coordinated Distributed Scheduling
Wireless Channel Conditions
Mesh Data Dropping, Queuing,...
Uncoordinated Distributed Scheduling Mesh Collision Avoidance Mesh Network Entry
are grouped as a transmission opportunity. The first several transmission opportunities are utilized for central schedule messages, including CSCH (central scheduling) messages and CSCF (central scheduling configuration) messages, while the remains are for coordinated distributed schedule messages: DSCH (distributed scheduling) messages. DSCH messages may also appear in the data subframes during uncoordinated distributed scheduling. Data transmissions scheduled uncoordinatedly should submit to those scheduled coordinately.
FUNCTIONAL ARCHITECTURE AND RESEARCH AREAS In order to focus on researching the key techniques of WMN, the suitable functional architecture, shown in Figure 3, is presented which is based on the cross-layer design and is to optimize the system performance. The proposed architecture consists of three main models: wireless mesh routing, mesh packet scheduling, and mesh network entry. There are three sub-parts in the mesh packet scheduling model: centralized scheduling, coordinated distributed scheduling, and uncoordinated distributed scheduling, which are defined to represent these three scheduling mechanisms respectively. Meanwhile, in order to support the mesh routing procedure, the wireless mesh routing model is involved. Since both scheduling and routing algorithms require the wireless channel conditions to improve the throughput and meet the quality of service (QoS) requirement to support the various applications, there is a cross-layer design between wireless mesh routing and mesh packet scheduling models. The wireless mesh routing model will forward the routing message to the mesh packet scheduling and make the scheduling model update its scheduling tree periodically. When
Mesh Network Synchronization
the radio channel condition is bad or there is not enough radio resource, the scheduling model can send a request message to the wireless mesh routing model to change the mesh routing. In order to guarantee the real-time requirement and minimize the delay, the directional and periodical signal exchange are necessary. Furthermore, the link measurement and establishment must be completed in the lower MAC layer, and the model mesh link establishment is added to assist the wireless mesh routing model. What is more, the mesh collision avoidance model is proposed to avoid collisions that may occur during the mesh network entry process and the uncoordinated distributed scheduling process. The mesh data dropping, queuing, etc., model is corresponding with the mesh scheduling mechanisms, while the mesh network synchronization model works together with the network entry procedure. Among functional parts described in Figure 3, collision avoidance mechanisms, scheduling techniques, and routing algorithms are three challenging research areas. Most existing MAC protocols based on conventional collision avoidance mechanisms solve partial collision problems, but raise others. Considering particular features of wireless multi-hop mesh networks, how to improve existing mechanisms to avoid collision that would happen both in network entry procedure and in uncoordinated distributed scheduling is a crucial issue according to the scalability of the entire network. Since resources needed to be scheduled are dependent on transmission techniques applied by PHY layer, scheduling techniques must make adaptations correspondingly, especially when advanced techniques such as MIMO and cognitive radios are introduced. Routing in WMNs is also a tough task due to the delay-sensitive applications, as well as higher throughput and bandwidth requirements which distinguish these networks from other wireless multi-hop 923
Standard-Based Wireless Mesh Networks
Figure 4. Collisions in IEEE 802.16 mesh mode
Collision during data transmission under non coordinated distributed scheduling Collision during the node ’ s entry
Mesh BS Nodes in the network
Message sending Coverage area of the node
Nodes wishing to enter the network Collision occur and data lose
networks like ad hoc and sensor networks. Therefore, novel routing algorithms need to be proposed to utilize the potential advantages of the wireless medium.
KEY TECHNIQUES FOR PROTOCOL IMPLEMENTATION Collision avoiding, scheduling, and routing can play a significant role in the implementation of IEEE 802.16 standard protocol. In this section we briefly review algorithms concerning those three key techniques and propose our new collision avoidance schemes, scheduling phases’ division, and routing metrics exploration respectively.
Collision Avoidance Schemes Although collision avoidance schemes have been presented in IEEE 802.16 mesh mode, collisions may still occur as demonstrated in Figure 4 during a new node entry process as well as during data transmissions controlled by uncoordinated distributed scheduling. Solutions for alleviating collisions in the above situations could increase the network throughput and reduce the average delay to some extent. Since both the start time and the duration of a transmission are unpredictable, taking into account the low-cost-requirement of normal mesh devices, we focus our attention on random access schemes for WMNs rather than fixed resource allocation schemes or dynamic resource allocations-on-demand schemes.
Wireless Scheduling Mechanisms IEEE 802.16 mesh mode MAC supports three scheduling mechanisms: coordinated centralized scheduling, coordinated distributed scheduling, and uncoordinated distributed 924
scheduling. Uncoordinated means performing in a partially, contention-based manner, while coordinated means using scheduling packets transmitted in a collision-free way within scheduling control subframes. Distributed means opportunity-based scheduling between two nodes, while centralized means mesh BS coordinates the radio resource allocation within the mesh network. In the centralized mechanism, every mesh SS sends its resource request to the mesh BS, and the mesh BS determines the resource allocation for each link and broadcasts the grants to SSs. A centralized mechanism is best for scheduling over links supporting persistent traffic streams, while distributed scheme is best for scheduling over links with occasional or brief traffic needs. Figure 5 depicts the scheduling tree for centralized scheduling and other possible connections for distributed scheduling. As shown in Figure 5, the centralized scheduled traffic only occurs on scheduling tree, in which different colors represent different hops from the mesh BS, while the distributed scheduled traffic occurs on both the scheduling tree and other links which are not scheduling tree branches. As most of the packet scheduling algorithms proposed for multi-hop wireless networks are TDMA based and could be further ameliorated for OFDMbased networks, our discussion will focus on the scheduling in a TDMA network. The scheduling mechanisms discussed as follows are based on the following assumptions: nodes are assumed to use omni-directional antennas and operate in half-duplex mode, which means a node cannot transmit and receive in the same moment. The wireless channel is free of non-collision-related errors.
Wireless Routing Algorithms In the case of IEEE 802.16 mesh mode, we consider routing in a relay-based network in which communication relations are limited to a few hops only. The choice of routing
Standard-Based Wireless Mesh Networks
Figure 5. Multiple scheduling mechanisms in IEEE 802.16 mesh mode
•
Mesh BS SS Scheduling Tree Branch Link which is not on Scheduling Tree
algorithm becomes challenging when considering multiple possible relays. When an intermediate node in the proposed WMN architecture receives traffic, its routing function decides the next hop on the path to the destination, and this node forwards traffic along to the next hop node. In this type of relay-based network, traffic at a node waiting for routing decision may be on behalf of other nodes that are not within direct wireless transmission range of a mesh BS. This is the mechanism that is generally used for routing algorithm designing in ad hoc and sensor networks. Compared to ad hoc networks, WMNs applying relaying via fixed nodes do not need complicated distributed routing algorithms, while retaining the flexibility of being able to quickly select another route as a link between relays breaks. Based on the performance of the existing routing algorithms for wireless multi-hop networks and the specific requirements of WMNs, we examine the following issues to design an optimal routing algorithm for WMNs: •
Routing Overhead: The routing overhead in wireless mobile multi-hop networks may become severe because of substantial memory requirements and wide bandwidth spending for route establishing and maintaining, thus it results in low efficiency in network throughput. Studies of routing algorithms for mobile ad hoc networks (MANETs) by the MANET subgroup of the Internet Engineering Task Force (IETF) have shown high routing overhead and reduced potential efficiency of multi-hop techniques. However, for WMNs, limiting the number of hops in a range (HRthreshold, which is a configuration value that need only be known to the mesh BS, as it can be derived by the other nodes from the MSH-CSCF message) would greatly simplify routing complexity. The other factor that would also simplify routing complexity is nodes’ minimal mo-
•
•
•
bility. Taking into account these two factors, routing overhead would be reduced greatly. Based on these simplicity results from inherent characters of WMNs, the routing algorithm must minimize routing overhead as much as possible. Fault Tolerance: Some links between neighboring nodes may fail or be blocked due to physical damage or environmental interference. The failure of links should not affect the overall task of the WMNs to ensure robustness as one of the objectives to deploy WMNs. If a link breaks, the routing algorithm should be able to quickly select another route to avoid service disruption. Scalability: Since the number of user nodes in WMNs may be much larger than that in ad hoc networks, and routing in a relatively large wireless network must take convergence time and end-to-end delay into account, scalability is also an important routing design parameter in WMNs. Load Balancing: Relay-based is the most distinguishing characteristic of the WMNs. Thus the potential substantial network resources could be shared among many users through relay-based communications. In this cooperative way, load-sensitivity is an important factor to measure the relative performance of different routing metrics. For example, when a part of a WMN experiences congestion, new traffic flows on behalf of other users should be routed through other routes avoiding this part. Routing Metrics: Many existing routing algorithms for ad hoc networks have traditionally attempted to find routes by using shortest path as a routing metric. It is known that the defining goal of conventional ad hoc networks is to function without any pre-existing infrastructure, while the fundamental goal of meshtyped multi-hop augmented networks is to provide wide-bandwidth coverage and high-data-rate throughput. Therefore, using shortest path metric for routing in multi-hop wireless networks is not sufficient to construct routes that are able to effectively transport traffic with reasonable delay, reliability, and throughput. In order to satisfy the original goal of WMNs, a routing algorithm must select better routes between given neighboring node pairs by explicitly using link quality measurements to explain the differences in quality of the paths. One obstacle to taking link quality into account is combining route metrics to form a path metric. This problem is not straightforward inherently, since several parameters might be used to give enough indication about the quality and behavior of a wireless link.
In order to exploit the potential advantages that the wireless medium defined by IEEE 802.16 offers, the approach 925
S
Standard-Based Wireless Mesh Networks
employing cross-layer design is a good way to explore new routing metrics. The objective is to turn lower layer drawbacks into advantages through use of MAC and PHY information to help routing selection and making the routing layer control some lower layer settings.
SUMMARY AND FUTURE WEOK The emergence of new multimedia and Internet applications for the wireless domain has spurred the study of WMNs for providing larger capacity and wider coverage. In this article, a novel functional architecture of WMNs based on IEEE 802.16 standard is proposed, and three key techniques demonstrated in this architecture, including collision avoiding, scheduling, and routing, are extensively studied. While discussing the desirable features and suitable techniques, in addition to the basic properties of wireless relay-based networks, the special characteristics of IEEE 802.16 mesh mode, such as fixed and hybrid-controlling, are also taken into account. Since the new node entry process in IEEE 802.16 mesh mode involves no resource allocation or link establishment and there would be rare new nodes required to enter the network simultaneously. Considering the multi-hop topology and multi-service applications in this kind of network, the three-phase scheduling mechanism is presented and analyzed. Finally, we discuss the approach of cross-layer design for integrating the link quality into the routing metrics. While we are designing the routing and scheduling algorithms, we should consider the physical limitations of each node and the interference between the transmissions of nodes in their neighborhoods. How to reflect the link condition from the physical layer to the network layer in time is a problem related to cross-layer design and needs more research work in the future. Besides that the path of a flow consists of multiple links, how to evaluate the overall performance of the route reasonably to choose the best routes is also an open issue. In addition, we need to consider more about the effective combination of scheduling and routing algorithms. Furthermore, there exist multi-services such as voice, minimum bandwidth required traffic flow, and best effort service. Therefore, QoS differentiation and guarantees must be supported. Wireless mesh will transmit data packets based on all-IP protocol, and with the emergence of new services, the user of WMAN based on IPv6 will wish not only to access the Internet through wireless links but also to remain online even while they are moving. The WMAN technology based on IPv6 is therefore widely envisioned to have tremendous market potential. With this background, it would be meaningful to investigate the network protocol of IEEE 802.16 mesh mode and its integration with IPv6.
926
REFERENCES Agis, E., Mitchel, H. et al. (2004). Global interoperable broadband wireless network: Extending WiMAX technology to mobility. Intel Technology Journal, 8(3), 173-187. Beyer, D., Waes, N. V., & Eklund, K. (2002, February). Tutorial: 802.16 MAC Layer Mesh Extensions. Proceedings of IEEE 802.16 Standard Group Discussions. Chen, J., Chi, C., & Guo, Q. (2005, October 3-5). A bandwidth allocation model with high concurrence rate in IEEE802.16 mesh mode. Proceedings of the 2005 Asia-Pacific Conference on Communications (pp. 750-754). Erceg, V. et al. (1999). An empirically based path loss model for wireless channels in suburban environments. IEEE/ACM Journal on Selected Areas in Communications, 17(7), 1205-1211. Erceg, V. et al. (2001). Channel models for fixed wireless applications. IEEE 802.16 Broadband Wireless Access Working Group, IEEE 802.16.3c-01/29r4. Retrieved from http://ieee802.org/16 Gupta, P., & Kumar, P. R. (2000). The capacity of wireless networks. IEEE Transactions of Inf. Theory, 46(2), 388-404. Iannone, L., Khalili, R., Salamatian, K., & Fdida, S. (2004, September 20-22). Cross-layer routing in wireless mesh networks. Proceedings of the 1st International Symposium on Wireless Communication Systems (pp. 319-323). IEEE Standard 802.16-2004. (2001). Revision of IEEE standard 802.16-2001: IEEE standard for local and metropolitan area networks, part 16: air interface for fixed broadband wireless access systems. Ko, Y.-B., Shankarkumar, V., & Vaidya, N.H. (2000). Medium access control protocols using directional antennas in ad hoc networks. Proceedings of IEEE INFOCOM 2000 (Vol. 1, pp. 13-21). Li, J., Blake, C., De Couto, D., Lee, H. I., & Morris, R. (2001). Capacity of wireless ad hoc networks. Proceedings of ACM SIGMOBILE. Nair, G., & Chou, J. (2004). IEEE 802.16 medium access control and service provisioning. Intel Technology Journal, 8(3), 213-228. Pabst, R. et al. (2004). Relay-based deployment concepts for wireless and mobile broadband radio. IEEE Communications Magazine, 42(9), 80-89. Sun, Z., Lu, Y., Zhou, Y., Peng, M., & Wang, W. (2005). A simulation model for the IEEE 802.16 broadband wireless access systems. OPNETWORK, (8).
Standard-Based Wireless Mesh Networks
Sun, Z., Lu, Y., Zhou, Y., Peng, M., & Wang, W. (2005). Research of uplink packet scheduling mechanisms based on GPSS in IEEE 802.16 systems. Proceedings of ICCI. Wu, Z., Liu, M., Wang, Y., Li, M., Peng, M., & Wang, W. (2005). Investigation of collision avoidance mechanisms in the IEEE 802.16 based wireless mesh networks. OPNETWORK, (8).
KEY TERMS CSCF: Central scheduling configuration message. CSCH: Central scheduling message.
HRthreshold: A configuration value that need only be known to the mesh BS, as it can be derived by the other nodes from the CSCF message. NCFG: Network configuration message. NENT: Network entry message. Network Frame: The frame containing the network control subframe to “network frame.” Schedule Frame: The frame containing the schedule control subframe to “schedule frame.” Super Frame: One network frame and several schedule frames constitute a super frame.
DSCH: Distributed scheduling message.
927
S
928 Category: M-Entertainment
Taxonomies, Applications, and Trends of Mobile Games Eui Jun Jeong Michigan State University, USA Dan J. Kim University of Houston Clear Lake, USA
INTRODUCTION Wireless communications and the distribution of cell phones have been rapidly extended with the expansion of mobile content services since the early 2000s. With such extension, mobile games have been viewed as a separate branch in game device platforms. While studies on mobile contents have increased for several years, research on mobile games is still in the early stages. Although mobile games have developed and expanded their ranges in game markets, there is little research on the classification and development trend of mobile games. Considering that game devices have been converged into ubiquitous communication/networking features and the range of games has been expanded from entertainment to education, health, and exercise, there is an urgent need to study mobile games’ taxonomy, application, and future trends. In this article, mobile games are classified by several criteria (i.e., contents, platforms, and multi-layer based). Examples of mobile games are summarized, along with taxonomies. Lastly, applications and the macro trend of mobile games will be presented. In addition, some insights in the design and development of mobile games will be discussed.
MOBILE GAME TAXONOMIES Games are different from other genres such as music, film, and literature in the participation of users. Interactivity and narratives are two important factors to categorize games. Aarseth, Smedtad, and Sunnana (2003) classified games with a number of basic dimensions such as space, perspective, time, and teleology. Klabbers (2003) suggested social systems such as actors, rules, and resources for the establishment of game taxonomy. Wolf (2005) considered some standards such as the games’ goals and objectives, and the nature of the games’ play-characters and control devices. The devices have been applicable to traditional games in PC or console games with wide monitors, gorgeous graphic environments, and broad structures of narratives. Nowadays, however, with the development of fusion games and the acceleration of
genre convergence, the clear division of game genres has been difficult because many cross-listed games emerged in two or more genres (Wolf, 2005). Considering such a trend and characteristics of mobile games, this article classifies mobile games into some basic genres. First, content-based taxonomy is conducted from the basic features of games such as the control range of gamers, the role of characters, and the degree of user participation. Second, platform-based is from mobile device platforms with which games are played. Third, multi-layer-based is from the capability of multi-player network and 3D graphic technology.
Content-Based Taxonomy The role of gamers is the essential element in game taxonomy. Gamers can take their own individual roles or become omnipresent beings in games. Role-playing games (RPGs) comprise the representative genre of individual role games; strategic simulation games are controlled by omnipresent gamers. These games can be divided by the environments of the role of gamers: gamers should be a shooting gunner in shooting games; gamers should be an adventurer in an unknown world in adventure games. In these games, users usually take on their own roles. Finally, such games could be classified by the degree of user participation: multi-player games are played by the collaboration of individual roles; team games are played by teams (or guilds) with enough members having individual roles. •
•
•
Role Playing Games: A gamer as a character takes on an individual role in accomplishing missions or quests. The user can upgrade his/her character and take items to complete missions effectively. Simulation Games: Users complete their missions in simulated environments by controlling resources such as objects, characters, and items with their own strategies. There are some types of simulation games such as construction, management, or war simulation games. Fighting Games: Gamers take a character and fight with skills of kicking or striking against other char-
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Taxonomies, Applications, and Trends of Mobile Games
•
• •
• • • •
acters to win the contests. There are some kinds of fighting games mixed with action marshals such as judo and boxing; they are sometimes referred to as action fighting games. Shooting Games: Users take on their missions as a shooter or artilleryman in a war or as an infiltrating agent in an operation. In these games, the perspective of a user is a very essential element in attracting the user’s involvement. Therefore, shooting games with the first-person perspective are usually referred to independently as first-person shooting games (FPSs). Adventure Games: Users take travels to unknown space or environment as travelers or warriors. Sports Games: Users take on a role or control teams in sports contests such as baseball, basketball, or football. Racing sports such as riding and car racing are usually called racing games. Board Games: Users compete with opponents in traditional board games such as chess, Tetris, puzzles, oriental chess, baduk, and so forth. Single-User Games: Only one user can participate with a role or mission. Team (or Guild) Games: Users should join a team with other users to complete missions or win a contest. Massively Multi-Player Online Games (MMOGs): A huge number of users can participate simultaneously with their roles or missions.
Platform-Based Taxonomy
terms of mobile devices, mobile games are classified into mobile phone, portable console, and PDA games. In terms of mobile platforms, mobile games are called Java games, Brew games, and WAP games. • •
•
Mobile Phone Games: Conducted in cell phones. Portable Console Games: Played in portable consoles; examples include PSP (Play Station Portable), NDS (Nintendo Dual Screen), and GBA (Game Boy Advance). PDA Games: Embedded or downloaded in PDAs.
Multi-Layer-Based Taxonomy Multi-layer-based games are classified with the adaptation of high technology including capability of multi-player network and 3D graphic technology. Mobile games have developed from text-based to 3D multi-user network games. Therefore, in terms of multi-layer-based taxonomy, mobile games can be classified into five types: text-based, 2D graphic, 2D network, 3D half network, and 3D multi-user network games. Table 1 summarizes the taxonomies of mobile games with genre examples by the criteria of division.
MOBILE GAME APPLICATIONS AND INDUSTRY Application Areas of Mobile Games
Mobile devices are also regarded as independent platforms. Each mobile device has its own features in containing mobile games. Thus, if producers want to transplant a game in a device into another one, they should restart the product processes from the beginning. For this reason, platformbased taxonomy is beneficially used in mobile games. In
The application areas of mobile games can be categorized into four areas: traditional game industry, mobile Internet applications, mobile advertisements, and new applicable areas. The most applicable area of mobile games concerns the traditional game industry. With the development of handheld
Table 1. Taxonomies of mobile games
Content-Based
Platform-Based
Multi-LayerBased
Criteria of Division
Genre Examples
Examples
Control range of gamers
RPGs Simulations
Doom RPG Real Estate Tycoon
Role of characters
Fighting (Action) Shooting, Sports Adventure, Board
Mortal Kombat Quake Mobile, FIFA06 Tomb Raider, Tetris
Degree of user participation
Single-user, Team, Multi-user
Deep Pocket Chess Samgukji, Undercover2
Device platforms
Cell phones, PDAs, Portable consoles, etc.
Game platforms
WAP, Java, Brew, etc.
Adaptation of 3D graphic and network technology
Text-based, 2D graphic, 2D network, 3D half network (3D games with two or several users) 3D multi-user network (3D MMOGs, etc.)
929
T
Taxonomies, Applications, and Trends of Mobile Games
mobile services, PC and console games have transplanted popular games into mobile devices: old big hit games have found their new hit area in game markets, so big game companies have expanded their business areas into mobile games. Owing to the advanced technology, some traditional game devices have included mobile capabilities. The boundary between mobile phone games and portable console devices has been revolutionarily converged, and mobile game cartridges or multimedia cards (MMCs) have been sold in game shops with game CDs and DVDs. Mobile games have provided various channels of game services. Internet portals and interactive TV service companies have provided mobile game services, and console game manufactures have opened mobile portal services for their customers of portable console devices with network capability. In particular, game producers developed new games such as LBS (location-based service) games using the features of mobile devices, and old mobile games have been upgraded to multi-user 3D games with the development of technology. Most of all, mobile games have been used by game companies to provide further services to users. Gamers can use their items and avatars saved in PC or console games by using mobile games, because PC and console network games can be used in mobile devices at any place, and they are linked to original devices without any loss of user information. Secondly, mobile games have great potential making mobile Internet applications flourish. Mobile game communities could expand mobile communities: the conversion of Internet games into mobile games could drive many Internet game communities to change their main space into mobile communities. Like the online shopping malls connected with Internet game portals, mobile game users could be excellent resources for shopping malls in mobile services; as with Internet chatting rooms in game portals, mobile chatting services could be developed with the connection of mobile games. With the division of mobile game users into heavy and light users, mobile portal sites would be differentiated: mobile game user types could be applied to analyze mobile user types for business, because game users could account for a high ratio of heavy mobile users. For the security of big hit mobile game information, mobile security technology and anti-duplication technology of mobile content sources would be applied into mobile games. Mobile game events such as mobile game exhibition, mobile game contests, and mobile game character shows could expand areas of mobile Internet applications. Thirdly, as mobile game users are not restricted in terms of age, gender, and social status, advertisements through mobile games could be as effective as public broadcasts. Companies can inform consumers of their logos or specific brands through mobile games. Recently, public relationships using the characters of popular games have been attracting the attention of companies, and information services about new products for target customers have been provided through 930
SMS (short messaging services), EMS (enhanced messaging services), or MMS (multimedia messaging services). Finally, mobile games could be applied to new areas such as education, exercise, and therapy in mobile services. Games could be a good way to enhance participation and involvement, while mobile devices could provide both easy accessibility and convenience. So, mobile services in the form of interactive games such as game-applied education, exercise programs with interactive games, and health programs with MMS could be developed.
Mobile Game Industry Mobile games comprise a rapidly developing industry in the game markets. In the market of mobile phone games, according to Datamonitor (2002) and In-stat/MDR (2004), the market size estimated around $500 million in 2002 is expected to grow over $5 billion in 2008. The size of the U.S. market was about $40 million in 2002, but it is expected to exceed $1 billion in 2008. Mobile game users are expected to reach over 70 million in 2008, which is about 10 times higher than in 2002. These figures exceed the growth ratio of other game markets such as PC-online and console games. As the economy of developing countries improves, and mobile devices continue to spread, mobile games are increasingly regarded to comprise a most fascinating market. The AsiaPacific market has captured over 50% of the total mobile phone game market, with Japan and Korea in the lead, but China, India, and South Asian countries have been expanding their market share at a speed parallel to the spread of mobile devices. In portable console games, according to NPD (2006), revenue in 2005 was $1.4 billion, notwithstanding the stagnation of console game markets in the U.S. Portable console markets have grown rapidly both in Europe and Asia with the development of network games. Mobile games have been developed with the spread of mobile devices, gradual upgrade of capabilities of mobile devices, development of graphic technology, introduction of the flat sum system, and the spread of mobile cultures. The mobile game market is no longer independent from other game markets, because most traditional games in PCs and consoles are translated into mobile devices, and games are converged without the division of game machines. The advent of game-specialized mobile phones and portable console game devices with network capability has accelerated the convergence of games. World game publishers have merged mobile game aggregators and extended the ratio of mobile game products. Leading game machine manufacturers (platform holders) such as MS, Sony, and Nintendo are competing for dominance of the future market, not only of game devices, but also of home entertainment devices. They are striving to take the world standard of new media with their game machines. Mobile game devices will be the ultimate destination for business triumph, because mobile
Taxonomies, Applications, and Trends of Mobile Games
devices are the most prevalent tools used by game users, and mobile games would be the most useful and profitable content in the entertainment market using state-of-the-art technology.
FUTURE TRENDS AND CONCLUSION Several trends in the mobile games industry are expected in the near future: the convergence of game devices, variation in game content, and prevalence of mobile game cultures. The first outstanding change will be the acceleration of device convergences into mobile game devices. As console game machines have been developed into portable devices, mobile phones will be enhanced with high-capability games, and new mobile game devices based on PC capabilities will be created. To conquer preoccupation of future mobile device markets, game publishers will focus on popular game brands and develop new technologies with high multimedia functions. Game producers will create less-expensive 3D network games and more various genre games. User interfaces in mobile devices will be enhanced, and devices will be focused on game functions that can massively run multi-user games. Long-lasting charge cells will be developed for mobile devices, and local area network (LAN) games with Wi-Fi and Bluetooth will be pervaded. High-definition screens will be able to run previous PC or console games with multimedia cards and security digital (SD) cards. The second trend is variation in game genres. Games will be differentiated for heavy and light users: for heavy users, hit MMORPG or strategic games such as World of Warcraft, Starcraft, and Lineage series will be created as mobile games. Mobile portals will be differentiated in terms of specified services, and their services will include generic mobile services such as shopping, chatting, and transactions. With the diffusion of flat sum services and the stabilization of the mobile network, 3D network games could be used in both mobile devices and PC or console devices simultaneously. The third trend is the expansion of the mobile game cultures. As game contests and exhibitions are popularized in game markets, mobile game events would gain more space among them. Mobile games will extend their ranges and lead mobile content services, and they will be regarded as one of the best ways to provide mobile interactive services such as education, medical consultation, and exercises. Therefore, new mobile games using such mobile cultures will be created beyond entertainment. In short, mobile games expand their areas not only with the convergence of game devices, but also with the expansion of their areas of application. With the development of state-of-the-art mobile devices and network technology, mobile games can include all the games played in other
device platforms such as PCs, consoles, and arcade game machines with the same environment. MMOGs in online computer games and high-definition 3D games in consoles can be adapted into mobile games. Furthermore, with the prevalence of mobile cultures, mobile games are thought of as a tool to communicate with others in various ways. As mobile cultures such as mobile communities expand, mobile games have expanded their ranges from entertainment to more applicable services such as education, exercise, and industry advertisements. Mobile games will expand their shares not only in game markets, but also in mobile-applied content markets.
REFERENCES Aarseth, E., Smedtad, S. M., & Sunnana, S. (2003). Multidimensional typology of games. Proceedings of the Level Up Conference (pp. 48-53). Utrecht: University of Utrecht. Brad, K., & Borland, J. (2003). The rise of computer game culture: Dungeons and dreamers from geek to chic. New York: McGraw-Hill. CESA. (2005). 2005 CESA game white paper. Tokyo: Computer Entertainment Suppliers’ Association. Datamonitor. (2002). Asia-Pacific mobile gaming: A study of best practice. Identifying success factors the Asia-Pacific markets. Datamonitor, (October). Gee, J. P. (2003). What video games have to teach us about learning and literacy. New York: Palgrave Macmillan. Hall, J. (2005). Future of games: Mobile gaming. In J. Raessens & J. Goldstein (Eds.), Handbook of computer game studies (pp. 47-55). Cambridge, MA: MIT Press. In-Stat/MDR. (2004). Mobile gaming services in the U.S., 2004-2009. In-Stat/MDR, (August). KGDI. (2005). 2005 game white paper. Seoul: Korea Game Development & Promotion Institute. Klabbers, H. G. (2003). The gaming landscape: Taxonomy for classifying games and simulations. Proceedings of the Level Up Conference (pp. 54-67). Utrecht: University of Utrecht. Newman, J. (2004). Videogames. London: Routledge. Nokia. (2003). Introduction to mobile game development. Retrieved from www.forum.nokia.com/html_reader/ main/1,,2768,00.html NPD. (2006). The NPD group reports annual 2005 U.S. video game industry retail sales. Retrieved January 17, 2006, from www.npd.com 931
T
Taxonomies, Applications, and Trends of Mobile Games
Ring, L. (2004). The mobile connection: The cell phone’s impact on society. San Francisco: Morgan Kaufmann.
Game Publisher: Game provider and copyright owner who connects game producers and content providers.
Schwabe, G., & Goth, C. (2005). Mobile learning with a mobile game: Design and motivational effects. Journal of Computer Assisted Learning, 21, 204-216.
Local Area Network (LAN) Game: A network game played with other users within a short distance using LAN equipment such as Wi-Fi and Bluetooth.
Wolf, M. (2005). Genre and the video game. In J. Raessens & J. Goldstein (Eds.), Handbook of computer game studies (pp. 193-204). Cambridge, MA: MIT Press.
Multi-Layer Game: A multi-user game where highdimensional graphic and network technologies are adapted, such as a 3D full-network game.
KEY TERMS
Platform Holder: A company that provides game platforms to developers for development of games. In console games, device manufacturers are the same as platform holders.
Game Genre: One of several classified game types in terms of storylines, role of players, or device platforms.
932
WAP Game: Game serviced by wireless application protocol (WAP).
Category: M-Business and M-Commerce 933
A Technology Intervention Perspective of Mobile Marketing Dennis Lee The University of Queensland, Australia and The Australasian CRC for Interaction Design, Australia Ralf Muhlberger The University of Queensland, Australia and The Australasian CRC for Interaction Design, Australia
INTRODUCTION In the last decade, the explosive growth and adoption of mobile phones has become commonplace in our everyday lives (Haghirian, Madlberger, & Tanuskova, 2005). In 1997, there were only 215 million people worldwide who used mobile phones as communication devices (Bauer, Barnes, Reichardt, & Neumann, 2005). Today, it is estimated that 2 billion people own a mobile phone worldwide and this number makes up a third of the entire human population (Wireless Intelligence, 2005). Mobile phones are no longer thought of as mere personal communication tools (Cheong & Park, 2005; Ito & Okabe, 2005). They have become a fashion symbol for teenagers and young adults (Katz & Sugiyama, 2005). Personalised ring tones, colours, display logos and accessories are individualised accordingly to suit individuals’ preferences (Bauer, Barnes, Reichardt, & Neumann, 2005). Furthermore, mobile phones are no longer just a platform for voice calls and sending and receiving text messages such as short messaging service (SMS). Photos, pictures and video clips can be attached as a multimedia message service (MMS) for communication purposes too (Okazaki, 2005a). With the recent introduction of 3G mobile technology, mobile phone users are able to perform more activities via their 3G enabled phone sets. They are able to browse the Internet fairly quickly, access online banking, play video games wirelessly, watch television programs, check for weather forecasts, allow instant messaging, and perform live video-conferencing (Okazaki, 2005b). The rapid growth of the mobile industry has created a foundation for mobile commerce (m-commerce). M-commerce facilitates electronic commerce via the use of mobile devices to communicate and conduct transactions through public and private networks (Balasubramanian, Peterson, & Jarvenpaa, 2002). The current emerging set of applications and services that m-commerce offers include mobile financial applications, mobile entertainment and services, product locating and shopping, wireless engineering, mobile auc-
tions, wireless data centres and mobile advertising (Malloy, Varshney, & Snow, 2002). Commercial research has indicated that consumers’ interest in m-commerce services and mobile payments have increased from 23% in 2001 to 39% in 2003 (Harris, Rettie, & Cheung, 2005). It is projected that by 2009 the global mobile commerce market will be worth at least US$40 billion (Juniper Research, 2004). Considering the projected worth of mobile commerce and the number of mobile subscribers, mobile marketing is increasingly attractive, as companies can now directly convey their marketing efforts to reach their consumers without time or location barriers (Barnes, 2002). The potential of using the mobile medium to market is now more attractive than before (Karjaluoto, 2005), as it can assist companies in building stronger relationships with consumers (Barwise & Strong, 2002), and can be used as a promotional channel to reach consumers directly (Barnes, 2002; Kavassalis, Spyropoulou, Drossos, Mitrokostas, Gikas, & Hatzistamatiou, 2003; Okazaki, 2004) anywhere and anytime. However, many aspects of mobile marketing are still in its infancy (Bauer, Barnes, Reichardt, & Neumann, 2005; Haghirian, Madlberger, & Tanuskova, 2005; Okazaki, 2004, 2005b; Tsang, Ho, & Liang, 2004). Research into mobile marketing is currently lacking, as this is a relatively new phenomenon. Very few studies have been conducted to demonstrate how the mobile phone channel can be successfully integrated into marketing activities of companies (Balasubramanian, Peterson, & Jarvenpaa, 2002; Haghirian, Madlberger, & Tanuskova 2005). Furthermore, no studies to date have compared the effectiveness of this mobile medium in delivering advertising and sales promotion with other more established media such as the print medium. The fundamental question that remains unresolved is, “What is the difference between mobile marketing and traditional marketing?” Will this new form of marketing be effective? How will consumers respond to this form of marketing? What will be the benefit to marketers when consumers receive this type of advertising? These are just some of the issues that marketers are concerned with in order to
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
T
A Technology Intervention Perspective of Mobile Marketing
evaluate the mobile channels for marketing purposes and are questions that are core to computer-supported collaborative work (CSCW) and technology intervention research.
MOBILE MARKETING Mobile marketing via SMS-based advertising and sales promotions is now being carried out by several multinational corporations (MNCs) in Europe and the United States of America. MNCs are very cautious in integrating such a new medium into their marketing mix (Mayor, 2005; Okazaki, 2005a). This is mainly because marketers are not fully convinced of the value of mobile channels as a marketing tool (Haghirian, Madlberger, & Tanuskova, 2005). Marketers are unsure whether their marketing efforts will cause positive or negative impacts on their consumers. Another issue is the difference in worldwide telecommunication networks and mobile handsets used in the last decade (Leppaniemi & Karjaluoto, 2005). The recent introduction of 3G mobile technology as a worldwide standard for telecommunication networks and mobile handsets has brought about a new level of investment safety for companies (Karjaluoto, 2005). Companies are beginning to test their marketing efforts via the mobile phone medium (Cheong & Park, 2005). This suggests the need for researchers to develop theories and models to inform how mobile marketing can work effectively in the mobile phone context (Karjaluoto, 2005). According to Tsang, Ho, and Liang (2004), mobile marketing can be classified as either permission-based, incentive-based or location-based. Permission-based marketing requires mobile users’ prior approval before specific marketing messages can be sent (Barwise & Strong, 2003). By getting the permission of the mobile users, the factor of irritation may be reduced when users read the advertisement. Incentive-based marketing provides specific rewards to individuals who agree to receive promotions (Tsang, Ho, & Liang, 2004). For instance, mobile phone users may get free connection time from their mobile service providers for retrieving and reading advertisements. Location-based marketing targets mobile users in a certain location. The advantage of location-based marketing is that advertisements are sent to those individuals who are present or near the location (Barnes, 2003). The incentive-based marketing approach is adopted because most consumers perceive the current mobile marketing as advertising, without making a distinction between sales promotions and advertising messages (Gogus, 2004). In other words, most consumers will generally term any marketing message received on their mobile phones as an advertisement, regardless of content (Gogus 2004). Moreover, the dominant form of mobile “advertising” appears to be in the form of promotion (Kavassalis, Spyropoulou, Drossos, 934
Mitrokostas, Gikas, & Hatzistamatiou, 2003; Haghirian, Madlberger, & Tanuskova, 2005; Mayor, 2005; Okazaki, 2004; Tsang, Ho, & Liang, 2004). In the marketing literature, a sales promotion can be defined as a more direct form of persuasion that may offer incentives to stimulate immediate purchase behaviour (Rossiter & Percy, 1998). Examples of sales promotional incentives include coupons, on-pack promotions, bonus packs, samples, premiums, and sweepstakes (Rossiter & Bellman, 2005; Shimp, 2003). Most of these promotional tools are based in print and termed as traditional promotional incentives (Belch & Belch, 2004). On the other hand, advertising can be defined as a relatively indirect form of persuasion that may cause a favourable mental impression and then create an inducement toward a purchase response (Rossiter & Percy, 1998). Advertising is considered as the placement of a message to either increase product awareness, promote sales of goods and services, or just disseminate information (Leppaniemi & Karjaluoto, 2005). Advertisements may also include the element of sales promotion, a common example of which is in the form of coupons. Coupons are considered to be some types of inducement that provide extra incentives to buy (Belch & Belch, 2004). Thus, in the context of mobile promotion, a mobile coupon is defined as an incentive that is paperless and electronic in nature (Wehmeyer & Müller-Lankenau, 2005). It is the fusion of the traditional print-based coupon with the mobile phone medium. A mobile coupon is delivered to a mobile phone handset as a message and is associated with mobile services and contents (Wehmeyer & Müller-Lankenau, 2005).
INTERACTION DESIGN AND THE LOCALES FRAMEWORK The Locales Framework is a comprehensive theoretical CSCW and interaction design framework in the field of information and computer science (Fitzpatrick, 2003). According to Fitzpatrick, Kaplan, & Mansfield (1998), this research framework is an approach that allows for the creation of shared abstractions among stakeholders (e.g., companies, individuals, consumers, marketers), and also to narrow the gap between social and computing concerns with a common language. Understanding the social phenomenon and designing a relevant application that can fit the social setting are the two important factors when applying the Locales Framework. It is the aim of Locales Framework analysis that more pragmatic design and systems applications are built to suit the social world (Fitzpatrick, 2003). The Locales Framework is based on five aspects, each of which are interdependent and overlapping, as they share various concerns with one another and are used to approach the domain to be studied from different perspectives—rather
A Technology Intervention Perspective of Mobile Marketing
than separating the domain into distinct subdomains to be studied independently. The locale foundation aspect portrays the social world and the locale it uses for its interaction (Fitzpatrick, 1998, p. 91). The social world can be characterised by a number of issues such as collective goal, memberships, duration, structure, culture and roles. A locale is the primary unit of analysis in the Locales Framework. A locale consists of the site and means that a social world uses in its pursuit of the shared purposes. According to Fitzpatrick (2003), a site is a place the social world uses and means are the objects within this place. The social world needs sites and means to facilitate their shared interactions. The civic structure aspect takes the locale of interest and considers its relationships and interactions with the wider community (Fitzpatrick, 1998, p. 92). In other words, it concerns the facilitation of interaction with the wider community within and beyond a person’s known social worlds and locales. The interaction with a wider community can possibly relate to an environment that is physical, spatial, geographical, organizational, informational, professional, legislative, and so on. The individual view aspect describes an individual’s single perspectives on one social world as well as on multiple social worlds (Fitzpatrick, 1998, p. 115). A single perspective is how an individual sees one social world, and is dependent on the level of engagement with the centre of that world, whereas multiple view sets incorporate the individual’s views of all the social worlds with which he or she is engaged. Individuals personalize their views to suit their tasks according to their current level of engagement. The interaction trajectory aspect identifies the dynamic, temporal aspects of the social world in action (Fitzpatrick, 1998, p. 122). This aspect identifies the actual interactions individuals have over time within the setting and with each other. Moreover, this aspect is not only concerned with the current action, but also with the past and projected futures. Awareness of past actions and outcomes, present situations, and visions for the future are important for creating plans and strategies. An important consideration to understand this aspect is to look at what perspective or point of view is applied to any particular domain. The mutuality aspect is a collaborative activity that draws specific attention to how the locale supports presence, and how awareness of that presence is supported for the achievement of shared activity. The mutuality aspect enables questions on who, what, when, where, why and how to be answered. When the Locales Framework is used, it involves a two-phase approach. This is iterative in order to better understand the nature of the given (Fitzpatrick, 1998). The first phase is to understand the current locales of interest from the view of the interaction needs. This could involve using qualitative data collected through an ethnographic study or a one-to-one interview. Generally the data collected could
then help to provide some relevant structure to designers when they engage in the design process of an application. It is argued that for designers who do not have any social science background, the Locales Framework could be applied as a sensitizing device to aid in formulating initial questions for the design process (Fitzpatrick, 1998). The second phase in applying this framework is to evolve new locales. The goal is to discover more possibilities for the existing locale of interest in order to better support the activities that take place there and to explore possible newer locales that can evolve as a result. This phase is to identify the advantages of any available medium, physical or computational, and the synergy among them, so that the needs of the social world are better meet. Specific questions that will help to drive this phase include: What interaction needs does the social world need that are lacking in this current locale? How can the existing locale be enhanced to support the aspects of the Locales Framework; namely, mutuality, individual views, civic structure and interaction trajectory? Can new technology be applied to the locale? Can new social worlds evolve if the resources are used in newer locales?
MOBILE MARKETING AS TECHNOLOGY INVENTION Prior research has identified the importance of coupons in affecting consumers’ cognitive, affective and conative behaviour during promotional campaigns (Raghubir, Inman, & Grande, 2004), but relatively little research has been conducted into the use of electronic coupons (Fortin, 2000; Suri & Swaminathan, 2004), particularly the form of mobile coupons (Okazaki, 2004). Most research in coupon studies is based mainly in the traditional medium of print (Coyle & Thorson, 2001; Liu & Shrum, 2002). Much of the current literature that has been mentioned is adapted purely from a marketing perspective. Since mobile marketing involves people, technology and applications, mobile marketing should also be investigated from a human-computer interaction (HCI) and CSCW perspective. This will perhaps provide a better understanding of how and what is best for mobile marketing. To better design mobile marketing strategies from a technological viewpoint, the use of the Locales Framework can be applied. The five inter-dependent characteristics of the Locales Framework guide study of the product or service to be marketed. An example may be a coffee shop:
Locale Foundation The social world will be portrayed by consumers trying to buy beverages or food at the cafes and the locale is the cafe. The means in this case will be the chairs, tables, coffee 935
T
A Technology Intervention Perspective of Mobile Marketing
machines, coffee counters, and the cashier’s machine found within this site. The new technologies, that is, mobile phones and systems to send mobile coupons, are also included. More broadly the cafe may be situated in a locale such as a shopping centre or University that has its own means.
Civic Structure The civic structure aspect considers issues such as the physical location of the café in its broader situation, store layout, and any competitors of café.
Individual View In the case of mobile marketing, the individual that comes into the café will be a consumer and thus his or her task may be to purchase a cup of beverage for enjoyment. When they leave the café, their perspective may change to follow the priority that they may have to engage in. Perhaps the perspective may change to acquire knowledge and thus attend lesson at a lecture theatre or maybe need to catch a ride home by become a passenger when boarding a public transport such as a bus.
Interaction Trajectory In this case study, the interaction trajectory aspect will determine the objective of the consumer coming to the café and how does the consumer interact with the surrounding environment. Despite receiving a discount coupon for cheaper beverages via the mobile phone, the consumer may come into the café with the purpose of meeting someone and not buy any beverages at all. To this consumer, the café has become a meeting venue and not a place for consumption. The café may become a place for taking a coffee break with fellow colleagues, and therefore the consumer may take advantage to purchase a cup of coffee at the special price for enjoyment.
Mutuality In this case scenario, the mutuality factor will look into how mobile marketing is supported and how applications can be created to support mobile marketing in the context of a café. Several advantages of the Locales Framework, as according to Fitzpatrick (1998, pp. 152,153), are listed as follow: •
936
It provides a common tool for understanding and designing of a social problem.
• • •
It has the potential to analyse issues from group to individual level, local setting to global context, and structure to process; It is independent of any one theoretical orientation when investigating into one phenomenon. It is a framework that is strong in identifying key elements of a collaborative environment but sufficiently generic, open and incomplete so as not to prescribe nor circumscribe all that is of interest.
Applying the Locales Framework approach to mobile marketing, we are able to build several “locales” (which can be defined as potential social world scenarios) in helping companies to consider before they actually implement their marketing plan. Moreover, companies that implement mobile marketing should consider the aspect of civic structure—the facilitation of interaction among various factors like physical, informative, geographical and technological parties. In the context of mobile marketing, companies should look at who their telecommunication service providers are, where their potential consumers are located, what applications should be used to generate response from consumers and what types of mobile phones should be able to receive mobile marketing. Companies need to understand that their potential consumers have many different perspectives and opinions, another aspect considered in the Locales Framework analysis. As mentioned earlier, the inducement of using a coupon to induce potential consumers to respond to mobile advertisements is a possible suggestion. Furthermore, companies need to understand that the locale does not stay static, as it is always changing and evolving. Therefore companies need to involve the interaction trajectory aspect that identifies actual interactions individuals have over time within a given context setting and with each other. The last aspect on mutuality involves companies to consider how best mobile marketing can be supported in a given location and how mobile marketing can create awareness for the companies. The Locales Framework does not attempt to account the findings of one particular phenomenon that is generalisable. It does, however, deliberately aim to characterise its findings that are open in many ways. In fact, one of its aims is to focus on providing an evolvable framework that can be made relevant for both understanding the social situation (in this case mobile marketing) and for improving better technology development to it. This approach, unlike many traditional marketing strategic research approaches, does not assume that the technology introduction has to accept static technologies, or unchanging user attitudes towards technology (rather than the product). A dynamic, multi-dimensional picture of clients and possible interactions allows more dynamic engagement models—supported by dynamic technology, not based on working around systems controlled by other developments.
A Technology Intervention Perspective of Mobile Marketing
FUTURE TRENDS Mobile marketing still lacks research. However, approaching this area from a technological perspective we can suggest several possible outcomes that a marketing and CSCW combined approach indicate: First, companies are increasingly able to understand potential consumers’ attitudes and behaviours in the context of mobile marketing from a more holistic perspective. In particular, the adoption of Locales Framework can provide insights to companies on how to further improve their design and concept for mobile marketing strategies in a particular situation. Second, there is a need to consider the aspect of interaction design in mobile marketing campaigns. Companies who intend to reach the target market effectively should consider factors like interactivity in their marketing materials, what types of technology (Wifi, Bluetooth, RFID, or global positioning system) can the companies adapt in mobile marketing and how best can the mobile medium fit in a given situation as well as to their potential consumers. The technology perspective of mobile marketing should be considered thoroughly. Third, situational factors like time and location are important issues that any given companies who decide to use the mobile channel for marketing need to consider. At present, there are no concrete solutions and applications for companies to fully adapt when they design their marketing materials. Fourth, marketing and technology are both at the forefronts of innovation and fashion. Technology introductionbased marketing methods may also drive technology R&D, when the integrated study approach suggests technological improvements. Lastly, companies may need to consider social factors like culture, values and norms prior to the launch of a mobile marketing campaign. A single set of mobile marketing materials cannot be carried out in different places, as the social factors are often different. Thus, companies operating across many countries may just need to create a general set of guidelines for mobile marketing with the ability to be tailored to specific context situations. The adoption of the Locales Framework is a suitable tool to be considered for such a multi-level guidelines and customisation approach.
CONCLUSION Using mobile phones as a medium for marketing is a new phenomenon. Companies need to understand the impact of this medium thoroughly before proceeding. Current research into mobile marketing begins with the traditional marketing perspectives and replaces existing media with mobile technol-
ogy. Such an approach is based on a historical perspective that doesn’t arise from the capabilities of human-computer interaction. The introduction of a theoretical-based research framework such as the Locales Framework is suitable in the investigation of this new medium for mobile marketing. A holistic perceptive of technology introduction in a business-to-client interaction to understand mobile marketing is described, with guidelines for supporting the development of mobile marketing strategies. Such a hybrid marketing/ interaction design approach generates new possibilities for both technology development and client engagement that either approach individually would not.
ACKNOWLEDGMENTS This work is supported by ACID (the Australasian CRC for Interaction Design) established and supported under the Cooperative Research Centres Programme through the Australian Government’s Department of Education, Science and Training.
REFERENCES Balasubramanian, S., Peterson, R. A., & Jarvenpaa, S. L. (2002). Exploring the implications of m-commerce for markets and marketing. Journal of Academy of Marketing Science, 30(4), 348-361. Barnes, S. J. (2002). Wireless digital advertising: Nature and implications. International Journal of Advertising, 21, 399-420. Barwise, P., & Strong, C. (2002). Permission-based mobile advertising. Journal of Interactive Marketing, 16(1), 1424. Bauer, H. H., Barnes, S. J., Reichardt, T., & Neumann, M. M. (2005). Driving consumer acceptance of mobile marketing: A theoretical framework and empirical study. Journal of Electronic Commerce Research, 6(3), 181-192. Belch, G. E., & Belch, M. A. (2004). Advertising and promotion: An integrated marketing communications perspective (6th ed.). Boston: McGraw-Hill/Irwin. Cheong, J. H., & Park, M-C. (2005). Mobile Internet acceptance in Korea. Internet Research, 15(2), 125-140. Coyle, J. R., & Thorson, E. (2001). The effects of progressive levels of interactivity and vividness in Web marketing sites. Journal of Advertising, 30(3), 65-77. Fortin, D. R. (2000, June). Clipping coupons in cyberspace: A proposed model of behavior for deal-prone consumers. Psychology & Marketing, 17, 515-534. 937
T
A Technology Intervention Perspective of Mobile Marketing
Fitzpatrick, G. (1998). The locales framework: Understanding and designing for cooperative work. PhD thesis. University of Queensland, Australia. Fitzpatrick, G. (2003). The locales framework: Understanding and designing for wicked problems. Kluwer Academic Publishers. Fitzpatrick, G., Kaplan, S. K., & Mansfield, T. (1998). Applying the locales framework to understanding and designing. Paper presented at OZCHI 1998, Australasian Computer Human Interaction Conference. Gogus, C. (2004) Understanding young adults’ participation in mobile sales promotions. Paper presented at the 13th EDAMBA Summer School. Soreze, France. Haghirian, P., Madlberger, M., & Tanuskova, A. (2005). Increasing advertising value of mobile marketing—An empirical study of antecedents. Paper presented at 38th Hawaii International Conference on System Sciences HICSS-38. Hawaii, USA. IEEE Computer Society Press. Harris, P., Rettie, R., & Cheung, E., (2005). Adoption and usage of m-commerce: A cross-cultural comparison of Hong Kong and the United Kingdom. Journal of Electronic Commerce Research, 6(3), 210-224. Ito, M., & Okabe, D. (2005). Intimate connections: Contextualizing Japanese youth and mobile messaging. In R. Harper, L. Palen, & A. Taylor (Eds.), The inside text: Social perspectives on SMS in the mobile age. London: Kluwer. Juniper Research. (2004). M-commerce market to grow to $40bn by 2009. Juniper Research. Retrieved November 20, 2006, from http://www.finextra.com/fullstory. asp?id=12605 Katz, J. E., & Sugiyama, S. (2005). Mobile phones as fashion statements: The co-creation of mobile communication's public meaning. In R. Ling & P. Pedersen (Eds.), Mobile communications: Re-negotiation of the social sphere (pp. 63-81). Surrey, UK: Springer. Karjaluoto, H. (2005). An investigation of third generation (3G) mobile technologies and services. Paper presented at BAI2005 International Conference on Business and Information. Hong Kong.
Liu, Y., & Shrum, L. J. (2002). What is interactivity and is it always such a good thing? Implications of definition, person, and situation for the influence of interactivity on advertising effectiveness. Journal of Advertising, 31(4), 53-64. Malloy, A. D., Varshney, U., & Snow, A. P. (2002). Supporting mobile commerce applications using dependable wireless networks. Mobile Networks and Applications, 7, 225-234. Mayor, T. (2005). The potential of mobile marketing is huge, but is there more to it than just fun and games? AlertAds. com. Retrieved November 20, 2006, from http://alertads. com/mobile-marketing-is-huge.html Okazaki, S. (2004). How do Japanese consumers perceive wireless ad? A multivariate analysis. International Journal of Advertising, 23, 429-454. Okazaki, S. (2005a). Mobile advertising adoption by multinationals - senior executives’ initial responses. Internet Research, 15(2), 160-180. Okazaki, S. (2005b). New perspectives on m-commerce research. Journal of Electronic Commerce Research, 6(3), 160-164. Raghubir, P. J., Inman, J., & Grande, H. (2004). The three faces of price promotions: Economic, informative and affective. California Management Review, 46(4), 1-19. Rossiter, J. R., & Bellman, S. (2005). Marketing communications: Theory and applications. Pearson: Prentice Hall. Rossiter, J. R., & Percy, L. (1998). Advertising communications and promotion management (2nd ed.). The McGraw-Hill Companies, Inc. Shimp, T. A. (2003). Advertising, promotion and supplemental aspect to integrated marketing communications (6th ed.). Thomson: South Western. Suri, R., Swaminathan, S., & Monroe, K. B. (2004). Price communications in online and print coupons: An empirical investigation. Journal of Interactive Marketing, 18(4). Tsang, M. M., Ho, S. H., & Liang, T. P. (2004). Consumer attitudes toward mobile advertising: An empirical study. International Journal of Electronic Commerce, 8(3), 65-79.
Kavassalis, P., Spyropoulou, N., Drossos, D., Mitrokostas, E., Gikas, G., & Hatzistamatiou, A. (2003). Mobile permission marketing: Framing the market inquiry. International Journal of Electronic Commerce, 8(1), 55-79.
Wireless Intelligence. (2005). Worldwide cellular connections exceeds 2 billion. GSM Association Press Release 2005. Retrieved November 20, 2006, from http://www.gsmworld. com/news/press_2005/press05_21.shtml
Leppaniemi, M., & Karjaluoto, H. (2005). Factors influencing consumers’ willingness to accept mobile advertising: A conceptual model. International Journal of Mobile Communications, 3(3), 197-213.
Wehmeyer, K., & Müller-Lankenau, C. (2005). Mobile couponing: Measuring consumers’ acceptance and preferences with a limit conjoint approach. Paper presented at the18th Bled eConference, eIntegration in Action. Slovenia.
938
A Technology Intervention Perspective of Mobile Marketing
KEY TERMS Computer-Supported Collaborative Work (CSCW): Combines the understanding of the way people work in groups with the enabling technologies of computer networking and associated hardware, software, services and techniques (Wilson, 1991). CSCW also addresses how collaborative activities and coordination can be supported by means of computer systems (Carstensen & Schmidt, 2002). Moreover, CSCW involves incommensurate perspectives as well as incongruent strategies and discordant motives (Schmidt & Bannon, 1992) to gain a better understanding of collaborative efforts within organizations so that this understanding can be used to effectively design collaborative technology that can be best deployed within the organizations. Human-Computer Interaction (HCI): The study of how people interact with computers and to what extent computers are or are not developed for successful interaction with human beings (ACM SIGCHI, 1996). In fact, HCI is a very broad discipline that encompasses different fields with different perspectives regarding computer development. For instance, HCI in psychology is concerned with the cognitive processes of humans and the behaviour of users, while HCI in computer science is concerned with the application design and engineering of the human interfaces. In sociology and anthropology, HCI is concerned with the interactions between technology, work and organization and the way that human systems and technical systems mutually adapt to each other. Interaction Design (ID): The study of designing interactive products to support people in their everyday and working lives (Sharp, Rogers, & Preece, 2002). One of the objectives in ID is to produce usable products that are easy to learn, effective to use and also provide an enjoyable experience. Generally users are engaged in the design process.
Locales Framework, The: The Locales Framework is a theoretical-based research framework for interaction design, with a key focus on CSCW. It approaches the study of a certain context or domain with the aim to discover findings that are open in many ways. The general set of guidelines derived from a context or domain is known as the five interdependent aspects. They are: locale foundations, civic structure, individual views, interaction trajectory, and mutuality. Mobile Advertising (M-Advertising): Advertising is a mass-mediated communication tool. Its aim is to communicate with the intended audience to buy into the desired message. In the context of mobile phones, mobile advertising aims to present the desired information to the consumers, hoping that consumers will react. Currently, most “advertising” contents found on mobile phones are considered as sales promotional materials, which aim to persuade consumers to buy the products. Mobile Marketing (M-Marketing): Marketing a company’s advertised and promotional materials via mobile phones through short message service (SMS) or multimedia messaging service (MMS) is known as mobile marketing. The mobile phone is to the adapted a marketing channel to reach consumers. Technology Intervention: Technology intervention is the intentional introduction of a technology, or method, into a context to alter that environment. Technology intervention may be targeted at improving information flow, communication, or other types of awareness. In mobile marketing, mobile phones can be seen as a technology intervention in the company-client relationship. Interaction design specifically focuses how to use technology intervention to improve interactions.
939
T
940 Category: 3G
3G Commercial Deployment Mugen Peng Beijing University of Posts & Telecommunications, China Shuping Chen Beijing University of Posts & Telecommunications, China Wenbo Wang Beijing University of Posts & Telecommunications, China
INTRODUCTION Currently, five terrestrial radio interfaces, which can be categorized as frequency division duplex (FDD) and time division duplex (TDD) modes, have been approved as the IMT-2000 radio interfaces. CDMA in TDD (TDD-CDMA) mode will be based on the harmonization between UTRA (UMTS terrestrial radio access) TDD and TD-SCDMA (time division-synchronous CDMA). Compared to UTRA TDD, which is 3.84Mcps in 5MHz bandwidth, TD-SCDMA, is also called low chip rate (LCR) TDD or Narrowband (NB) TDD for its 1.28Mcps in 1.6MHz bandwidth. TDD uses a combined time division and code division multiple access scheme. Hence the signals of different users are separated in both time and code domains (Chen, Fan, & Lu, 2002). Jointly developed by the China Academy of Telecommunications Technology (CATT) and Siemens, TD-SCDMA is one of the five IMT-2000 standards accepted by the ITU. The main benefits of TD-SCDMA are that it can be implemented less expensively than comparable 3G systems since it is much more spectrum efficient and is compatible with the current deployment of GSM network elements in China, allowing 3G asymmetric services without installation of completely new infrastructure. Compared with WCDMA and CDMA, TD-SCDMA (Peng et al., 2005) adopts TDD duplex mode, uses the same frequency band for the uplink and downlink, and makes full use of the asymmetrical frequency resource. Meanwhile, the TDD mode has the adjustable switch point between uplink and downlink, which can adapt to the asymmetrical service in uplink and downlink and makes full use of the spectrum. Furthermore, the symmetrical channel feature of TDD systems makes it very flexible and convenient for TDSCDMA to adopt the advanced technologies such as joint transmission, smart antenna, and so on, which can improve the system capacity and spectrum efficiency. Because the uplink and downlink of TD-SCDMA use the same carrier frequency, the channel propagation features and channel
impulse response in downlink and uplink have strong correlation when the interval between uplink reception and downlink transmit is less than the channel coherent time. So the channel information estimated in uplink can be directly used for downlink transmission, and this provides a good condition for implementation of smart antenna. Therefore, compared with the FDD system, channel estimation, power control, and smart antenna in the TD-SCDMA system become more simple and feasible. As the CDMA system is a self-interfering system, interference among users is the key factor that limits the system capacity (Klein, Irwin, & Roberto, 1991). Using joint detection (JD), inter-symbol interference (ISI), and multiple access interference (MAI) can be effectively eliminated, and as a result, the system capacity is improved (Peng et al., 2004). With the spatial location information, smart antenna can focus the transmit signal power in the direction of the target user, through which users will receive more useful signal power, and interference is highly compressed. This is another way to increase system capacity. However, what is the difference of the radio network planning between TDSCDMA and WCDMA/cdma2000? Based on the TDD mode, TD-SCDMA covers all application scenarios: voice and data services, packet and circuit-switched transmissions for symmetric and asymmetric traffic, pico, and micro and macro coverage for pedestrian and high mobility users. However, the adopted key techniques, such as time division duplex, smart antenna, multi-user joint detection (MUD), dynamic channel distribution, uplink synchronization, and baton handover, make the network design in TD-SCDMA have a significant different from WCDMA and cdma2000 (Peng et al., 2005). Unfortunately, to the best of our knowledge, there are still no papers investigating and proposing network design solutions for TD-SCDMA in which the impacts of key techniques have been considered. In this article, the key techniques impacting on the radio network design of TD-SCDMA are introduced. Meanwhile, some special radio network design issues for TD-SCDMA are presented.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
3G Commercial Deployment
Figure 1. TD-SCDMA frame structure
T
Figure 2. System block of TD-SCDMA base band data processing TX data
TX data TX data
Multiplexing Channel coding Interleaving Data mapping
Spreading Modulation
Digital combination
DAC DAC
Signaling Signaling RX data RX data
De-multiplexing
De-channel coding De-interleaving
De-spreading Demod Rx power estimation
SYSTEM BLOCK TD-SCDMA makes use of both TDMA and CDMA techniques such that channelization in TD-SCDMA is implemented using both time slots and signature codes to differentiate mobile terminals in a cell. The frame structure of TD-SCDMA is shown in Figure 1, where the hierarchy of four different layers—super-frame, radio frame, sub-frame, and time slot—are depicted. A sub-frame (5 ms) consists of 7 normal time slots and 3 special time slots, where TS0 is reserved for downlink and TS1 for uplink only, whereas the rest (TS2-TS6) should form two groups, the first (whose size can vary from 1 to 4 slots) for uplink and the second (whose size can vary from 4 to 1 slots) for downlink. The slot number ratio of the two groups can take 1/4, 2/3, 3/2, and 4/1 to suit particular traffic
RF Trx
Carrier Recovery
ADC
Synch
ADC
requirements. The agility in support of asymmetric traffic is a very attractive feature of TD-SCDMA and of particular importance for Internet services with rich multimedia content in 3G applications. The other three special time slots are downlink pilot (DwPTS), guard period (G), and uplink pilot (UpPTS), respectively. DwPTS and UpPTS are used as a synchronization channel (SCH) for downlink and uplink, respectively, which should be encoded by different pseudonoise (PN) codes to distinguish different base stations and mobiles. Meanwhile, the subscribers use the midamble part of every burst to estimate the channel impulse response. The subscribers in the same cell are assigned the same basic midamble with different time shift. Using the Steiner estimate principle, the base station can estimate the channel impulse response of all the subscribers simultaneously. The base band data processing is described in Figure 2. 941
3G Commercial Deployment
Figure 3. Explaining of TDD and FDD D D D D D Downlink
Frequency Division Duplex: UL band seperated from DL band
With asymmetric loads. Portions of the spectrum are occupied but not used
Uplink U
Time Division Duplex: UL band same as DL band D/U = Resource units
Up/Downlink symmetry can be adapted efficiently according to data load
Up/Downlink D D D D D U
Greater spectrum efficiency
KEY TECHNIQUES IN TD-SCDMA
Smart Antenna
The wireless access technique of TD-SCDMA is based on the TDMA (time division multiple access)/TDD, while WCDMA and cdma2000 are both based on the FDD. Some advanced techniques, such as smart antenna, MUD, and dynamic channel allocation (DCA), can be utilized more conveniently in TD-SCDMA than WCDMA and cdma2000, which can be categorized as the advantage of TD-SCDMA. Meanwhile, the uplink synchronization and handover have a stricter requirement, which can be regarded as the disadvantage in TD-SCDMA.
A smart antenna system is composed of N antenna elements, N related feed cables, and N coherent radio frequency (RF) transceivers. By using the A/D converters or D/A converters in the analog base-band (ABB), the receiver and transmitter analog signals are interfaced to the digital base-band (DBB) part over the high-speed data bus. The beam-forming which points to a particular user equipment (UE) can be obtained through smart antenna. Due to the inherent robustness of CDMA and the space diversity realized by smart antenna, the interference for multi-path propagation is greatly overcome, and the inter-symbol interface can be greatly reduced. For downlink, beam-forming can also reduce the interference to the other co-channel UEs. These performances can lead to the higher capacity of the TD-SCDMA system. Smart antennas employed by TD-SCDMA technology are not conventional diversity beam-switching antennas but advanced beam-forming (and beam-steering) bi-directional adaptive antenna arrays. The maximal individual directivity between base stations and mobile terminals is achieved by a concentric array of eight antenna elements with programmable electronic phase and amplitude relations. The terminals tracking is performed by the fast angle of arrival (AOA) measurements in intervals of 5 ms 200 times per second. The basic idea of the smart antenna is to track the user’s mobility, make the interference among different users compressed by spatial filtering, and enhance the desired signal received or focus the energy of the signal on the direction of the desired user location. Thus the system coverage, power efficiency, and system capacity can be improved. The most common antenna array geometry structures are uniform linear array (ULA) and uniform circle array (UCA), which is described in Figure 4.
CDMA Plus TDMA/TDD The TDD mode allows uplink and downlink on the same frequency band and does not require the pair frequency bands. In TDD, uplink and downlink are transmitted in the same frequency channel but at different times, and the difference between TDD and FDD is described in Figure 3. It is possible to change the duplex switching point and move capacity from uplink to downlink or vice versa, thus utilizing spectrum optimally. It allows for symmetric and asymmetric data services. TDMA is a digital technique that divides each frequency channel into multiple time-slots and thus allows transmission channels to be used by several subscribers at the same time. CDMA increases the traffic density in each cell by enabling simultaneous multiple-user access on the same radio channel. In order to decrease the interference sourcing from the MAI, smart antenna and MUD are utilized. Since TD-SCDMA is based on TDMA/TDD, the air interface for both uplink and downlink is interoperable, which make smart antenna efficiency.
942
3G Commercial Deployment
Figure 4. Basic principle of uniform circle array
beamforming. For the phased array, the weighted combining only do some phase rotation on the received signals. Switch-beam is a simplification of the phased array, in which the whole cell is divided into several sectors and a set of pre-defined beam-forming weights are used to cover these sectors one by one. The rationale of the Eigen-beamforming is that the maximum Eigen value and Eigen vector of the spatial correlation matrix of the user is found by Eigen value decomposition as the power gain and beamforming weight. The performance of the switch-beam is little worse than that of phased array, and the Eigen-beamforming is expected to be the best. In the current TD-SCDMA system, the phased array smart antenna is utilized to suppress the interference from both the inter-cell and intra-cell. However, if the air condition is too bad and the AOA estimation is incorrect, the smart antenna cannot work efficiently. In this way, the multi-user joint detection technique is adopted to suppress only the intra-cell interference.
Code 1
Code 2
Code 3
Smart antenna
Code 16
Rf, BB, JD_Unit and DSP-Unit Code 1 Code 2 Code 3 ....... Code 16
Multi-User Joint Detection The beam pattern of the ULA is symmetrical along the boresight of the antenna array, that is to say ULA cannot distinguish the users that locate at the symmetry location along the boresight of the antenna array. The more the user location deviates the boresight of the antenna array, the wider the beam and the lower the resolution of the antenna array will be. For UCA, as the distance between the elements increases, the aperture of the antenna array becomes larger, and the resolution of AOA increases; meanwhile the amplitude of the side lobe becomes larger. Compared to the ULA, no phase blur happens. So except the sector scenario, UCA will be adopted. Three types of beam-forming antennas are introduced in TD-SCDMA: phased array, switch-beam, and Eigen-
Multi-user joint detection (MUD) allows the receiver to estimate the radio channel and works for all signals simultaneously. Through the parallel processing of individual traffic streams, MUD eliminates the MAI and minimizes intra-cell interference, thus increasing the transmission capacity. Figure 5 describes the basic principle of MUD. The first step is that all signals from the carious UEs are received in the node B receiver through the radio frequency (RF)/baseband (BB). The second step is to detect the signals barely emerging from the MAI with a low signal to noise ratio (SNR). In the last step, using a specific algorithm, a DSP thus extracts all CDMA channels in parallel and removes the interference caused by the undesired CDMA channels. The result is a clear signal (high signal to noise ratio) for each CDMA code.
Figure 5. Basic principle of MUD UE 1
1
3
2
UE 2 UE 3 UE 4
RF/BB Unit
Matched Filter Correlator
Joint Detection Unit
Data Output Signals
UE 5 UE 6 UE 7 UE 8
Channel Estimation
Code Program
943
T
3G Commercial Deployment
Note that the efficiency of MUD in TD-SCDMA technology is based on the TDMA/TDD operation and on the limited number of codes employed. The total number of users per radio carrier is distributed over the different time slots of the basic TDMA frame, so that a maximal number of 16 codes per time slot per radio carrier can be easily processed in parallel and detected. However, due to the huge number of spread codes used by WCDMA and cdma2000, the implementation of an optimal multi-user receiver in these systems is difficult, since the implementation complexity is an exponential function of the numbers of codes. In order to combat MAI, both WCDMA and cdma2000 systems employ the suboptimal detection schemes, such as the rake receiver, which do not extract all CDMA codes in parallel.
Dynamic Channel Allocation A further minimization of inter-cell interference is achieved by dynamic channel allocation (DCA). There are different radio resource dimensions for TD-SCDMA: TDMA, FDMA, CDMA, and SDMA (space division multiple access). Making an optimal use of these degrees of freedom, DCA provides an adaptive allocation of the radio resources according to the interference scenario, minimizing the inter-cell interference (Peng et al., 2003). In TD-SCDMA, the DCA is divided into two parts: Slow DCA allocates resources to cells, while Fast DCA allocates resources to the bearer services, balances the load in the different slot/frequency, and congregates the radio resource for supporting the high bit rate services. Both UEs and node Bs perform the periodic monitoring and reporting to support DCA. Fast DCA is always terminated at the node B, but slow DCA can be terminated at any network entity above the node-Bs that forms the seamless coverage area. The slow DCA algorithm allocates the radio resource units in a cell-related preference list for Fast DCA to acquire them for different bearers. In the first phase the cell-related preference list is a fixed table that is given as a parameter for each base station. Fast DCA resides in each node B and is responsible for utilizing the slots assigned to the node B in the most efficient manner possible. This involves: (1) assigning each UE to the slot(s) best suited in each particular case at the start of a connection, and (2) reshuffling/reallocation of UEs when traffic and/or environmental conditions have changed or if the requirements of an existing connection have been altered by the UE. According to the asymmetry of transmission bits, the switched point between uplink and downlink is different in the adjacent cells in which the cross slot interference sourcing from the node B-node B is huge. In order to suppress the cross slot interference, the Fast DCA combines the smart antenna technique to allocate the radio resource in the cross slots to the UEs close to the serving node B. In this way, the
944
total transmission power in the cross slots decreases and the interference is suppressed.
Uplink Synchronization Like all TDMA systems (e.g., GSM), TD-SCDMA needs an accurate synchronization between mobile terminal and base station—that is, the UEs’ spread signals arriving at the node B at the same time can effectively simplify the demodulator in node B, decrease MAI, and increase the system capacity. This synchronization becomes more complex through the mobility of the subscribers, because they can stay at varying distances from the node B, and their signals present varying propagation times. An uplink synchronization procedure includes two stages: synchronization establishment and synchronization maintenance. The synchronization establishment is often associated with UE’s access procedure, and the synchronization maintenance is often associated with the dedicated communication procedure between UE and network. Due to the multipath and shadow fading, however, the establishment and maintenance of the uplink synchronization between different UEs in a cell are difficult. The simple, effective, and low-cost establishing and maintaining uplink synchronization solution will bring benefits to TD-SCDMA. The UL synchronization is equivalent to a very high precise timing advance according to the physical layer specifications. Therefore, an extended time advanced (TA) option by means of a sub-chip granular operation is utilized in TD-SCDMA. The granularity of TA, in the case of UL synchronization, is ±1/8 chips.
Baton Handover The conventional hard handover has some shortages, for example the dropping ratio of the hard handover is high, and the efficiency of the radio resource usage of soft handover is low. With smart antenna and uplink synchronization, the handover strategy will change and the performance will be improved in TD-SCDMA. The UE position information and AOA are provided to predict the handover request, prepare for handover and pre-synchronize for handover, shorten handover time, decrease handover blocking, simplify handover procedure, and improve handover confidence. This is named baton handover in TD-SCDMA (Peng et al., 2003). The basic principles of baton handover are: (1) the system knows the position of all UE, (2) the system knows and determines the target cell for handover, (3) the system informs mobile the information about the node B in neighboring cells, (4) mobile measurement helps the system to make the final decision, and (5) after the cell search procedure, the mobile UT has already established. Since UE can synchronize to the node-B in target cell, the baton handover costs a shorter handover time period
3G Commercial Deployment
for both inside the TD-SCDMA system and between TDSCDMA and the different systems. In the TD-SCDMA system, the parameters that UE measure are not only the received signal power level, but also their transmission time offset and so on.
TD-SCDMA RADIO NETWORK DESIGN Since the TD-SCDMA system adopts many advanced techniques, there are some unique issues when doing TDSCDMA radio network designs (Peng et al., 2005). First, the scramble codes in TD-SCDMA are only 16 chip lengths, and the orthogonality between different scramble codes is easy to lose which will cause difficulty in pilot searching and reception. So, the scramble code planning is especially important in a TD-SCDMA system. Second, as TD-SCDMA is a narrowband system, its bandwidth is one-third of that of WCDMA, and the capacity of single carrier TD-SCDMA is low compared to WCDMA. That means if TD-SCDMA occupies the same bandwidth as WCDMA, the multi-carrier system can be adopted to form a higher capacity (Peng, Chen, & Wang, 2006). Besides, multi-carrier with careful configuration of carrier can make scramble codes planning much easier. Multi-carrier TD-SCDMA layout is proposed to form a network and provide many kinds of mobile services. Third, coverage of a certain system is of great importance. It can be analyzed in a theoretical way by link budget. Link budget of TD-SCDMA differs from that of CDMA2000 and WCDMA, due to its own properties and the advanced techniques that TD-SCDMA employs. Finally, coexistence of TD-SCDMA with other 3G systems is another important issue both for TD-SCDMA and other related system performances (Peng et al., 2004).
CONCLUSION In TD-SCDMA only 16 codes for each timeslot for each carrier are used. The intra-cell interference is eliminated by MUD, and inter-cell interference is minimized by the joint use of smart antennas and DCA. Meanwhile, since most interference is suppressed under the condition of smart antenna working efficiently, the capacity is radio resource limited, and the cell breathing effect is not an issue anymore. However, some new problems occur due to the key techniques and features of TD-SCDMA, such as the N frequency planning, scrambling code planning, multi-operators coexistence planning, and advanced radio resource management. In order to successfully deploy the commercial TDSCDMA network, the key techniques impacting on the network performance and the novel network strategies should be proposed. The special network planning method
and project steps should be configured for TD-SCDMA. This article discussed all these confusing problems in TDSCDMA network planning and presented the principle solutions. Some special issues should be investigated, such as the handover strategies between TD-SCDMA and other mobile communication systems, indoor network design, and repeater design specified for TD-SCDMA.
REFERENCES Chen, H. H., Fan, C. X., & Lu, W. W. (2002). China’s perspectives on 3G mobile communications and beyond: TD-SCDMA technology. Wireless Communications, 9(2), 48-59. Klein, S. G., Irwin, M. J., & Roberto, P. (1991). On the capacity of a cellular CDMA system. Vehicular Technology, 40(2), 303-312. Peng, M. G., Bao, W., Hu, W., & Wang, W. B. (2004). Investigation of uplink admission control schemes for TDD-CDMA systems. Proceedings of the International Conference on Communications, Circuits and Systems (pp. 443-446). Chendu, China. Peng, M. G., Chen, S. P., & Wang, W. B. (2006). TD-SCDMA evolution and multi-carrier techniques. Telecommunication Science, 22(5). Peng, M. G., Hu, W., & Wang, W. B. (2004). Investigation of uplink capacity based on the background noise floor in TDD-CDMA systems. Proceedings of the International Conference on Signal Processing (Vol. 3, pp. 1918-1921). Beijing, China. Peng, M. G., Huang, B., & Wang, W. B. (2004a). TDDCDMA capacity loss due to adjacent channel interference in the macro environment employing smart antenna techniques. Proceedings of the 2004 Asia-Pacific Radio Science Conference (pp. 146-149). Qingdao, China. Peng, M. G., Huang, B., & Wang, W. B. (2004b). Investigation of TDD and FDD CDMA coexistence in the macro environment employing smart antenna techniques. Proceedings of the 5th International Symposium on Multi-Dimensional Mobile Communications, 2004 Joint Conference of the 10th AsiaPacific Conference (Vol. 1, pp. 43-47). Beijing, China. Peng, M. G., & Wang, W. B. (2003). Novel approaches for downlink performance analysis in CDMA networks. Proceedings of PIMRC 2003 (Vol. 3, pp. 2533-2537). Beijing, China. Peng, M. G., & Wang, W. B. (2004a). Advanced HARQ and scheduler schemes in TDD-CDMA HSDPA systems. Proceedings of the 5th International Symposium on Multi945
T
3G Commercial Deployment
Dimensional Mobile Communications, 2004 Joint Conference of the 10th Asia-Pacific Conference (Vol. 1, pp. 67-70). Beijing, China.
frequency channel assignment schemes. Proceedings of the International Conference on Communication Technology (Vol. 2, pp. 803-807). Beijing China.
Peng, M. G., & Wang, W. B. (2004b). An analysis of resource allocation and management in TDD-CDMA systems employing smart antennas. Emerging Technologies: Frontiers of Mobile and Wireless Communication, 2, 753-756.
Peng, M. G., Zhang, J. W., Zhu, X. M., & Wang, W. B. (2003). A novel dynamic channel allocation scheme to support asymmetrical services in TDD-CDMA systems. Proceedings of the International Conference on Communication Technology (Vol. 2, pp. 794-798). Beijing, China.
Peng, M. G., & Wang, W. B. (2004c). TDD-CDMA uplink capacity investigation in the background noise floor. Proceedings of the IEEE International Conference on Multimedia and Expo (pp. 233-236). Peng, M. G., & Wang, W. B. (2005a). A framework for investigating radio resource management algorithms in TD-SCDMA systems. IEEE Communication Magazine, 43(6), 12-18. Peng, M. G., & Wang, W. B. (2005b). Comparison of capacity between adaptive tracking and switched beam smart antenna techniques in TDD-CDMA systems. Proceedings of Mape2005, Beijing, China. Peng, M. G., & Wang, W. B. (2005c). Investigation of handover strategies in TDD-CDMA cellular networks. Proceedings of the ACM SIGCOMM Asia Workshop 2005, Beijing, China. Peng, M. G., & Wang, W. B. (2005d). Investigation of the distributed antenna scheme for multi-cell environment in TDD-CDMA systems. Proceedings of ITC2005, Beijing, China. Peng, M. G., & Wang, W. B. (2005e). Investigation of uplink performances based on the switched beam antenna scheme in TDD-CDMA systems. Proceedings of ITC2005, Beijing, China. Peng, M. G., & Wang, W. B. (2005f). TD-SCDMA mobile communication system. Beijing: China Machine Press. Peng, M. G., & Wang, W. B. (2005g). TD-SCDMA network planning. Telecommunication Technology, 5, 6-9. Peng, M. G., Wu, Y. C., & Wang, W. B. (2004, May 1719). Joint and advanced proportionally fair scheduling and rate adaptation for multi-services in TDD-CDMA systems. Proceedings of VTC 2004 (Vol. 3, pp. 1630-1634). Milan, Italy. Peng, M. G., Zhang, J. W., Hu, C. J., & Wang, W. B. (2003). Handover performance analysis in TDD-CDMA cellular network. Proceedings of WCNC 2003 (Vol. 2, pp. 806-811). Peng, M. G., Zhang, J. W., Liu, Y., & Wang, W. B. (2003). On the capacity of a cellular TDD-CDMA system employing
946
KEY TERMS Code Division Multiple Access (CDMA): A kind of multi-access method. Users are distinguished in code domain. Other well-known multiple access methods are FDMA (frequency division multiple access), in which frequencies are used to distinguish different users; and TDMA (time division multiple access), in which time is used to distinguish different users. DCA: Another key technique that a TD-SCDMA system holds, by which radio resources are allocated to different traffic bears of different cells. Multiple Access Interference (MAI): The main kind of interference in a CDMA system. It has a bad impact on CDMA system capacity. Multi-User joint Detection (MUD): A kind of detection method at receiver. Other common detection methods are rake receiver and matching receiver. Radio Network Design: Design of network framework, layout, and so on. This is a key for a system to be commercially feasible. Smart Antenna: A key technique of a TD-SCDMA system which can enlarge received signal power and suppress interference, thus improving received SNR (signal to noise ratio) and enlarging system capacity. System Capacity: The number of users of certain traffic that the system simultaneously holds. Other definitions are the number of Erlang that a system holds simultaneously. Time Division Duplex (TDD): A kind of duplex mode. Another well-known duplex is FDD (frequency division duplex) mode. Uplink (from user to network) and downlink (from network to user) in TDD mode are divided by time and in FDD by frequency. Time Division duplex-Synchronous Code Division Multi-Access (TD-SCDMA): A 3G standard proposed by the China Academy of Telecommunications Technology (CATT) and Siemens jointly, and accepted by the ITU in 1999.
Category: Mobile Software Engineering 947
Transaction Management in Mobile Databases Ziyad Tariq Abdul-Mehdi Multimedia University, Malaysia Ali Bin Mamat Universiti Putra Malaysia, Malaysia Hamidah Ibrahim Universiti Putra Malaysia, Malaysia Mustafa M. Dirs College University Technology Tun Hussein Onn, Malaysia
INTRODUCTION Recent advances in wireless communications and computer technology have provided users the opportunity to access information and services regardless of their physical location or movement behavior. In the context of database applications, these mobile users should have the ability to both query and update public, private, and corporate databases. The main goal of mobile software research is to provide as much functionality of network computing as possible within the limits of the mobile computer’s capabilities. Consequently, transaction processing and efficient update techniques for mobile and disconnected operations have been very popular. In this article, we present the main architecture of mobile transactions and the characteristics with a database perspective. Some of the extensive transaction models and transaction processing for mobile computing are discussed with their underlying assumptions. A brief comparison of the models is also included
TRANSACTION MANAGEMENT IN MOBILE DATABASES A mobile database system is a special multi-database system on a mobile computing environment. It allows mobile hosts to access and manipulate data stored on several per-existing, autonomous, and heterogeneous local database systems located on different parts of the wired network. Transactions in a mobile database system may access data from several local databases at different sites. Management of these transactions requires different approaches in mobile databases than in a multi-database. This is mainly due to the fact that a mobile host is not suitable to manage a global transaction by itself due to the described nature of the mobile computing environment. Usually this management is done by the mobile host’s base station or by coordination of it.
Due to the described nature of the mobile computing environments, transaction management has to be reevaluated for mobile databases. The transactions in mobile computing environments are usually long-living transactions, possibly covering one or more disconnected durations. Supporting disconnected operation (i.e., allowing a mobile host to operate autonomously during disconnection) raises issues in consistency. Providing disconnected operation also requires some pre-caching of data that will be required for the necessary operations to be performed during disconnection. The moving behavior of the transactions in mobile computing environments also requires new mechanisms. As a mobile host moves from a cell to another cell, its transactions might need to migrate from one base station to another. In general, transactions in mobile databases require relaxed ACID properties. There are several works on mobile transactions, each addressing some of the issues in mobile transaction management. We will explain some of them in the following sections.
Kangaroo Transactions Kangaroo transactions (KTs) are introduced in Dunham and Helal (1997). As the name suggests, this model mostly addresses the moving behavior of the mobile transactions. As the transactions hop from one site to another, the management of the transaction also moves. In addition to the mobile computing environment we have described, these systems introduce a couple of other terminologies. The term source system represents a collection of systems that offer information services to mobile users. These systems could be any type of system that exists in the mobile computing environment. One good example is a distributed database system. The term data access agent (DAA) represents an agent that is hosted by each base station. Mobile hosts reach data in source systems by sending their transactions to DAAs. When a handoff occurs, the DAA at
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
T
Transaction Management in Mobile Databases
the new base station receives the transaction information from the DAA in the old base station. A mobile transaction is defined as the basic unit of computation in the mobile environment. The management of a mobile transaction might hop through different base stations, which are not known until it completes its execution. DAAs at base stations are responsible for management of the mobile transactions. One part of the DAA responsible for the management of the transactions is called the mobile transaction manager (MTM). The main responsibilities of an MTM are maintaining the status of mobile transactions in execution, logging recovery information, and performing needed check pointing. In this model it is assumed that a mobile transaction issued by a mobile host to a DAA might include several subtransactions that require access to data at several global database systems (GDBSs) and DBMSs residing at different places of the fixed network. As a result DAAs serve as a mobile transaction manager built on top of GDBSs and DBMSs. DAAs also keep log information about the mobile transaction parts that have executed on them. Remember that a mobile transaction changes its DAA as it moves from one cell to another. Since mobile transactions are long lived and include possible disconnected durations; the atomicity of a mobile transaction in this model is not always guaranteed. The time between an interruption of a transaction and its resume could be quite long. As a result, it is valuable to commit early on some portions of a mobile transaction, while breaking the atomicity property of the transactions. These early commits enable the release of possible important resources, instead of holding them for a long time. A mobile transaction in this model, which is called a kangaroo transaction, is an extension to global transactions (GTs). Figure 1 shows a global transaction, which consists of subtransactions called local transactions (LTs). Each local transaction is assumed to be issued to a DMBS. A kangaroo transaction can be composed of both GTs and LTs. The mixture of GTs and LTs are grouped under a transaction type called Joey transactions (JTs) based on the DAA on which they have initiated. Figure 2 shows an example kangaroo transaction. Each JT represents the unit of execution at one base station. When a mobile host makes a transaction request to
the DAA on its associated base station, a KT is formed. In addition to that, a JT is formed for managing subtransactions that originate from the mobile host when the KT is under the control of the first DAA. When a mobile host hops from one cell to another, the control of its KTs changes to the new DAA on the new base station. The new DAA creates a new JT for handling the future subtransactions the mobile host might request to this DAA. The old JT is committed independently from the new one. Note that this breaks atomicity. To enable a KT to be completely undone, previously committed transactions should be compensated. Kangaroo transactions have two different modes of execution. The first mode is called compensating mode where a JT falls and all KT is undone. However, this mode of operation requires compensating transactions for undoing operations of the previous JTs, since they are independently committed building compensating transactions requiring input from the user. As a result this mode is rarely used. Note that this mode tries to preserve atomicity of the KT, which breaks the durability of the subtransactions. The second mode, which is the default mode, is called the split mode. In this mode of operation, when a JT fails, no new JTs are created, but the previously committed Ms are also not undone. This mode breaks the atomicity. None of the modes ensures serializability.
Clustered Model In Pitoura and Bhargava (1995), the sites are grouped in clusters if they are connected by strong network links or they are pre-grouped in clusters. In a cluster, full consistency is always enforced, however over the clusters a bounded inconsistency is permitted. A site can be a member of a cluster or leave one dynamically based oil the network conditions, for example, a mobile host forms a cluster itself when it is disconnected. Mobile transactions are grouped into two types: strict and weak transactions. Weak transactions access data in the same cluster, and they have two commit points: cluster and global. Global commit can only be made after clusters merge. Strict transaction can only access data that is ac-
Figure 2. Figure 1. KT
GT JT 1
LT 1
948
Hop
JT 2
JT 3
LT 2 GT11
op11
Hop
op12
op21
op22
op23
GT12 LT12
GT21
GT22
GT31
Transaction Management in Mobile Databases
cessed by a strict transaction or globally committed weak transaction. They are not allowed to access inconsistent data. Applications updating private data-datamostly updated by the mobile host on itselfcan use weak transactions during disconnections. Applications accessing global data can use weak transactions also if they are tolerant of inconsistent data to some level.
A Pre-Serialization Transaction Management Model This model (Dircke & Gruenwald, 2000) also assumes a fixed and attached mobile network like the kangaroo model. MUs connect to MSSs. MUs access the mobile multi-database system (MMDBS) via an interface called global transaction manager (GTM). GTM is responsible for providing consistent and reliable units of computing. GTM is a group of global transaction coordinators (GTCs) and site transaction managers (STMs). A GTC runs on every MSS. There exists an STM on each local database. GTC handles disconnections and logging, and delivers the transactions and transaction compensators to the STMs. The STM returns the commit or abort results to the GTC once the transaction submitted to it is completed. In case of a global abort, the compensators on STMs are used to roll-back the transactions. GTC tries to forward the result to the MU or saves it if the MU is disconnected. GTC also tries to conclude if an MU is disconnected for a short time or failed for a prolonged period. Transactions of MUs that are failed are aborted as they become obstacles to other transactions. A transaction is a group of compensatable, open-nested, vital and non-vital subtransactions. All subtransactions can commit independently. After all the vital subtransactions are committed, a transaction is checked for atomicity and isolation (A/I). After this point if that transaction is not aborted, it can submit only non-vital subtransactions. If any vital subtransaction aborts, then the transaction is aborted. An algorithm called partial global serialization graph (PGSG) algorithm is used to check the A/I constraints. Each STM maintains a local serialization graph for vital transactions. These graphs are forwarded to GTCs upon request. After merging these graphs if there is no cycle then the transaction is toggled (marked checked), else it is aborted and all subtransactions are compensated. A toggled transaction is guaranteed to commit if its mobile station is alive or only disconnected for a small amount of time. If a catastrophic failure has occurred, then a toggled transaction can be aborted if it obstructs other transactions. This model allows subtransactions to commit independently, thus allowing them to release resources as they are not needed. Also a suspended state where transactions can be aborted is introduced. In case of a catastrophic failure, an MU is assumed to be in the suspended state. However,
determining if an MU has failed should be done carefully, so not to abort transactions of a disconnected host.
Deno In Cetintemel and Kelender (2002), Deno, a replicated object storage system designed for use in mobile and weakly connected environments, is discussed. Deno is designed to support weak connections and limitations of mobile hosts, such as limited processing power and limited coverage area. Thus, it is a lightweight, peer-to-peer, fully decentralized, and asynchronous system. An MU does not need to know the other hosts, but it needs to be in contact with at least one other node. Pair-wise anti-entropy sessions are used to spread informationvotes in this case; update transactions are always voted to commit after they are committed locally on a node. This scheme works as follows: if two nodes x and y can communicate with each other, x can ask y if it knows any globally committed transactions it does not know, if so it copies this information. If y does not know anything, it just sends commit candidates it voted for to x and casts its own vote. Any voter keeps track of a number of votes for an object k and the number of unknown votes. Since currencies (weights) are distributed such that the total is 1.0, any node can keep track of unknown votes. If on any MU the votes for k are more than the votes for j and unknown, then it can commit k. This method ensures that all updates are committed in the same order globally. In the case of planned disconnections (e.g., sleep mode), a node can appoint a proxy for itself and transfers its currency to that proxy. After returning to the network, the formerly disconnected node can claim its currency back. In case of an unplanned disconnection, a proxy for the disconnected node is selected using the voting scheme described above. If the proxy election is globally committed, the currency (weight) of the disconnected node is transferred to the elected proxy. In the case of network partitioning, only proxies for the partition that has the smaller total currency can be selected because the proxy election is done using the voting scheme described above. As in the planned case, any re-connected MU can claim its currency. Weight assignment is crucial to designing a good working scheme; basically, assigning more weights to strongly connected hosts (e.g., stable and more powerful hosts) is a good strategy. Currency distribution is discussed in detail in Cetintemel and Kelender (2002).
Pro-Motion Model The pro-motion model (Walborn & Chrysanthis, 1996) is designed to support disconnected transaction processing. The motivation is that disconnected MUs can execute transactions if they have the data and methods required. 949
T
Transaction Management in Mobile Databases
The fundamental building block is the compact, which is the basic unit of data replication for caching and hoarding. A compact is not only a piece of data, but it is an object that includes restrictions (allowable operations), obligations (such as an expiration time), and methods. In other words compact is a mini database that is moved to the local hosts upon request. A database server delegates an MU as the controller of a compact for local transaction processing. The MU should agree on all obligations and restrictions that the database server sets by the database. Pro-motion uses open-nested split transactions. When an MU is connected to the network, it identifies a group of compacts that are updated by locally committed subtransactions. Those transactions are split from the uncommitted ones and sent to the owner of the compact. Those transactions are then committed on the database, making the updates visible to all other transactions. An MU first caches the compacts it needs, then disconnects and processes the transactions. When reconnected it resynchronizes with the fixed database. Transactions can be of two typeslocal and traditional. Local transaction results are made visible when they commit to the other transactions on the same MU. A transaction can also be traditional, which means its results are invisible until resynchronization. To allow the resources to be released in a timely manner, each compact is assigned a deadline. An MU can request an extension if the deadline has passed. If the compact is free, then a new deadline is negotiated; if not, this compact is marked invalid and all transactions accessed are aborted. Also the other compacts written by these transactions are marked unavailable, and transactions read from unavailable compacts are aborted. The valid compacts are resynchronized with the databases. The operations made on the compact are sent to the database servers. These operations are executed as a single transaction. If this transaction can complete, then a commit message is returned to the MU, else the MU receives an abort message and executes compensating procedures for the committed local transactions’ owners. A 10-level scale of correctness is defined in pro-motion from serial to no-guarantee. Each method in a compact is assigned a level. Each transaction is also assigned a minimal level for READ and WRITE operations, and it can only perform operations at that level and READ operations on a higher level than WRITE operations.
TCOT Protocol Kumar, Prabhu, Dunham, and Seydim (2002) present a timeout-based commitment protocol, TCOT. Like Deno, it is designed for weakly connected, less powerful MUs. TCOT tries to minimize the communication, since bandwidth is scarce and is non-blocking since MUs can disconnect un-
950
predictably. It assumes that MU has a cache and transaction processing power. A coordinator (CO) is responsible for a transaction submitted by an MU. COs are either base stations (BSs) or a node on the fixed network. The MU extracts the subtransaction ti,j; it will run from transaction Ti = {ti,j|l ≤ n ≤ n} and sends Ti - ti,j to the CO. MU sends extra information such as how long it will take to process the subtransaction and to ship the updates to the CO. The CO distributes the subtransaction to the fixed nodes on the network and starts its timer according to the values it received from the MU. Any node executing a subtransaction can request a time extension. But if the CO does not receive a message from the MU or the other participating nodes, then it aborts the transaction and sends an abort message to all of the nodes. Else, if it receives commit messages from all nodes and updates from MU, it does not send any further messages. Since some nodes commit without a global commit, compensating transactions are necessary to undo the globally aborted but locally committed subtransactions. Choosing or calculating timeout values is important through this scheme; extension messages are minimized and the transaction restar MUs should be able to choose values for initial timeout values based on the network conditions. Moreover they can request extensions from those incremented with each request. The CO should be able to reject an extension request if the system throughput falls tinder-desired value.
MANET Model A mobile ad-hoc network (MANET) has restrictions, which make models like TCOT and kangaroo transactions infeasible to use. MANETs lack nodes which are on fixed networks and BSs which can act as CO or DAA. Moreover, since all nodes route traffic, if they fail the network starts to weaken, partitions can occur, and usable bandwidth between the nodes starts to drop. Gruenwald and Banik (2001) propose a model to tackle these problems. It is assumed that there are two types of nodes, large mobile hosts (LMHs) and small mobile hosts (SMHs). LMHs have more processing power, storage capacity, and power source than SMHs. An SMH sends its transaction to an LMH depending on the transaction type. If the transaction is a firm transaction, it should be finished before its deadline or it has no value. If the transaction is soft, then it has two deadlines. The earlier deadline can be violated, but the value of the transaction starts to decrease to zero towards its second deadline. A fixed transaction is always processed by the nearest LMH, whereas a soft transaction is processed by the LMH with the highest remaining energy level. LMH distributes the subtransactions to other LMHs upon submission. The
Transaction Management in Mobile Databases
subtransactions are also grouped into two sets: vitals and non-vitals. After all vital subtransactions finish, the transaction is verified against atomicity and isolation using PGSG (Dircke & Gruenwald, 2000), as described earlier. If the transaction does not violate these conditions, it would be marked as ready and toggled to commit, and it starts to wait for the non-vital subtransactions. A toggled transaction is guaranteed to commit unless it blocks another global transaction while it is in the suspended state. A transaction is in suspended state if its MU disconnects. The LMH tries to send the result to the originating the SMH, but if it fails while sending the result of a firm transaction, it aborts the transaction. However, if it fails while trying to send a soft transaction, it retries until the second deadline is reached. While a transaction is executing, the SMH can be in doze mode or sleep mode to conserve energy. If it is waiting for a result of fixed transaction, it should not go into the sleep mode for a long time, otherwise the transaction would be aborted. If it is waiting for a soft transaction, it can sleep until the end of the second deadline, saving energy.
Planned Disconnection Modes Planned disconnection modes (Datta et al., 1999; Demers et al., 1994) involve informing the distributed system of the intention to disconnect and may include the appointing of a proxy. The purpose of a planned disconnection procedure is to enable the remaining connected sites to continue processing with minimal disruption. There are a number of different ways that a disconnection can affect the database. In this section, we explore the possible kinds of planned disconnection and provide appropriate terminology. In basic sign-off mode (Datta et al., 1999), an MH decides to disconnect and informs the system, consisting of the currently connected sites, of its intention. The database of the disconnected MH becomes read-only while the access capabilities of the remaining connected sites are unaffected. In check-out mode (Demers et al., 1994), the MH wants to disconnect and be able to update a set of data items X. There are three variations to this mode that determine what type of access to non-checked-out items is allowed. The first variation is DB partition. In DB partition mode, the database is partitioned into X and DB−X. The disconnected site has complete and unlimited access to X and nothing else, while the remaining system has complete access to DB−X and nothing else. The second variation is check-out with mobile read. This mode allows the disconnected site to have read access to all database items in addition to the read/write access to the checked-out items X. The remaining connected sites in the system have complete (read/write) access to DB−X and no access to X.
The third variation is check-out with system read. This mode allows the connected sites in the system to have read access to all database items in addition to the read/write access to the non-checked-out items DB−X. In check-out mode, when the MH checks out an object, either the MH or the remaining connected sites are prevented from accessing to read some of the objects in the database. Although this is necessary to preserve serializability, many database systems operate on lesser degrees of isolation (Chrysanthis et al., 1994). Therefore, we define a relaxed check-out mode in which the remaining sites can read the items that other sites have checked out, while disconnected sites can read items they have not checked out.
COMPARISON OF MODELS Kangaroo, pre-serialization, TCOT, and pro-motion models are designed for multi-tiered networks. It is assumed that there is a fixed and reliable wired network that can support mobile hosts. Moreover, all local databases are assumed to be located in the fixed network, which makes replication management and deadlock detection algorithms easier to implement. These models mentioned above are all making use of a similar notion called DAAs. DAAs execute the transaction for mobile units, and since they reside on the fixed network, they are never disconnected. Pro-motion is different from the other terms in that mobile units can execute transactions by themselves, which is an advantage if they are working on small units of data and they need user interaction. In other systems user interaction is not considered. If a transaction needs user input, it should wait for the originating mobile node to be disconnected. A drawback of pro-motion is that it assumes that wireless bandwidth is not a scarce resource, since it moves data back and forth. Kangaroo is the only modal that addresses the mobility issue. Although in the other models mobility can be handled similarly to kangaroo, it is not explicitly addressed. In the models mentioned the problem of failed hosts is handled differently. TCOT uses time-outs to abort transactions that can hold the resources for a long time. In pre-serialization, transaction of unresponsive nodes are not aborted immediately, rather they are aborted as new transactions need the resources held by those suspended. In pro-motion the data sent to mobile hosts are attached an expiration time so it is guaranteed that they will be released without blocking the other transactions. While using time-outs can be a problem because of the unpredictable transaction durations and messages exchanged for extension, it would be easier to predict if a node is dead for a long time or not. None of these models tries to minimize the power used by mobile nodes. Moreover, in pro-motion mobile nodes 951
T
Transaction Management in Mobile Databases
execute transactions and exchange large amounts of data compared to the other ones. The other models try to minimize the messages passed on the wireless network, which in turn means mobile nodes will use less messages and power. Models like Deno, clustered, and MANET are different from the models mentioned above in that they do not assume that there is a fixed network. Deno is the most interesting model among all the models. It does not make any assumptions about the network topology, node capacities, and connection characteristics. It is a fully distributed model and it can handle network partitions seamlessly. However, it does not take advantage of the more powerful nodes if there are any in the network. The MANET model assumes that there are more powerful nodes in the network. This is not an unreasonable assumption for the mobile networks. MANET can be used to take advantage of the wired network if there is any. It also tries to distribute the energy usage homogeneously, which is unique for all the models. Since this model assumes more powerful and always-connected nodes, disconnections can be handled similarly to the models mentioned above. The clustered model is similar to Deno for that: it does not make assumptions about network topology. It also defines two types of transactions: weak and strong. Weak transaction can proceed when the network is partitioned. The planned disconnected model supports transaction management in a disconnection mobile database, and increases flexibility and allows the distributed database to take advantage of mobility and use new ways. The planned disconnected protocol produces executions that are one-copy serializable. Briefly, it can be argued that, since all of the transactions of the disconnected site are read-only, the values of the data that are read are those of a snapshot taken at the time of disconnection. All of the readonly transactions of the disconnected site can be serialized at the time of disconnection. The check-out mode in a planned disconnected model with system reads produces executions that are one-copy serializable. The locked data items are modified by the transactions at the disconnected site, and these transactions are serialized with respect to each other because of local two-phase locking. They are serialized with respect to the transactions of the rest of the system at the point in time of reconnection (rather than the point of disconnection, as for check-out with mobile read).
space, processing speed, and the fact that the computing is distributed made mobile databases subtle and prone to more difficulties. The ACID model suggests that, in some situations, a client requests a sequence of separate requests to a server to be atomic, provided that they are free from interference by operations being performed on behalf of other concurrent clients; and either all of the operations must be completed successfully or have no effect at all in case of server crash. Many of the transaction management models start from these concepts that ACID suggests. They normally use relaxing ACID as an architectural model. The kangaroo model, however, suggests that a new mobile transaction definition is needed, specific to the mobile computing proposed. Each of the transaction management models makes its own assumptions about the infrastructure needed to support the respective model. For example the pre-serialization transaction management model and kangaroo model both give a very general architecture by which mobile computing can be performed in a heterogeneous multi-database environment. The constantly decreasing price of mobile devices is leading to a revolution in the field of mobile computing. The set of enhanced services currently provided to the owners of PDAs is expected to reach the users of the next generation of mobile phones and other mobile devices. A set of very powerful applications is expected to support this revolution, ranging from simply text transfer up to even multimedia transfer.
REFERENCES Cetintemel, U., & Kelender, P. (2002). Lightweight currency management mechanism in mobile and weakly-connected environment,. Journal of Distributed and Parallel Database, 11(January), 53-71. Dircke, R., & Gruenwald, L. (2000). A Pre-Serialization transaction management technique for mobile multi-database. ACM Mobile Networks and Applications, 5(December), 311-321. Dunham, M. H., & Helal, A. (1997). A mobile transaction model that captures both the data and the movement behavior. ACM/Baltzer Journal on Special Topics in Mobile Networks and Applications, 2, 149-162.
SUMMARY AND CONCLUSION
Gary, J., & Reuter, A., (1993). Transaction processing: Concepts and technique. San Francisco: Morgan Kaufman.
The aim of this article was to present a comparison between a few of the transaction management models in distributed mobile databases. The study suggests that a few improvements still need to be accomplished in this context. Limitations such as low and inconsistent connections, low storage
Gruenwald, L., & Banik, S. M. (2001, September). A poweraware technique to manage real-time database transactions in mobile ad-hoc networks. Proceedings of the 4th International Workshop on Mobility in Database and Distributed
952
Transaction Management in Mobile Databases
Systems, part of the International Conference on Database and Expert systems Applications (DEXA).
Skeen, D. (1985). Determining the last process to fail. ACM Transactions on Computer Systems, 3(1).
Kumar, V., Prabhu, N., Dunham, M., & Seydim, Y.A. (2002). TCOTA timeout-based mobile transaction commitment protocol. IEEE Transactions on Computers, 51(October).
Unland, R., & Schlageter, G. (1992). A transaction manager development facility for non-standard database systems. In A. K. Elmagarmid (Ed.), Database transaction models for advanced applications (pp. 400-466). San Francisco: Morgan Kaufmann.
Pitoura, E., & Bhargava, B. (1995). Maintaining consistency of data in mobile distributed environments. Proceedings of the IEEE Workshop on Mobile Systems and Applications. Reihar, P., Heidemann, J. S., Ratner, D., Skinner, G., & Popek, G. J. (1994, June). Resolving file conflicts in the Ficus file system. Proceedings of the USENIX Summer Conference (pp. 183-195).
Walborn, G., & Chrysanthis, P. (1996). Transaction processing in promotion. Proceedings of the ACM Symposium on Applied Computing.
953
T
954 Category: Mobile Software Engineering
Ubiquitous and Pervasive Application Design M. Bakhouya The George Washington University, Washington DC, USA J. Gaber Université de Technologie de Belfort-Montbéliard, France
INTRODUCTION The recent evolution of network connectivity from wired connection to wireless to mobile access together with their crossing has engendered their widespread use with new network-computing challenges. More precisely, network infrastructures are not only continuously growing, but their usage is also changing and they are now considered to be the foundation of other new technologies. A related research area concerns ubiquitous and pervasive computing systems and their applications. The design and development of ubiquitous and pervasive applications require new operational models that will permit an efficient use of resources and services, and a reduction of the need for the administration effort typical in client-server networks (Gaber, 2000, 2006). More precisely, in ubiquitous and pervasive computing, to be able to develop and implement applications, new ways and techniques for resource and service discovery and composition need to be developed. Service discovery is the process of locating which services are available to take part in a service composition. The service composition process so far concentrates on combining different available existing services as a result of the service discovery process. Most research to date in service discovery and composition is based on the traditional client/server interaction paradigm (CSP). This paradigm is impracticable in ubiquitous and pervasive environments and does not meet their related needs and requirements. Gaber (2000, 2006) has proposed two alternative paradigms to the traditional client/server interaction paradigm to design and implement ubiquitous and pervasive computing applications: the adaptive services/client paradigm (SCP) and the spontaneous service emergence paradigm (SEP). Bio-inspired approaches are adequate to carry out these new paradigms for designing and implementing ubiquitous and pervasive applications (Gaber, 2000). Indeed, the adaptive servers/client paradigm, considered as the opposite of CSP, could be implemented via a self-adaptive and reactive middleware inspired by a biological system like the natural immune system. The service emergence paradigm could also be implemented by a natural system that involves selforganizing and emergence behaviors (Gaber, 2000). Recently, agent-based approaches, with self-adapting and self-organizing capabilities, have been proposed in Bakhouya
(2005) and Bakhouya and Gaber (2001, 2006a, 2006b) to implement SCP and SEP respectively. More precisely, these approaches, inspired by the human immune system, provide scalable and adaptive service discovery and composition systems for ubiquitous and pervasive environments.
UBIQUITOUS COMPUTING In ubiquitous computing (UC), the objective is to provide users the ability to access services and resources all the time and irrespective to their locations (Weiser, 1993). Service discovery and access systems can be classified into three categories as depicted in Figure 1: structured systems, unstructured systems, and self-organized systems. Structured systems can be classified also in indexation-based architectures and hashing-based architectures. In indexationbased architectures, there are two categories: centralized and decentralized systems. In centralized indexation-based systems, typical resource discovery architectures (Bettstetter & Renner, 2000), such as Jini (2001), consist of three entities: service providers that create and publish services, a broker that maintains a repository of published services to support their discovery, and services requesters that search the service broker’s repository. Centralized approaches scale poorly and have a single point of failure. To overcome the scalability problem, decentralized approaches, such as m-SLP (Zhao, Schulzrinne, & Guttman, 2000) or Secure Service Discovery Service (Xu, Nahrstedt, & Wichadakul, 2001), traditionally have a hierarchical architecture consisting of multiple repositories that synchronize periodically. In hashing-based architectures (Wang & Li, 2003), proposed primary to file-sharing, distributed hash tables (DHTs) are used to assign files to specific nodes. This technique allows the implementation of direct search algorithm to efficiently locate files. However, hashing-based architectures require overlay networks between nodes that are generally hard to maintain. In unstructured systems, the most typical localization mechanisms are flooding and random walk. There are two main flooding techniques: the push and the pull technique. In the first technique, the server advertises periodically its services across the network. The clients receive the service
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Ubiquitous and Pervasive Application Design
Figure 1. Classification of service discovery systems according to their architectures and their operating modes
U
Service discovery systems
Structured systems
Self-organized systems
Affinity networks
Indexation
Centralized
Unstructured systems
Distributed hash tables
Decentralized
advertisement and cache the information. This information must have a time period associated with it, and must be flushed out from the cache when this time period expires. Hence, the user has a complete knowledge of the available services, and no request resolution process is required. In the pull technique, the client has no knowledge of services present in the network. In this case, a service request is broadcast to all neighbors within a certain radius with a TTL (time to live) tag (Wang & Li, 2003). A random walk is a stochastic process that evolves in the following manner (Gaber & Bakhouya, 2006b). A client sends its query message (i.e., a walker) to a randomly chosen neighbor. At each step, the query message is forwarded to a neighbor of its current location, and the process continues this way by taking random steps that are independent of all the previous ones until meeting the required service. Consequently, the random walk technique avoids message duplication inherent to the flooding mechanism (Wang & Li, 2003). More precisely, by using one walker, it cuts down the message overhead significantly. Nevertheless, the delay for a successful request resolution could be high. To decrease this delay, a requester could send k parallel query messages, and each query message takes its own random walk. However, it is difficult to determine a priori a suitable value for k. In other words, if this number k is big enough, the message traffic could increase considerably. An alternative approach to avoid this problem uses both random walks together with an adaptive cloning agent-based technique for service discovery (Gaber & Bakhouya, 2006b).
Flooding
Push
Pull
Random walk
Parallel random walk
Agent cloning
It should be noted that the fundamental aspect of these systems is the process of service discovery based on the traditional client to server paradigm. More precisely, it is the user who should initiates a request, should know a priori that the required service exists, and should be able to provide the location of a server holding that service. This is why the use of repositories is essential in these discovery systems. However, ubiquitous environments have the potential ability to integrate a continuously increasing number of services and resources that can be nomadic mobiles and partially connected. A user can be mobile or partially connected, and its ability to use and access services will no longer be limited to those that she/he currently has at hand or those statically located on a set of hosts known a priori. Therefore, the ability to maintain, allocate, and access a variety of continuously increasing numbers of heterogeneous resources and services distributed over a mixed network (i.e., wired, wireless, and mobile network) is difficult to achieve with the traditional client/server approaches (Gaber, 2000, 2006). More precisely, these architectures cannot meet the requirements of scalability and adaptability simultaneously. The way in which they have typically been constructed is often very inflexible due to the risk of bottlenecks, the difficulty of repositories updating, or the network loading problem. This is particularly true for the cases where some services could be disconnected from the network and new ones may join it at any time. An appropriate model was proposed originally by Gaber in Gaber (2000) as an alternative to the traditional client/ 955
Ubiquitous and Pervasive Application Design
server paradigm. This model can be viewed as opposed to the client/server model and is denoted adaptive servers/client paradigm. In this model, it is the service that comes to the user. In other words, in this paradigm, a decentralized and self-organizing middleware should be able to provide services to users according to their availability and the network status. As pointed out in Gaber (2000), such a middleware can be inspired from biological systems like the natural immune system. The immune system has a set of organizing principles such as scalability, adaptability, and availability that are useful for developing a distributed networking model in a highly dynamic and instable setting. In Gaber (2000, 2006), the immune-based approach operates as follows: unlike the classical client/server approach, each user request is considered as an attack launched against the global network. The immune networking middleware reacts like an immune system against pathogens that have entered the body. It detects the infection (i.e., user request) and delivers a response to eliminate it (i.e., satisfy the user request). Recently, an immune approach using mobile agents with cloning capabilities was proposed in Bakhouya (2005) and Bakhouya and Gaber (2006a, 2006b) to implement SCP. A mobile agent is a software program that may move from location to location to meet other agents or to access resources provided at each location. Using a mobile agent that can clone itself in order to increase system robustness and performance is an attractive idea. The clone operation creates multiple instances of an agent that runs on different machines. For example, an initial mobile agent starts on the requesting machine and, after a local step, creates replications (i.e., clones) that initiate parallel walks to further machines. This would allow agents to cover a much wider area of machine space in a reasonable amount of time (Gaber & Bakhouya, 2006b). However, it should be noted that increasing agent population with cloning operation will increase resource demands in the network, which would indirectly affect network performance. Since mobile agents operate in a dynamic and distributed environment, it is difficult, even impossible, to estimate a priori an appropriate number of agents in the network (Bakhouya, 2005). Also, changing the population dynamically in response to its environment is a complex issue in the absence of a central controller. A distributed approach for the regulation of mobile agent population and inspired by the immune system is proposed in Bakhouya (2005) and Bakhouya and Gaber (2002, 2006b). The immune system consists primarily of lymphocytes that circulate through the body in the blood and lymph system. There are two categories of lymphocytes, the B-cells and T-cells. The B-cells are developed in the bone marrow and the T-cells are developed in the thymus. The principle function of T-cells is to potentate the immune response by the secretion of specialized factors that activate other cells to fight off infection. The major function of the B-cell is the production of antibodies in response to foreign antigens. 956
According to Jerne, B-cells are interconnected by affinity relationship against foreign antigens and form idiotypic networks (Jerne, 1974). The mapping between the immune system entities and the middleware agents is done in the following manner. T-cells represent servers while B-cells represent services and resources. Antigens correspond to client requests while antibodies correspond to delivered responses. Servers are organized into communities by the creation of affinity relationships in order to represent services in the network. The establishment of relationship affinities between servers allows the solving, by collaboration, of user requests. A reinforcement learning mechanism is used to adjust and reinforce dynamical relationship affinity values according to delivered responses (Bakhouya, 2005). This reinforcement mechanism permits coping with dynamic changes in the network, the services availability, and the user requests. Similar to the natural immune system, new communities may be created or modified according to a dynamically changing environment. In other words, servers may acquire new or drop current servers through establishing or deleting the affinity relationship.
PERVASIVE COMPUTING Pervasive computing (PC), often considered the same as ubiquitous computing in the literature, is a related concept that can be distinguished from ubiquitous computing in terms of environment conditions. We can consider that the aim in UC is to provide any mobile device access to available services in an existing network all the time and everywhere, while the main objective in PC is to provide spontaneous services created on the fly by mobiles that interact by ad hoc connections (Gaber, 2000, 2006). Service composition systems can be classified into three categories as depicted in Figure 2: proactive composition systems, reactive composition systems, and spontaneous emergent service systems. The first category refers to off-line composition of available services to form new ones. Services that may be used for proactive service composition can be considered stable, widely and always available (Chakraborty & Joshi, 2001). The second category refers to the process of creating a composite service on demand. In other words, a composite service is created only when a user requests the execution of that service. Most known reactive and proactive service composition systems, such as the eFlow system (Casati, Ilnicki, Jin, Krishnamoorthy, & Shan, 2000), are based on a centralized broker which manages the service composition process (Chakraborty & Joshi, 2001). The drawback is that if a huge number of users attempt to access a variety and increasing number of services distributed over the network, the broker quickly becomes a bottleneck. It should be noted also that these systems are based on the client/server paradigm; it is the user who should
Ubiquitous and Pervasive Application Design
Figure 2. Classification of service composition systems according to their architectures and their operating modes
U
Service Service composition composition systems
Spontaneous emergent service
Propitient multi-agent systems (e.g. affinity networks)
Proactive/Reactive Composition
Ontologies-based methods
initiate a request, and moreover, services and future demands are known in advance. Gaber (2000, 2006) has proposed a second alternative paradigm to the client/server one for service composition that suits pervasive environments. This paradigm involves the concept of spontaneous emergence and is called the spontaneous service emergence paradigm. This paradigm can also be carried out by an inspired natural immune middleware that allows the emergence of ad hoc services on the fly according to dynamically changing context environments such as computing context and user context (Gaber, 2000). More precisely, in this model, ad hoc or composite services are represented by an organization or group of autonomous agents. Agents correspond to the immune system B-cells. Agents establish relationships based on affinities to form groups or communities of agents in order to provide composite services. A community of agents corresponds to the idiotypic network in the human immune system (Gaber, 2000). More generally, agents together with their affinity relationships as a whole form a propitient multi-agent system (Bakhouya & Gaber, 2006c). A propitient system is a system with the ability to self-organize in order to adapt towards the most appropriate agent organization structures according to unpredictable changes in the environment. This emergent behavior is delivered as a result of agent-to-agent and agentto-environment interactions that adapt until the system hits a most suitable affinity network. Therefore, a propitient multi-agent system implements the SEP. A self-organizing approach assumes that individual agents are autonomous agents, while multi-agent organizations are emerged structures that are not represented explicitly, but they exist through the affinity relationships between agents. In other words, agents cooperate equally rather than being assigned subordinate and supervisory relationships. It is worth noting that this multi-organization based on dynamic affinities supported by relationships provides a highly decentralized system while remaining adaptive in dynamic and open environments. More precisely, this decentralized organizational structure offers a high degree of resilience against an agent
Learning-based approaches
Workflow-based approaches
leaving the organization. For example, when an agent leaves an organization, all the peer affinity relationships with other agents are removed without additional messages since it does not rely on any overlay control structure. An affinity-driven clustering learning mechanism could be used to adjust the affinity relationships between nodes to cope with the user context and provoke or produce an emergent service (Gaber, 2000; Gaber & Bakhouya, 2006a). More precisely, an ad hoc emergent service is created spontaneously on the fly for a user or between a group of users in an unpredictable manner (i.e., without a priori intention).
CONCLUSION The design and development of ubiquitous and pervasive applications require alternative operational models to the traditional client/server paradigm. The adaptive servers/client paradigm and spontaneous service emergence paradigm are more adequate to ubiquitous and pervasive computing respectively. Service discovery and composition systems based on these three paradigms and proposed in the literature are presented with emphasis on elf-organizing and self-adapting approaches inspired by the immune system to implement SCP and SEP. Self-adaptation and self-organization are crucial issues in systems that operate in an open and dynamic environment.
REFERENCES Bakhouya, M. (2005). Self-adaptive approach based on mobile agent and inspired by human immune system for service discovery in large scale networks. PhD thesis NO. 34, Université de Technologie de Belfort-Montbéliard, France. Bakhouya, M., & Gaber, J. (2002) Distributed autoregulation approach of a mobile agent population in a network (Research Report RR-12-02, pp. 1-14). Université de Technologies de Belfort-Montbéliard, France. 957
Ubiquitous and Pervasive Application Design
Bakhouya, M., & Gaber, J. (2006a). Adaptive approaches for ubiquitous computing. Mobile networks and wireless sensor networks (pp. 129-163). Hermes Science. Bakhouya, M., & Gaber, J. (2006b). Adaptive approach for the regulation of a mobile agent population in a distributed network. Proceedings of the 5th International Symposium on Parallel and Distributed Computing (ISPDC’06) (pp. 360-366). Timisoara, Romania: IEEE Press.
Gaber, J., & Bakhouya, M. (2006b). Mobile agent-based approach for resource discovery in peer-to-peer networks. Proceedings of the 5th International Workshop on Agents and Peer-to-Peer Computing (at AAMAS) (pp. 1-9). Hakodate, Japan. Hofmeyr, S. A., & Forrest, S. (2000). Architecture for an artificial immune system. Evolutionary Computation, 8(4), 443-473.
Bakhouya, M., & Gaber, J. (2006c). Self-organizing approach for emergent multi-agent structures. Proceedings of the Workshop on Complexity Through Development and Self-Organizing Representations (GECCO’06) (pp. 1-5). Seattle, WA: ACM Press.
Jerne, N. (1974). Towards a network theory of the immune system. Annals of Immunology, 125, 125-373.
Bettstetter, C., & Renner, C. (2000). A comparison of service discovery protocols and implementation of the service location protocol. Proceedings of the 6th EUNICE Open European Summer School. Retrieved December 19, 2006, from http://www.bettstetter.com/publications/bettstetter2000-eunice-slp.pdf
Robert, M. (2000). Discovery and its discontents: Discovery protocols for ubiquitous computing. Research Report UIUCDCS-R-99-2132, Department of Computer Science, University of Illinois Urbana-Champaign, USA.
Casati, F., Ilnicki, S., Jin, L., Krishnamoorthy, V., & Shan, M. (2000). Adaptive and dynamic service composition in eFlow. Technical Report HPL-200039, Software Technology Laboratory, Palo Alto, CA. Chakraborty, D., & Joshi, A. (2001). Dynamic service composition: State-of-the-art and research directions. Technical Report TR-CS-01-19, Department of Computer Science and Electrical Engineering, University of Maryland, Baltimore County, USA. Retrieved from http://citeseer.ist.psu.edu/ chakraborty01dynamic.html Gaber, J. (2000). New paradigms for ubiquitous and pervasive computing (Research Report RR-09). Université de Technologies de Belfort-Montbéliard, France.
Jini. (2001). Jini technology core platform specification. Retrieved from http://www.sun.com/jini/specs
Wang, C., & Li, B. (2003). Peer-to-peer overlay networks: A survey. Technical Report, Department of Computer Science, HKUST. Retrieved from http://comp.uark.edu/cgwang/ Watanabe, Y., Ishiguro, A., & Uchkawa, Y. (1999). Decentralized behavior arbitration mechanism for autonomous mobile robot using immune system. Books artificial immune systems and their applications. Berlin: Springer-Verlag. Weiser, M. (1993). Hot topics: Ubiquitous computing. IEEE Computer. Xu, D., Nahrstedt, K., & Wichadakul, D. (2001). QoS-aware discovery of wide-area distributed services. Proceedings of the 1st IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGrid) (pp. 92-99). Brisbane, Australia.
Gaber, J. (2006). New paradigms for ubiquitous and pervasive applications. Proceedings of the 1st Workshop on Software Engineering Challenges for Ubiquitous Computing, Lancaster, UK.
Zhao, W., Schulzrinne, H., & Guttman, E. (2000). mSLPmesh-enhanced service location protocol. Proceedings of the International Conference on Computer Communications and Networks (ICCCN 2000) (pp. 504-509). Retrieved from draft-zhao-slp-da-interaction-07.txt
Gaber, J., & Bakhouya, M. (2001), A middleware for large scale networks inspired by the immune system (Research Report RR-11-01, pp. 1-6). Université de Technologies de Belfort-Montbéliard, France.
KEY TERMS
Gaber, J., & Bakhouya, M. (2006a). An affinity-driven clustering approach for service discovery and composition for pervasive computing. Proceedings of the IEEE International Conference on Pervasive Services (ICPS’06) (pp. 277-280). Lyon, France.
958
Adaptive Services/Client Interaction Paradigm (SCP): Adaptive Services to Client interaction paradigm is the opposed model to the traditional Client/Server interaction paradigm in which it is the most appropriate service that comes to the user in response to a request. This most suitable service can be computed by the network itself via an intelligent middleware.
Ubiquitous and Pervasive Application Design
Mobile Agents: A mobile agent is a software entity which may move with its own code and execution context from node to node to meet other agents or to access resources provided at each location. Pervasive Computing: The main objective of pervasive computing is to provide spontaneous emergent services created on the fly by mobiles that interact by ad hoc connections Propitient System: An agent-based system with the ability to self-organize in order to adapt towards the most appropriate organization that copes with the unpredictable changes in the environment. Reinforcement Learning: A mechanism used to adjust and reinforce dynamically relationship affinity values according to delivered responses and in order to cope with dynamic changes in the network, the service's availability and user requests. Service Composition: Service composition process concentrates on combining different available and selected
services via the service discovery process to deliver new ones. Service Discovery: Service discovery is the process of locating which services are available to take place in a service composition. Spontaneous Service Emergence Paradigm (SEP): Allows the emergence of ad hoc services on the fly for a user or group of users without a priori planning or intention Traditional Client/Server Interaction Paradigm (CSP): The traditional Client to Server paradigm is based on the following fundamental aspect: it is the user who should initiate a request, should know a priori that the required service exists and should be able to provide the location of a server holding that service. It should be noted that push, pull and P2P systems are still based on the Client/Server interaction paradigm. Ubiquitous Computing: the main objective of ubiquitous computing is to provide users with the ability to access services and resources all the time and irrespective to their location.
959
U
960 Category: Mobile Software Engineering
The “Umbrella” Distributed Hash Table Protocol for Content Distribution Athanasios-Dimitrios Sotiriou National Technical University of Athens, Greece Panagiotis Kalliaras National Technical University of Athens, Greece
INTRODUCTION During the past few decades, the Internet has blossomed due to the immerse growth of the telecommunication backbone, making it one of the key players in a wide area of fields. Even traditional players such as television or radio are now being challenged by the new entertainment media, the home computer. The increase of share communities such as Weblogs (Drezner & Farrell, 2004) or MySpace (www. myspace.com) and content-sharing software proves that people want to share their content with their global Web community. The need for such content-sharing software is therefore undisputable. Such attempts have been introduced in many ways during the past, with perhaps the most common example being Napster (www.napster.com) and Gnutella (http://gnutella.wego.com). The technical and ethical issues of these systems proved to be their weak point. Systems that have no central point of control and distribute functions among all users seem better fit for sharing and distributing content. A solution has been proposed in the form of distributed hash-tables (DHTs). This article proposes an alternative architecture for content distribution based on a new DHT routing scheme. The proposed architecture is well structured and self-organized in such a way as to be fault-tolerant and highly efficient. It provides users with content distribution and discovery capabilities on top of an overlay network. The novelty of our proposed architecture lies in its routing table which is maintained by each node and is of constant size, as opposed to other algorithms that are proportional to the network’s size (usually O(logN)). All operations in our architecture are of O(logbN) steps (entry, publishing, and lookups) and degrade gracefully as up-to-date information of the routing table decreases due to numerous node failures.
BACKGROUND The firsts to introduce routing algorithms that could be applied to DHT systems were Plaxton, Rajaraman, and Richa (1997). The algorithm was not developed for P2P systems,
and thus every node had a neighborhood of Ο(logN) and inquires resulted in Ο(logN) steps. It was based on the ground rule of comparing one byte at a time until all bytes of the identifier (or best compromise) were met. Our scheme meets the logarithmic growth of inquiries introduced by Plaxton et al. (1997), even though nodes are not placed within constant distance from each other. A variation of the Plaxton algorithm was developed by Tapestry (Zhao et al., 2004), properly adjusted for P2P systems (where overall state is not available). The algorithm once again tackles one digit at a time, and through a routing table of β*logβN neighbors routes to the appropriate node, resulting in a search of logβN maximum steps. Pastry (Rowstron, 2001) is similar to Tapestry, but added a leaf set of neighbors that the node first checks before referring to the routing table. Also a different neighbor set is maintained for tolerability issues. Each node maintains a neighborhood of log2bN rows with (2b-1) elements in each row and requires a maximum of O(log2bN) steps for enquires. Proper routing is maintained as long as (L/2) nodes are available in the neighborhood of each node. Once again, the variable size of each node’s table (which must be maintained up-to-date) limits the algorithm’s scalability. In Chord (Stoica, Karger, Kaashoek, & Balakrishnan, 2001) a different approach was applied, placing nodes in a circular space and maintaining information only for a number of successor and predecessor nodes through a finger table. Routing is established through forwarding queries to the correct successor based on the identifier. Even though the basic Chord mechanism only requires the knowledge of one successor, modifications where needed in order for the system to be applicable to a robust environment, introducing a finger table of O(logN) size. Finally, Kademlia (Maymounkov & Mazieres, 2002) bases nodes in a binary-tree through identifiers. Each node of the tree retains information concerning one node from each leaf, other than the one in which it resides. It also differentiates by applying an XOR comparison on identifiers instead of the casual comparison of each bit, adopted by all other algorithms.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
The “Umbrella” Distributed Hash Table Protocol for Content Distribution
SYSTEM ARCHITECTURE
Figure 1. The Umbrella architecture
In this section we will give an overview of the Umbrella architecture. The main functions consist of the insertion of nodes, the assignment of keys to corresponding nodes, and the routing mechanisms for three principle operations, namely the insertion of nodes, publication of content, and lookup of keys. Our architecture is based on an overlay network, and thus we assume that node connectivity is both symmetric and transitive.
U Level-3
Level-2 Level-1 Ring
Hashing Function The proposed architecture is based on the creation of an overlay network, where all inserting nodes are identified by a unique code, asserted by applying the SHA-1 (NIST, 1995) hash-function on the combination of IP and computer name, which returns a 160-bit identifier. This hash-function has been proven to distribute keys uniformly in the 160-bit space and thus provides the desired load balancing for both the user space and the content space, as the same function is applied to each content destined for distribution in the system.
Structure Overview The main objective of the architecture is to insert and retain nodes in a simple and well-structured manner, thus querying and fetching of content is both efficient and fault-tolerant. In addition, each node will need only to retain up-to-date information of a limited, constant number of neighboring nodes, allowing the system to escalate in population of both users and content. Each node is inserted into the system through an existing node, which announces the new entrance. When this procedure has ended successfully, the new node can, having acquired and informed all neighboring nodes, continue to publish all of its content. The publishing procedure is similar to the insertion mechanism, as content is characterized by a number of keys, which after being hashed can be forwarded in the same manner. All keys are published in an existing node whose identifier is the closest match to the key identifier. In a similar fashion, querying is performed by routing the request to the node with the identifier closest to the desired key. If no such node exists, it is assumed that the desired content is not available. The overlay network is constructed in the form of a loose B-Tree, where each node is placed in a hierarchy tree with a parent node and b child nodes, which in our initial architecture is of the value 16 ( in order to classify the 160-bit identifiers to a maximum B-Tree of height ≤ 40). All nodes are placed along the tree structure, without being required to fulfill pre-defined ranges as in a proper B-tree structure, and are responsible for updating their connections with
Ring Level-1 Level-2 Level-3
neighboring nodes that reside on either the parent, sibling, or child level. Along with obvious connections (parent, child, and sibling level of each node), further links to a limited number of nodes in the near vicinity are kept in record for fault-tolerant operations. Figure 1 illustrates the structure of this loose B-Tree. Routing in the umbrella protocol is simple and constitutes the forwarding of messages to either a parent or child node until the appropriate node is reached. In the rest of the article, with the term-appropriate node we will refer to either the exact or closest match alike.
Key Mapping Each level n of the structure is capable of withholding bn+1nodes. Each node has a unique parent node, which is always one level higher, and a maximum of b children at a lower level. The Umbrella overlay network is configured with the following simple rule. The relation between a parent node at level n and a child node (which must by default reside on level n+1) is defined as such and only such that: • •
The n+1 first (from left to right) digits of the parent’s identifier are equal to the corresponding digits of the child’s identifier. The n+2 digit of the child’s identifier determines the child’s position in the parent’s child list. Thus all children of the same parent share the first n+1 digits and all differ in the n+2 digit.
The above simple rule is obeyed by all nodes entering the Umbrella overlay network, with only the exception of the first node that actually initiates the network and is considered to be positioned at level -1. As already stated, the SHA-1 hash 961
The “Umbrella” Distributed Hash Table Protocol for Content Distribution
Table 1. Fields of the neighborhood table
logarithmic overlay steps to the total size of the network. This is stated and proved within the following two theorems: Theorem 2. Given an Umbrella network of N nodes with identifiers of base b acquired by a consistent hash function, the maximum height of the loose B-tree structure is of logarithmic scale.
function is used to assign identifiers to both nodes and content, offering a uniform distribution in the 160-bit space along with non-voluntary placement anonymity (Milojicic et al., 2002) of the published content. In order to apply the routing algorithms, we define a comparing function for identifiers as comp, which compares two identifiers and calculates their difference as a long integer, with importance given to digits from left to right. The consistent hash function balances key distribution among nodes, as stated in Karger et al. (1997) in the form of the following theorem: Theorem 1. Given a set of nodes N and keys K, then with high probability each node is responsible for an average of K/N keys, with a maximum of (1+α)K/N, whereas α is a parameter with bound of O(logN).
Routing Table As in most DHT systems, a routing table is maintained by each node in order to route incoming messages. Each node is responsible for keeping the table up-to-date by issuing messages to all nodes in its table at different intervals. The routing table in our architecture consists of three different setsa basic, an upper, and a lower set. The basic set stores nodes and information needed for basic routing operations under fault-free conditions. The upper and lower sets store additional indexes to nodes in the upper and lower levels, correspondingly, which are utilized when nodes in the basic set become unreachable. These three sets constitute the node’s neighborhood table and are presented in Table 1. The above elements are sufficient to maintain proper routing in our architecture even in the case of sudden failure of nodes. The upper set allows routing to nodes of higher level (when the parent node is unreachable) and the lower set to nodes of lower level (when child nodes fail). Each node is responsible to modify or fix its routing table when nodes enter/leave the network or a failure to communicate with another node is detected, respectively. Our architecture’s structure and routing table described so far ensure that a published key can be located by an appropriate query within 962
Proof: Let b denote the base of our identifiers, N the total number of nodes, and k a particular level in the Umbrella structure. Then according to the Umbrella protocol, in each level a maximum of bk nodes can reside, with b0=1 as stated for the first node that creates the network. Thus, if m denotes the number of levels required for the above population of nodes, we acquire the following relation, in respect to Theorem 1 that provides, with high probability, a uniform distribution of identifiers to our space: b 0 − b m +1 1 − b m +1 = ⇔ 1− b 1− b k =0 m =[N(b–1) {log b [N+(b −1 ) + 1]} − 1 ⇔ m =⇔ {log 1]} –1 N =
m
∑b
k
= b
Thus the maximum height m of our structure is of O(logbN). Theorem 3. A successful lookup in an Umbrella network requires, with high probability, O(logbN) steps. Proof: Suppose that a node p that resides at level lp is seeking a specific key k that resides within our network in another node f at level lf. If m denotes the number of levels of the current network, N the nodes, and b the base of identifiers, then we could argue that the worst-case scenario would require both nodes to reside at level m and with maximum distance between them (thus node p is an m-depth child of the first child at level 0 and on-forth, and node f is the m-depth child of the b child at level 0 and on-forth). In this case, the lookup must first ascend all the way to the top of our structure (thus m steps) and then descend to the bottom (m steps again). In total, a maximum of 2m steps are required. Hence, from Theorem 2, the required maximum steps for a successful lookup is, with high probability, of O(logbN) steps.
ALGORITHMS AND IMPROVEMENTS Main Algorithms During the creation of the overlay network, the first node to enter creates the new network by placing itself on the top of the system. As new nodes arrive, they are placed according to their identifier, as described in the previous section. A node only needs to contact an existing node in the system
The “Umbrella” Distributed Hash Table Protocol for Content Distribution
in order to be inserted (special mechanisms for fetching existing nodes by outside contacts are not of the scope of this article, as our architecture can embody any of numerous such techniques already proposed (Francis, 1999)). Only the first node is automatically inserted regardless of its identifier; all subsequent nodes are placed within the system based on the insertion algorithm. The insertion mechanism is quite simple, intentionally, as with all of the system’s mechanisms, and consists of the following steps: • • • • •
Contact an already connected node and issue a request for insertion. The established node checks if the n+1 first digits of the identifier match its own, where n is the level the node resides. If not, the insertion message is forwarded to the node’s parent. If yes, the message is forwarded to the child with the n+2 digit common with that of the new node. If such a child does not exist, then the new node is placed as a child to the current node and is informed of its new neighbors and via versa.
The publish procedure is similar to insertion and is therefore suppressed. Conversely, the search mechanism is executed as follows: • • • • • •
The node first checks for the keyword in its list of published keywords. If it exists then the search terminates. If not, then it checks whether the first n+1 digits are identical to its own identifier. If not, the message is forwarded to its parent. If yes, then it is forwarded to the child with the corresponding n+2 digit matching. If no such child exists, then the search fails.
The pseudo code of the above algorithm is given in Figure 2. The final mechanism provided by our protocol is that of node departure from the system. When a node issues a departure, the following steps are followed: • •
•
If the node has no children, then all of its keywords are forwarded to its parent and it informs all its neighbors of its departure. If it has any child, then it randomly picks one and copies all of its neighborhood and keyword information to it before departing. The chosen child moves up a level and substitutes the departing node. If the chosen child has any child, then the previous step is repeated recursively until a node with no children is reached and the first step is then executed ending the algorithm.
Figure 2. Pseudo code for search mechanism
U
Enhanced Algorithms The algorithms presented in the previous section embody the main mechanisms of our architecture and are capable of maintaining the system stable and fully functional under normal conditions, as will be validated by our simulation results in the following article. The system is however liable to node departures, either intentional or due to network disconnections. In the next set of mechanisms, we will concentrate on sudden departures of nodes, which we will call “node failures.” These are due to either voluntary departure without calling the appropriate mechanism, sudden departures due to client errors, or nodes becoming unavailable due to network disconnections. We will treat all of the above cases in the same manner, and through changes in the algorithms already presented, we will allow the system to bypass node failures. Most changes are based on using the upper and lower set of our neighborhood table to bypass nodes that are not responding. The upper set is utilized to forward messages to nodes of a higher level, while the lower set is for nodes on a lower level. In the first case, when a node is unable to contact its parent node, it attempts to forward requests consequently to: • • •
the parent’s parent node (field Up2 on the upper set), the node to the right of the parent node (field Right2 on the upper set), and the node to the left of the parent node (field Left2 on the upper set).
Whichever of the above succeeds first will terminate the mechanism. The only exception to the above algorithm is that of the node on level -1, which triggers a variation since it has no parent of sibling nodes, and attempts are made toward nodes on either the left or the right of the child node. In the latter case of a child node failure, the corresponding nodes are contacted in the following order:
963
The “Umbrella” Distributed Hash Table Protocol for Content Distribution
Table 2. Corresponding forwarding of repair messages in order to reach a parent node Failing Node Up Node2 Left3 or Right3 Left2 or Right2 Left or Right Up2
• • • • •
Action Taken Contact Up2 Contact Node Contact Left or Right respectively Contact Up2 Contact Up Contact Up
one of the child’s child (field Umbrella2 on the lower set), the node on the right of the child (field Umbrella on the basic set), the node on the left of the child (field Umbrella on the basic set), a child of the node right of the issuing node (field Right3), and a child of the node left of the issuing node (field Left3).
Repair Mechanism In order to address the problem of node failures even further, we have designed a repair mechanism, which is invoked whenever such a failure is detected. The algorithm utilizes the delete algorithm presented in our main mechanisms section in order to repair a failure to a child node. It can be proven that all other failures can be transformed into a child failure through contacting nodes in the neighboring table and forwarding a repair message to higher or lower levels until the parent node of the failing node is reached. More details of the actions involved in order to resolve the appropriate node are given in Table 2. Once the appropriate node is reached and informed of the child failure, a variation of the delete algorithm is evoked in order to repair the failure by substituting the failed node with one of its children or by deleting it if none is available. Each node is responsible for checking its neighborhood table periodically by issuing ping messages to all node entries and invoking the repair mechanism whenever a failure is detected.
SYSTEM EXTENSIONS Having presented the core structure and logic behind our routing protocol, we will continue with a number of extensions that improve the system’s performance. The first extension introduces the use of replication schemas, which have been shown to increase the robustness of content distribution systems (Ghodsi, Alima, & Haridi,, 2005). In this article we have implemented three additional replication schemas. 964
Table 3. Singular identifier assignment functions Name
Details
Instance
OI
This is the base fuction
a b c d e fg h
II
This function inverses the identifier
hgfedcba
IP
The identifier’s digits are inversed by pair
badcfehg
IPW
All digits are inversed by pair as in the case of IP and the result is inversed as a whole as in II
ghefcdab
IH
The II function is applied to the first and second half of the indentifier indenpendently
dcbahgfe
SH
The first and second halves are switched without being inverted
efghabcd
RR
A random reordering of the identifier’s digits
daecghbf
SRR
Same as the case of RR but with different random generator
cfbahgde
Our core routing protocol publishes a keyword in a single node, the one with the closest identifier to that of the keyword. All three replications schemas retain this quality and enhance it by also publishing the keyword to a number of additional nodes, from which one can recall a successful lookup. Our three variations follow: 1. 2. 3.
Local Spread Replication (LSR): The keyword is also published in all nodes residing in its neighboring table. Inverse Replication (IR): This mechanism publishes keywords to the closest match and to the inverse closest match. Local Spread Inverse Replication (LSIR): It implements a local spread in both the closest and the inverse closest match.
The second extension implemented allows nodes to participate in a number of virtual networks, with a different identifier in each one. This allows each node to have a different set of neighbors and thus increase its tolerability substantially. In order to achieve this, we have defined a number of singular identifier assignment functions that transform the original identifiers into a new set of identifiers. This new set is then used to allocate nodes and route requests in the virtual networks. We have defined seven different such functions, which are given in Table 3.
CONCLUSION Through the course of this article, we presented the Umbrella protocol, a novel protocol based on a distributed hash table that supports key publishing and retrieval on top of an overlay network for content distribution. We have analyzed our protocol and its algorithms through theoretical means, and provided a number of algorithms and extensions. Its main novelty lies in its fixed-size routing table sustained by each node, which is able to provide efficient routing even under contrary conditions. The protocol is also highly scalable due to its low traffic load demands.
The “Umbrella” Distributed Hash Table Protocol for Content Distribution
REFERENCES Drezner, D., & Farrell, H. (2004). Web of influence. Foreign Policy Magazine. Francis, P. (1999). Yoid: Extending the multicast Internet architecture. White Paper. Retrieved from http://www.aciri. org/yoid Ghodsi, A., Alima, L.O., & Haridi, S. (2005). Symmetric replication for structured peer-to-peer systems. Proceedings of the 3rd International Workshop on Databases, Information Systems and Peer-to-Peer Computing, Trondheim, Norway. Karger, D., Lehman, E., Leighton, F., Levine, M., Lewin, D., & Panigrahy, R. (1997). Consistent hashing and random trees: Distributed caching protocols for relieving hot spots on the World Wide Web. Proceedings of the 29th Annual ACM Symposium on Theory of Computing, El Paso, TX. Maymounkov, P., & Mazieres, D. (2002). Kademlia: A peer-to-peer informatic system based on the XOR metric. Proceedings of IPTPS’02, Cambridge, MA. Milojicic, D., Kalogeraki, V., Lukose, R., Nagaraja, K., Pruyne, J., Richard, B., et al. (2002). Peer-to-peer computing. HPL-2002-57R1, HP Labs Technical Report. NIST (National Institute of Standards and Technology). (1995). FIPS Pub 180-1: Secure Hash Standard (SHA-1). Federal Information Processing Standards Publication. Plaxton, G., Rajaraman, R., & Richa, A. W. (1997). Accessing nearby copies of replicated objects in a distributed environment. Proceedings of the 9th Annual ACM Symposium on Parallel Algorithms and Architectures (SPAA).
Stoica, M. R., Karger, D., Kaashoek, F., & Balakrishnan, H. (2001). Chord: A peer-to-peer lookup service for Internet applications. Proceedings of SIGCOMM. Zhao, B. Y., Huang, L., Stribling, J., Rhea, S. C., Joseph, A. D., & Kubiatowicz, J. D. (2004). Tapestry: A resilient global-scale overlay for service deployment. IEEE Journal on Selected Areas in Communications.
KEY TERMS DHT: Distributed hash table. Hashing: Producing hash values for accessing data or for security. A hash value (or simply hash), also called a message digest, is a number generated from a string of text. The hash is substantially smaller than the text itself, and is generated by a formula in such a way that it is extremely unlikely that some other text will produce the same hash value. Overlay Network: A network built on top of one or more existing networks. This network adds an additional layer of indirection/virtualization and also changes properties in one or more areas of underlying network. P2P: Peer-to-peer. Replication: A duplicate copy of similar data on the same or a different platform. Routing: The process of moving a packet of data from source to destination. Routing in P2P networks refers to finding a path to the desired node. SHA-1: Secure hash standard.
Rowstron, D.P. (2001). Pastry: Scalable, decentralized object location, and routing for large-scale peer-to-peer systems. Proceedings of Middleware 2001.
965
U
966 Category: Mobile Software Engineering
Understanding Multi-Layer Mobility Sasu Tarkoma Helsinki Institute for Information Technology, Finland Jouni Korhonen TeliaSonera Corporation, Finland
INTRODUCTION Mobility is an important requirement for many application domains, where entities change their physical or logical location. Physical location denotes the real-world location of a device, whereas logical location is not necessarily dependent on the physical environment. Mobility support may be divided into several technical layers and also categories depending on the nature of mobility. In this article, we consider mobility protocols starting from the network layer (layer 3 in the OSI stack) and ending at the application layer (layer 7), and focus on physical mobility. The most fundamental network-level protocols for supporting mobile hosts are the Mobile-IP protocols standardized by the IETF (Perkins, 2002; Johnson, Perkins, & Arkko, 2004). Another related network-level solution is network mobility (NEMO) (Devarapalli et al., 2004), in which complete sub-networks may change location as well as single hosts. Mobility can also be handled on the transport layer. Transport layer seamless handover (TraSH) (Fu et al., 2004), datagram congestion control protocol (DCCP) (Kohler, 2006), and mSCTP (Xing, Karl, Wolisz, & Müller, 2002) are recent examples of such solutions. Yet another way of managing host mobility is with mobility-aware virtual private networks (VPNs) such as MOBIKE-based IPSec VPNs (Kivinen, 2006). Protocols such as wireless CORBA (WCORBA) (OMG, 2004) and the session initiation protocol (SIP) (Schulzrinne & Wedlund, 2000) provide more finegrained mobility than host based, and they do not assume underlying transport- or network-level mobility support. Middleware support for mobility is required in order to provide location transparency for objects, agents, and other components; support efficient and reliable communication in wireless environments; and buffer messages and other data for disconnected operation. In addition, the middleware may support scalability and availability of resources and services. Mobility is inherently tied with the way nodes are addressed in a distributed network. In this article, we examine three different ways to address mobile nodes and components: addresses with location and identity, locator/identity split, and content-based addressing. The first addressing model is used by the IP protocol. The second model is an extension of
the first and used, for example, in the host identity protocol (HIP) and the i3 overlay (Stoica, Adkins, Zhuang, Shenker, & Surana, 2002). The third model has been proposed for expressive communication in ubiquitous environments. The aim of this article is to examine the addressing models and investigate cross-layer interactions of different mobility protocols. One of the interesting questions is how mobility should be handled and coordinated when there are multiple layers offering support for mobility. We also consider the case of the hop-by-hop routed layer-7 environment, implemented typically using SOAP (W3C, 2003), CORBA, or SIP in the telecommunications sector. These three technologies are the most frequently used, have differing characteristics and product bases, and contain the essence of middleware/application layer communication. SOAP is an abstract and generic messaging framework with extendable header system, allowing rich facilities for hop-by-hop propagation of messages.
ADDRESSING MODELS The way mobile and stationary nodes are addressed is crucial in how mobility is supported in a distributed system. We define three different addressing models for mobile systems (see Figure 1): 1.
2,
3.
Address with Both Location and Identity: This form of addressing couples the communicating end-points to specific locations in a network. For example the IP address is used in both identifying a node and routing packets to it. This form of addressing typically uses a mediating stationary node to handle the mobility management and location updates for the mobile nodes. Address with Locator/Identity Split: This way of addressing separates the identity of a node and the location of the node. This allows more flexible mobility support since the identity may be used to lookup the physical location of a node. For example the Internet Indirection Architecture (i3) and the HIP are based on this form of addressing. Content-Based Addressing: This goes beyond locator/identity split, because it decouples the destination
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Understanding Multi-Layer Mobility
Figure 1. Three addressing models
MOBILITY-ENABLING PROTOCOLS
U
Content-based address
A Taxonomy
Locator/identity split
Host mobility happens when a host relocates to a new location in the network, thereby possibly causing a change of the underlying IP address. Since IP addressing is tied to the location, this may cause a fundamental change in routing for the relocated host. This host relocation is commonly referred to as a handover. Handovers are usually divided into two main categories: horizontal handovers and vertical handovers. A horizontal handover is commonly understood as a handover that takes place within the same access network technology. A vertical handover is handover that takes place across different access network technologies (and from the host’s point of view, between different networking interfaces). There are also two ways of doing the handover: breakbefore-make or make-before-break. The difference of these two approaches is whether the mobility-enabling protocol or the terminal implementation (in hardware, point of view) allows creating connectivity to the new access network or router before leaving the old access network or router. The host may also have several active IP addresses, which is called multi-addressing. Multi-addressing may also be used to realize multi-homing, which generally means that the client is connected to two independent networks for increased reliability. Multi-homing is also needed when several different access network technologies are used simultaneously. Server-side resiliency is commonly realized by connecting services to multiple network providers. This is called site multi-homing. User mobility happens when a user changes the host device or access host, which causes a change in the underlying physical address of the user. The device characteristics may also change, for example when the user changes from a PDA (personal digital assistant) to a laptop. An important subcategory of user mobility is session mobility, which allows the relocation of user sessions from one host to another. Session mobility is an important requirement for current and future mobile applications, in which instant messaging (IM), multimedia, and voice sessions, for example, are moved from one device to another. Service or application mobility happens when a service relocates or resides on a mobile host that moves. Service mobility may be triggered by factors not related with a user, for example load balancing.
IP address (location, identity)
from both identity and location. The destination is no longer defined by a single identity, such as the IP address or a cryptographic public key, but rather it is defined by logical rules set by applications running on the destination host. The rules are applied on messages or packets in order to make forwarding decisions. This means that using content-based addressing, we have decoupled many-to-many communication. On the other hand, the realization of content-based communication is more complex and costly. The cost of mobility in content-based routing is high when compared with the other forms of addressing. Research systems such as Siena (Carzaniga, Rosenblum, & Wolf, 2000) and Rebeca (Mühl, Ulbrich, Herrmann, & Weis, 2004) use content-based addressing. These addressing models are not orthogonal and may be applied on different layers of the communications stack. Since the current Internet is based on the IP protocol, it provides the baseline addressing with location and identity contained in the IP address. Above that, we may implement the locator/identity split using HIP or an overlay network such as i3. Content-based addressing is also implemented above IP using application-level routers. Identity-based mechanisms may be extended to support anonymous communication and multicast. For example, i3 supports multicast using triggers and anonymity by chaining private and public triggers. Content-based routing may, on the other hand, be extended to support identity-based communication by subscribing public keys, for example. The addressing models have differing notions of the addressing space, in which addresses are defined. These differences can be used to characterize the difference between identity-based addressing and content-based addressing. The identity vector (public key) is a point in the flat onedimensional addressing space of an overlay system. The content-based address, which is defined using a logical rule, is a subspace of a multi-dimensional addressing space. This illustrates the main difference, which is the expressiveness of the communication. In essence, for IP mobility there is a single, fixed indirection point; for locator/identity split there is a single indirection point; and with content-based there are multiple indirection points.
Network Layer Solutions The current solutions being standardized by IETF for network-layer mobility support are the Mobile IPv6 (MIP6) and Mobile IPv4 (MIP4) protocols. MIP is a layer-3 mobility protocol for supporting clients that roam between IP net967
Understanding Multi-Layer Mobility
works. Upper-layer protocols and applications are unaware of possible changes in network location and thus can operate uninterrupted while the host moves. MIP6 mobility support consists of the triangle of the home agent (HA), correspondent host (CH), and the mobile node (MN). MIP4 has an additional optional networking node called foreign agent (FA) that has been left out from the MIP6 specifications. In both MIP versions the HA serves as an anchor point for MNs, and any CN may communicate and initially reach the MNs through the HA. The basic MIP routing is triangular. A CN sends packets to an MN via an HA, and then the HA tunnels packets to MN’s current location. Finally the MN sends packets directly to the CN. In practice, triangular routing is inefficient and generally also impossible due to widely used ingress filtering. Practical MIP deployments either route all packets via HA (reverse tunneling) or the MN and CN communicate directly (route optimization, which can be negotiated with the help of HA). The distance between the MN and the HA may also be long both topologically and geographically. Thus routing packets between the MN and the HA may cause considerable delay. However, to improve the situation, an HA may also be allocated from the network (Calhoun, Johansson, Perkins, Hiller, & McCann, 2004) the MN is currently visiting. A similar way of optimizing IP mobility is utilizing some form of localized or hierarchical mobility management. The hierarchical mobile IPv6 mobility (HMIPv6) (Soliman, Castelluccia, Malki, & Bellier, 2005) management solution introduces local mobility anchor points (MAPs) that are essentially Home Agents. MAPs can be located at any level in a hierarchical network of routers, including the access routers. The aim of the HMIPv6 is to minimize the signaling latency and reduce the number of required signaling messages. As long as the MN stays inside one MAP domain, it only needs to update its location with the MAP. The localized mobility management can also be completely handled on the network side without MN’s involvement at the IP mobility protocol level. In these cases the network side needs to employ some kind of tunneling or local routing solution that is transparent to the MN. Another network-layer mobility solution being standardized by IETF is the NEMO. The technical solution of NEMO is close to that of MIP6. NEMO allows complete sub-networks to change their location in a network instead of single hosts. This is realized with a mobile router that manages the mobile network. Hosts behind the mobile router do not need to be aware of mobility in any way. All packets destined to hosts behind the mobile router get routed towards the virtual home network. Then an HA managing the virtual home network tunnels all packets to the mobile router managing the mobile network. One practical application of IPSec-based VPNs is to extend the user’s home network environment to be accessible from any location. In a tunneled mode, VPNs tunnel 968
all packets between the mobile host and a security gateway (that is usually located at the edge of user’s home network). Until recently IPSec VPNs have not survived the change of underlying IP addresses that are also used as the outer IP addresses of the VPN tunnel. Both IKE (Kaufman, 2004) and IPSec SAs (security associations) had to be rekeyed after IP addresses of either end of the tunnel changed. This practically caused all existing connections to drop. Recent developments in standardization have addressed this issue. For example, MOBIKE aims to support a way to update the IKE SA and IPsec SA endpoint addresses without rekeying the SAs. This would allow keeping the existing IKE and IPsec SAs in place even when the IP address changes.
Transport Layer Solutions The recent need for multi-homing support for transport protocols has made it possible to provide limited mobility support at the transport level. Examples of recent such transport protocols are DCCP with multi-homing and mobility extension to the base protocol (Kohler, 2006), TraSH, and mSCTP. The latter two are based on mobile SCTP, which is defined as SCTP with the ADDIP extension (Stewart et al., 2004). The basic idea of transport layer mobility is to maintain the end-to-end connectivity at the transport layer, and solve the mobility problem without additional infrastructure and functionality at the network layer. When a host’s point of attachment to the network changesthat is, the underlying IP address changesthe transport-layer mobility protocol needs to refresh the association between the MN and the CN using some protocol-dependent mechanism. This approach is very appealing because it requires no additional tunneling and does not interfere the natural routing of IP packets. It has also been shown that transport-layer protocols are capable of smooth handovers (Fu et al., 2004). The biggest downside of current transport-layer mobility solutions is the lack of proper mobility management. As long as everything is MN initiated, the proposed solutions work. If a CN needs to locate an MN or both communicating ends are mobile (so called double jump problem), current transport-layer mobility solutions most probably fail to work properly. Proposals to solve the mobility management are still open research issues. MIP suffers from a number of limitations, such as packet loss, high handover latency, packet encapsulation overhead, and conflict with network-level security solutions (Fu et al., 2004; Schulzrinne & Wedlund, 2000). MIP requires that location management resides on the HA. TraSH decouples location management from data traffic forwarding (Fu et al., 2004), and thus a lookup server, such as DNS, may be used for location management. TraSH leverages the multi-homing capabilities of SCTP. The difference between TraSH and MIP is that TraSH sends packets directly to the MN without using the HA.
Understanding Multi-Layer Mobility
Between Network- and Transport-Layer: HIP Above the network level, we have various requirements for mobility in the transport and application layers. Transportlevel mobility support needs to cope with changing subnets and prevent, for example, socket errors during mobility. HIP is located between the network and transport layers, and provides this kind of functionality by associating each socket to a public cryptographic key instead of an IP address. The fundamental idea behind HIP is to separate the address of a network-addressable node to two parts: the identity and locator parts. The identity part uniquely identifies the host using a cryptographic namespace, and the locator part uniquely defines the location of the node. The former part is assumed to be a long-living identifier. The latter is typically the IP address of the mobile node. Additional benefits of HIP are authentication and support for denial of service (DoS) attacks through cryptographic puzzles in the initiation phase of the protocol (Moskowitz, Nikander, Jokela, & Henderson, 2006).
Application Layer SIP mobility support is similar in nature to the HA mechanism used in MIP. SIP mobility support is based on the home registrar, which is a rendezvous point for information for a particular user. SIP mobility is simplest for pre-call mobility that only requires updating of the home registrar. In addition, SIP supports midcall mobility, which requires the mobile host to send an INVITE request with the new IP address to the correspondent host (Schulzrinne & Wedlund, 2000). SIP supports session mobility, in which a media session can be maintained while changing hosts. Moreover, the end point of an active session may be changed to another device. The Wireless CORBA specification was designed to provide a minimal useful functionality for CORBA applications. The specification defines extensions and protocols for applications in which clients and servers are executed on hosts that can move. The specification introduces the mobile IOR (interoperable object reference), which is a relocatable object reference that identifies the access bridge and the terminal on which the target object resides (OMG, 2004). An entity called the home location agent (HLA) keeps track of the access bridge to which the terminal is currently connected. The mobile IOR provides mobility transparency and contains either the home location agent’s address or the last known access bridge of the mobile host. In the former case the HLA will provide the new address of the mobile host. In the latter case, the last known access bridge provides the current address or forwards the invocation. Each terminal is identified using a unique terminal identifier.
The access bridges may support handoff and the specification defines two different cases of handoff: the backward and forward handoff. The former is the normal case and the latter is used to re-establish connectivity after a sudden disconnection. The backward handoff (or simply handoff) may be network initiated or terminal initiated, whereas the forward handoff is always terminal initiated. The Internet Indirection Infrastructure (i3) (Stoica et al., 2002) is an overlay network that aims to provide a more flexible communication model than the current IP addressing (Stoica et al., 2002). In i3 each packet is sent to an identifier. Packets are routed using the identifier to a single server in the distributed system. The server, an i3 node, maintains triggers that are installed by receivers that are associated with identifiers. When a matching trigger is found, the packet is forwarded to the associated receiver. An i3 identifier may be bound to a host, object, or a session, unlike the IP address, which is always bound to a specific host. The robust overlay architecture for mobility (ROAM) builds on i3 and allows end-hosts to control the placement of rendezvous points (indirection points) for efficient routing and handovers (Zhuang, Lai, Stoica, Katz, & Shenker, 2002). ROAM uses trigger server caching and trigger sampling, and supports fast handovers and multicast-based handovers for make-before-break. ROAM supports legacy applications using a user-level proxy that encapsulates IP packets within i3 packets and manages trigger-related operations. Application mobility may require special mobility protocols, for example for applications that use or participate in overlay networks. Content-based routing and publish/subscribe (pub/sub) networks (Eugster, Felber, Guerraoui, & Kermarrec, 2003; Tarkoma et al., 2003) are examples of this kind of behavior. These systems build large-scale multicast event distribution trees over point-to-point communication links. When an application providing or subscribing certain events moves, a part of the routing topology needs to be updated to reflect this change (Burcea, Jacobsen, de Lara, Muthusamy, & Petrovic, 2004). Pub/sub systems require their own mobility protocols in order to update the event routing topology and optimize event flow. In content-based routing of information, event brokers forward notifications based on a routing configuration established by advertisement and subscription messages. The main motivation for a pub/sub mobility protocol is the avoidance of triangle routing through a designated home broker, which may be inefficient. Experimental results shows that home-broker-based approaches do not perform well (Tarkoma, Kangasharju, & Raatikainen, 2003; Burcea et al., 2004). Mobility protocols are also needed for load balancing subscribers and advertisers between brokers. Efficient mobility protocols for pub/sub are currently an active research topic.
969
U
Understanding Multi-Layer Mobility
Table 1. Differences in mobility-aware systems and protocols I MIP
HIP
SIP
Target
MN
MN
Session
Mechanism
HA
DNS/Overlay
Home registrar
Buffering
No
No
Yes (stateful proxy)
Update Point
1 Fixed
1
1 Fixed
Location Privacy
Yes (not /w route opt)
No
No
Authentication
Yes
Yes
Yes
Table 2. Differences in mobility-aware systems and protocols II WCORBA
i3
Pub/Sub
Target
Object
Any
Subs/Advs
Mechanism
Home bridge
Rendezvous point
Hop-by-hop
Buffering
Yes
No
Yes
Update Point
1 Fixed
1
>1
Location Privacy
Yes
Yes
Yes
Authentication
-
-
-
DISCUSSION Tables 1 and 2 illustrate the differences between different mobility-aware systems and protocols discussed in this article. The target denotes the nature of the mobile entity. The target is the mobile node for MIP and HIP. Higher-level mobility protocols allow more fine-grained mobility, SIP supports session mobility, WCORBA object mobility, i3 leaves the entity unspecified, and mobile pub/sub systems support the mobility of subscriptions and advertisements to various degrees. The mechanism term denotes the type of handover protocol, for example the HA- based scheme of MIP or DNS/Overlay update of HIP. SIP uses the home registrar for location updates, WCORBA has the home bridge. The i3 overlay uses a rendezvous point that manages the triggers, and mobile pub/sub systems typically have to update the whole routing path between the source and destination of mobility. The update point denotes the number of indirection points and whether or not they are fixed. MIP, WCORBA, and SIP have a single, fixed indirection point: the HA, home bridge, or the home registrar. HIPwith overlay-based address resolutionand i3 have a single, non-fixed indirection point. Finally, pub/sub systems have typically multiple indirection points. Buffering of packets and messages is a useful functionality for supporting disconnected operation. MIP and HIP do 970
not support this feature. SIP supports disconnected operation through stateful proxies and buffering messages for delivery. SIP also copes with network partitions using retransmission (Schulzrinne & Wedlund, 2000). WCORBA access bridges maintain a list of pending invocations, and pub/sub systems buffer notifications for disconnected clients. The i3 overlay, on the other hand, does not buffer packets. Location privacy hides the current IP address of the mobile entity. MIP supports location privacy with the exception of the route-optimization option. HIP and SIP do not support location privacy. A HIP host has access to the IP address corresponding to the public key (host identity) of a mobile node. SIP does not hide the IP address; it is disclosed in the session description. WCORBA provides support for location privacy. Terminals are addressed using the terminal identifier, and the access bridge hides the transport address of the terminal. On the other hand, the location of an access bridge is revealed. The i3 overlay supports the hiding of source addresses with private triggers. Typically mobile pub/sub systems, such as Siena and Rebeca, provide anonymous communication, and only the edge brokers know the transport addresses. Authentication of terminals and users is also an important functionality for mobile systems. MIP provides authentication of signaling messages using various authorization extensions. IPSec and Internet Key Exchange (IKE) can be used to protect the integrity and authenticity of MIP signaling. On the other hand, IPSec and IKE were not initially
Understanding Multi-Layer Mobility
designed for multi-homed operation, and currently, multihomed operation has overhead due to additional database entries and key negotiations for each pair of source/destination address. HIP supports authentication through the host identity, which is essentially a public key. SIP supports three authentication styles: HTTP-style basic authentication, digest authentication, and S/MIME. WCORBA may be used with the CORBA security specification (OMG, 2002), and i3 and pub/sub systems may be extended to support various forms of authentication.
CROSS-LAYER INTERACTIONS Mobility support is currently available on many layers. Assuming that the base network-level routing technology is the Internet protocol (IP), mobility solutions are also needed on many layers. First, the lower layers need to be able to detect mobility and activate higher-level protocols. These mechanisms are outside the scope of this article. Second, after physical mobility the IP address will change, and after user mobility it may change; both instances require solutions for informing others that the address of the host has changed. This may be accomplished by pure network-level solutions (mobile IP), a hybrid approach (HIP), or through purely application-level solutions (overlay, DNS update). The mobility of sessions and objects is more subtle than the mobility of hosts, and we need mechanisms above L3 to cope with session and object mobility due to reconfiguration and, for example, load balancing decisions. Lower-level mobility protocols are in some cases complementary to higher-level protocols. Both MIP and HIP are complementary, as is the case for MIP and middleware mobility protocolsfor example, WCORBA and multi-hop application-level overlays such as i3 and Siena. If SIP or WCORBA provides mobility support, a higher-level mobility solution is not required. In general, a higher-level handover protocol is required for more fine-grained mobility support and optimizations. If only host mobility support is required, higher-level protocols are not needed. MIP may still be useful in scenarios where application-level mobility protocols are used and the client needs to be contacted using the home IP address. For example, pub/sub and many overlay systems hide the IP addresses of the communicating entities. In these systems the IP address of the terminal (care-off address) is not used for end-to-end communication, but only for the lasthop at the edge of the network. If MIP is not used, clients cannot address the mobile node directly since its IP address changes with the access location. If MIP or HIP are not used, higher-level protocols need to provide mobility support. In this scenario there are no interactions between network-level mobility protocols and higher-level protocols. The transport, middleware, or overlay layer handles the locator/identity split and indirection. There
are many possibilities for supporting mobile applications, for example SCTP, mSCTP, TraSH, and the middleware protocols. HIP provides useful features for the higher layers, such as persistent transport-layer connections, multi-homing, and authentication. On the other hand, the timeliness and efficiency of the HIP key/address distribution mechanism is still open and requires interaction with an application-level lookup service, such as DNS or i3. The use of HIP with SCTP, SIP, and other protocols are currently open issues. Each networking layer operates mostly on its own. Due the lack of cross-layer interactions, similar tasks may be repeated multiple times on each layer. We can consider access authentication and service authorization as an example of this. Several networking layers, including the final application-level service, may initiate similar authentication and authorization transactions that all end up at the same AAA (authentication, authorization, accounting) backend system located in the user’s home network or the service provider’s premises.
CONCLUSION We presented taxonomy of mobility-enabling protocols on different layers of the networking stack starting from the network layer. Three useful addressing models were identified that are inherently related with mobility: addresses with both location and identity, locator/identity split, and content-based addressing. The first is currently used in the IP network architecture, and mobility solutions typically use triangle routing with optimizations. The second model has been proposed, because of its mobility-friendly characteristics—mainly a non-fixed point of indirection. The last model has also been proposed for expressive communication and is believed to be a good candidate for supporting mobile and distributed applications. The distinguishing feature of the models is the number of indirection points and whether or not they are fixed. Contentbased addressing is most flexible, but may also require the update of several indirection points, which is more costly than updating a single point. We identified several useful cross-layer interactions, but noted that some protocols are complementary. Especially if middleware mobility support is used, the benefit of using lower-level mobility protocols is uncertain. Middleware mobility support is needed for object, session, and pub/sub mobility.
REFERENCES Burcea, I., Jacobsen, H.-A., de Lara, E., Muthusamy, V., & Petrovic, M. (2004). Disconnected operation in publish/sub-
971
U
Understanding Multi-Layer Mobility
scribe middleware. Proceedings of the IEEE International Conference on Mobile Data Management.
Stewart, R. et al. (2004, June). SCTP dynamic address reconfiguration. Internet Draft.
Calhoun, P. R., Johansson, T., Perkins, C. E., Hiller, T., & McCann, P. J. (2004, August). Diameter Mobile IPv4 application. Standards Track RFC 4004.
Stoica, I., Adkins, D., Zhuang, S., Shenker, S., & Surana, S. (2002). Internet indirection infrastructure. Proceedings of the 2002 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications (pp. 73-86). ACM Press.
Carzaniga, A., Rosenblum, D. S., & Wolf, A. L. (2000, July). Achieving expressiveness and scalability in an Internet-scale event notification service. Proceedings of the 19th ACM Symposium on Principles of Distributed Computing (PODC2000). Devarapalli, V. et al. (2004, June). Network mobility (NEMO) basic support protocol. Standards Track RFC 3963. Eugster, P.T., Felber, P. A., Guerraoui, R., & Kermarrec, A.-M. (2003). The many faces of publish/subscribe. ACM Computing Surveys, (2), 114-131. Fu, S., Atiquzzaman, M., Ma, L., Ivancic, W., Lee, Y.-J., Jones, J. S. et al. (2004, January). TraSH: A transport layer seamless handover for mobile networks. Johnson, D., Perkins, C., & Arkko, J. (2004, June). Mobility support in IPv6. Standards Track RFC 3775. Kaufman, C. (2004, September). Internet Key Exchange (IKEv2) protocol. Standards Track RFC 4306. Kivinen, T. (2006, March). Design of the MOBIKE protocol. Internet Draft. Kohler, E. (2006, January). Datagram congestion control protocol mobility and multi-homing. Internet Draft. Mühl, G., Ulbrich, A., Herrmann, K., & Weis, T. (2004). Disseminating information to mobile clients using publish/ subscribe. IEEE Internet Computing, (May), 46-53.
Tarkoma, S., Kangasharju, J., & Raatikainen, K. (2003). Client mobility in rendezvous-notify. Proceedings of the International Workshop on Distributed Event-Based Systems (DEBS’03). W3C. (2003, June). SOAP version 1.2. W3C Recommendation. Xing, W., Karl, H., Wolisz, A., & Müller, H. (2002, October). M-SCTP: Design and prototypical implementation of an end-to-end mobility concept. Proceedings of the 5th International Workshop on the Internet Challenge: Technology and Applications, Berlin, Germany. Zhuang, S., Lai, K., Stoica, I., Katz, R., & Shenker, S. (2002). Host mobility using an Internet indirection infrastructure. Technical Report, University of California at Berkeley, USA.
KEY TERMS Break-Before-Make: A mobility solution feature where the connection to the old point of attachment must be torn down before establishing a connection to the new point of attachment to the network during handover. Content-Based Routing: The process of forwarding messages based on their content.
Moskowitz, R., Nikander, P., Jokela, P., & Henderson, T. (2006, March). Host identity protocol. Internet Draft.
Horizontal Handover: A handover within the same access technology.
OMG. (2002). CORBA Security Service v.1.8. Object Management Group.
Locator/Identity Split: Form of addressing, in which the identity and location of a node have been separated.
OMG. (2004, April). Wireless access and terminal mobility in CORBA v.1.1. Object Management Group. Perkins, C. (2002, August). IP mobility support for IPv4. Standards Track RFC 3344.
Make-Before-Break: A mobility solution feature that allows establishing connectivity to the new point of attachment in the network prior to tearing down the connection to the old point of attachment during handover.
Schulzrinne, H., & Wedlund, E. (2000). Application-layer mobility using SIP. SIGMOBILE Mobile Computing Communication Review, 4(3), 47-57.
Mobile IP: A classical layer-3 mobility solution based on tunneling and a stable anchor point representing the mobile node.
Soliman, H., Castelluccia, C., Malki, K. E., & Bellier, L. (2005, August). Hierarchical Mobile IPv6 mobility management (HMIPv6). Experimental RFC 4140.
Service or Application Mobility: The mobility of an application or service from one physical location to another.
972
Understanding Multi-Layer Mobility
Session Mobility: The seamless transfer of an ongoing communication session from one device to another.
U
User Mobility: The ability of an end user to send and receive information regardless of mobile terminal and current location. Vertical Handover: A handover between different access technologies.
973
974 Category: M-Business and M-Commerce
Using Mobile Devices for Electronic Commerce Raul Fernandes Herbster Federal University of Campina Grande, Brazil Hyggo Almeida Federal University of Campina Grande, Brazil Angelo Perkusich Federal University of Campina Grande, Brazil
INTRODUCTION Electronic commerce is becoming the most used mechanism for non-traditional commerce. However, several popular delivery services are still accessed via telephone, which enables commerce anytime, anywhere. Such telephony-based services have several problems: they do not offer a more detailed description of available products; users may ask the attendant to repeat the description of a certain product, directly affecting the time of product selling; the number of concurrently attended clients is limited to the number of attendants; and the product list must also be continuously updated, by adding or removing products, but the user cannot be automatically informed about that. Mobile devices offer a sophisticated interface that allows better user interaction by means of lists, menus, multimedia features such as images, and much more. A user can indefinitely explore product categories very fast. It is possible to offer a more detailed description of products, with visual elements such as pictures or even videos. Besides, the number of concurrent accesses depends only on the number of connections supported by the server. In this article, we describe an architecture for mobile commerce which allows the use of mobile devices for electronic commerce. The architecture enables the development of applications to be executed on a mobile device, which lists selling products having their own textual descriptions and pictures. We discuss architectural modules and the implementation of an application for selling fast food called Mobile Menu. We begin with the main background concepts related to our proposed architecture.
BACKGROUND Electronic commerce has attracted significant attention in the last few years (Varsghney & Vetter, 2002). The continuously increasing number of users of mobile devices, such
as mobile phones and personal digital assistants (PDAs), and advances in wireless network technology provide an ideal scenario for offering personalized services to mobile users and give place to the rapid development of mobile electronic commerce (MEC) (Tsalgatidou, Veijalainen, & Pitoura, 2000). The way MEC operates is partially different from Internet e-commerce due to special characteristics and constrains of mobile terminals and wireless networks. The context, situation, and circumstances under which people use their mobile devices are also different (Tsalgatidou et al., 2000). Wireless and mobile networks are increasing in exponential rate in terms of capabilities of mobile devices and user acceptance (Varsghney & Vetter, 2002). Today, more than 1 billion cell phones and other mobile devices are in use worldwide. MEC also has more advantages than traditional e-commerce applications: location-awareness, adaptivity, ubiquity, personalization, and broadcasting (Tsalgatidou et al., 2000). Applications for mobile devices are also easier to use, because the user interface of such devices is very intuitive. Mobile devices have less resources than desktop and mainframes computers: limited memory, disk capacity, and computational power. The user interface of such devices also has some constrains: for example, small screens and small multi-functional keypads (Tsalgatidou et al., 2000). These constrains restrict the variety of applications for mobile devices and must be taken into account when designing new systems for such platforms. Applications that demand a considerable quantity of system resources are harder to develop for mobile devices. For example, applications that need a large database to constantly perform queries and update the data are very difficult to develop for mobile devices, because the limited memory of devices does not support a database management system (DMS). Distributed architecture shares tasks among the elements of it, so that harder activities which demand memory and computational power can be allocated to those which have more resources.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Using Mobile Devices for Electronic Commerce
Figure 1. Mobile menu architecture
U Manager
Internet Client
Database Server
Client-server architecture is largely used as a distributed design; it shares the tasks of the elements and provides a certain level of decoupling. It has two elements that establish communication with each other: the front-end or client, and the back-end or server. The client makes a service request to the server whereas the server provides service to the request. The client-server architecture allows an efficient way to interconnect programs that are distributed at different places (Jorwekar, 2005). However, client-server architecture is more than just a separation of a user from a server computer (Fastie, 1999). Each portion also has its own modules: presentation, which handles inputs from devices and outputs to screen display, application, and data; application, which has the rules of the business; and data, which provides services for storing the data of the application (Fastie, 1999).
AN ARCHITECTURE FOR MOBILE COMMERCE We propose an architecture that enables mobile commerce for mobile devices. The architecture is illustrated in Figure 1. The application has three elements: the client, which requests the information about selling products; the manager, which updates information on the products; and the server, which receives requests and sends responses. To start with, the user accesses the service anytime and establishes a connection with the server. Then, a list with pre-defined categories of products is sent to the user. These categories help the user to browse through the list of products. After selecting the product, the user can obtain more specific information about it or purchase the item, if more detailed information about the product is requested. At the other side, the server receives the request of purchasing or
obtaining more information about the product, such as name, description, and price. Other more elaborate elements that describe the products, such as pictures and videos, can be attached. The application can be accessed anytime. The client application can run on a mobile device and establishes a connection with the server. It requests services to the server and receives the data. The server has two modules: the network layer, which manages the network connection of the mobile devices; and the database layer, which establishes connection with the database and manages data. Another important element of the application is the manager side. It is a desktop application which interacts with the system by modifying the database: it inserts or removes products and also modifies information about them, like price, name, and description. This architecture provides an interesting solution by delegating tasks for members of the system: client, server, and manager. The tasks performed by each one does not demand a considerable amount of resources from each system. For example, a mobile device cannot store a large amount of data. Thus, the architecture delegates the storing/processing tasks to the server that is supposed to have more resources. The layered architecture decouples the modules, so each one can be modified interchangeably, and also provides reuse of code.
Mobile Menu Food delivery is a frequently used service usually accessed via telephone. It has some problems that affect the business, for example: detailed descriptions about the products are not easy available; for each client, an attendant is allocated; and a request can be mistakenly made. Based on the architecture described, we developed Mobile Menu, which is an application for fast food mobile commerce. 975
Using Mobile Devices for Electronic Commerce
Figure 2. Client-server protocol Client
Server request_connection
1
accept_connection send_menu
2
subcategorie
*
3
send_list_products
request_description(ID)
*
4
send_description
It was implemented using Java™ Platform–J2SE for desktop applications and J2ME for mobile applications. The functioning of Mobile Menu is represented in Figure 2. The client connects to the server and requests the menu, which is sent by the server (1, 2). The first part of the menu has only categories and subcategories of the products (pastas, salads, drinks, beverage, etc.) (3). If the client wants to see the products of a specific subcategory, it sends a request to the server and the latter sends a list of all products (3). This list contains the name of each product and its ID. If the user wants to know a detailed description of the product, the client sends a request to the server and receives additional information (4). The menu can be updated anytime through the manager side of the application. Mobile Menu adds considerable business value in food delivery services, usually accessed via telephone. The user can indefinitely explore the categories very quickly. It is possible to offer a more detailed description of products, with visual elements like pictures or even videos on them. The number of accesses depends only on the quantity of connections supported by the server.
FUTURE TRENDS Mobile e-commerce has several classes of applications (Timmers, 1999) with different characteristics. For example, Internet banking is a very different kind of mobile commerce which is not explored here. As for future works, we propose enhancing the architecture so more classes of mobile applications can use it. For example, the architecture does not solve problems like selling multimedia products such as video to mobile devices.
976
These devices have limited memory/processing. Thus, storing and playing large videos with high quality, for example, is not viable. Another aspect that was not explored is data security: the messages are not encrypted and the channel is not secure. The architecture can be adapted by adding other modules to solve some security problems.
CONCLUSION Mobile e-commerce is an area that creates opportunities for many players in the field. Mobile devices have some constraints that must be taken into account when designing the application for such platforms. However, these devices also have important characteristics that make them an interesting channel of commerce. The number of services available through mobile devices is increasing: banking, purchasing of images and music, and much more. These applications make services easily accessible to the client, providing consumption of a large variety of products and services in a convenient way. In this article we proposed an architecture for mobile devices that enables the use of mobile devices for electronic commerce. We applied the proposed architecture to develop Mobile Menu, an application for food delivery based on mobile commerce. Although this application is specific for such a domain, the architecture is generic and could be adapted to any mobile commerce application.
REFERENCES Fastie, W. (1999). Understanding client/server computing. PC Magazine, 229-230. Jorwekar, S. (2005). Client server software architecture. Pitoura, E., & Samaras, G. (1998). Data management for mobile computing. Kluwer Academic. Timmers, P. (1999). Electronic commerce: Strategies and models for B2B trading. New York: John Wiley & Sons. Tsalgatidou, A., Veijalainen, J., & Pitoura, E. (2000). Challenges in mobile electronic commerce. Proceedings of the 3rd International Conference on Innovation Through E-Commerce, Manchester, UK. Varsghney, U., & Vetter, R. (2002). Mobile commerce: Framework, applications and networking support. Mobile Networks and Applications, 7, 185-198. Varsghney, U., & Vetter, R. (2000). Emerging wireless and mobile networks. Communications of the ACM.
Using Mobile Devices for Electronic Commerce
Wesel, E. K. (19998). Wireless multimedia communications, networking video, voice and data. San Francisco: Addison-Wesley.
KEY TERMS Client-Server Architecture: A basic concept used in computer networking, wherein servers retrieve information requested by clients and clients display that information to the user. Distributed Network: A system where resources are spread among many computers, instead of being stored in a single location Electronic Commerce (E-Commerce): The buying and selling of information, products, and services electronically over the Internet.
Mobile Devices: Any portable device used to access a network (Internet, for example). Multimedia Application: Applications that support the interactive use of text, audio, still images, video, and graphics. Protocol: A set of rules and procedures governing communication between entities connected by the network. User Interface: The means by which an individual communicates with a computer through a software application. The common methods for such communication are, commands, menus, and icons. Wireless Network: Networks without connecting cables, that rely on radio waves for transmission of data.
977
U
978 Category: Converging Technology
Using Mobile Devices to Manage Traffic Infractions Stefânia Marques Federal University of Campina Grande, Brazil Sabrina Souto Federal University of Campina Grande, Brazil Miguel Queiroga Federal University of Campina Grande, Brazil Hyggo Almeida Federal University of Campina Grande, Brazil Angelo Perkusich Federal University of Campina Grande, Brazil
INTRODUCTION
BACKGROUND
Mobile computing is one of the recent technologies with the most impact on people’s lives. Several research and industrial applications are benefiting from mobile computing, supporting various human daily activities. Transit law enforcement officials can benefit from the availability of powerful mobile devices, such as smart phones and PDAs, to help them to execute their daily tasks. In such a scenario, an official can verify a driver’s data record and issue tickets online. In this article we describe the SM-FIT system that makes it possible for transit law enforcement officials to perform online queries about potential infractions of a driver of a vehicle by using a mobile device. Queries are performed based on a unique identifier: the driver’s license number. The system is implemented based on a client-server paradigm, where mobile devices are clients and servers are base stations. Clients must have a local database to store each result of a query, when needed. Each registry stored has the following attributes: a unique identifier, the number of the vehicle’s plate, the date and time that the officer registered the infraction, and the status of the infraction. Besides, photographs can be stored, digitally signed, and transmitted to a database for future prosecution. The remainder of this article is organized as follows. We first outline some background concepts related to the system’s development. We then present the proposed system architecture and functioning, and discuss some trends related to future research in this area. We close with some final remarks.
This section describes briefly some concepts related to J2ME and, more specifically, MIDlets. Such technologies have been used for developing the proposed system.
J2ME J2ME is a development platform based on Java Technology for developing mobile and embedded applications. It focuses on two types of devices: •
•
High-End Consumer Devices: • CDC (connected device configuration); • interactive TVs, videophones, wireless devices; • a large variety of user interfaces; • typical memory of 2 to 4 Mb; and • persistent connection, generally TCP/IP. Low-End Consumer Devices: • CLDC (connected limited device configuration); • cell phones, bidirectional pagers, PDAs, and so forth; • limited processors (8 to 32 MHz); • limited memory; • lazy connection, intermittent (9600bps) and generally not based on TCP/IP; and • powered by batteries.
The J2ME platform includes flexible user interfaces, a robust security model, a broad range of built-in network Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Using Mobile Devices to Manage Traffic Infractions
Figure 1.
U
2a 2a
1
2b
2
2c
protocols, and extensive support for networked and off-line applications. Besides this, applications based on J2ME specifications are written once for a wide range of devices.
resources, like images, and a manifest file. This manifest file contains a list of attributes and definitions to be used by application managers to install the JAR files in the device.
MIDlets
Security in MIDlets
Java applications running on MIDP devices are known as MIDlets, which consist of at least one Java class and have to be derived from the abstract class javax.microedition.midlet. MIDlet. These MIDlets use an execution environment within the Java Virtual Machine to control the application’s lifecycle through a set of methods implemented by this MIDlet. MIDlets can also use methods to obtain services from the environment. A group of related MIDlets can be put together in a MIDlet suite, which is packaged and installed in (or removed from) a device as a unique entity. MIDlets in a suite share all static and dynamic resources in their environment:
The JAVA security model in its standard edition (J2SE) is too expensive in terms of costs for memory allocation, and it requires configuration knowledge that is not present in users of mobile devices. Thus, neither CLDC nor MIDP include these functionalities. Cryptography of public key and certifiers are not available as default, so it is necessary to pay attention when installing MIDlets and, preferentially, only accept software from trustable fonts. MIDP 2.0 included the https protocol that helps to diminish these problems.
• •
Execution data can be shared by MIDlets, and the usual Java conventions of synchronization can be used to control data access. Persistent data can also be accessed by all MIDlets in a suite.
All files in a MIDlet suite must be within a JAR package. These packages contain the classes of the MIDlet and other
SYSTEM DESCRIPTION The SM-FIT is a system that makes it possible for transit law enforcement officials to register transit irregularities in a local database. Each record of this database consists of the number of the vehicle’s plate, a code of the infraction, date and time that the infraction occurred, and the photography of the vehicle involved in the corresponding infraction.
979
Using Mobile Devices to Manage Traffic Infractions
The system is implemented based on a client-server paradigm, where mobile devices are clients and servers are base stations. At the end of the day, transit law enforcement officials send the registers of their mobile devices to the server in order to make available more space in their devices for local databases. Consider Figure 1 to understand the system functioning. Through the starting screen of the system (1), it is possible to access the system main menu (2). This menu of options includes: Add a record, Send records, and View records. If the user chooses the first option (2a), to add a record, it is shown a screen with the following data to be filled: the vehicle’s plate involved in the infraction, the code of the infraction, date and time of the infraction (this information is taken automatically from the mobile device), and the photography of the vehicle involved in the infraction. If the user wants to send all the records stored in the mobile device to the server (2b), it is shown a screen that requests the IP address of the server. Then, the process of sending the information to the database is initiated. Finally, if the user chooses to visualize the records of the mobile device (2c), it is shown a screen with a list of records contained in the local database. This visualization is taken in a way that each record is shown individually on the screen.
FUTURE TRENDS Future research and development of new technologies may make possible the transferring of streams from mobile devices to the server. In this way, the system described here could store videos of infractions or even the voice of the transit law enforcement official describing how the infraction occurred. For this kind of improvement, it would be necessary to consider efficient lower battery consumption mechanisms. Another point to be considered is cryptography in mobile devices, which would increase security during the transference of records from the mobile devices to the server. Considering that it is an industrial application, security is an essential requirement. In the context of the driver, a module for drivers to consult the situation of their licenses or cars could be made available. This module would make available all the information contained in the server database.
CONCLUSION The usage of mobile devices to manage traffic infractions can bring some benefits, such as saving time on queries made to check if drivers have any previous infractions. Another benefit is that using a mobile device that can take pictures, when the 980
transit law enforcement official registers a new infraction, he/she can also take a picture of the vehicle involved in the infraction to prove in the future that it really occurred. Regardless of application domain, there is a widespread use of mobile devices and an increasing need for real-time answers wherever a person is. This makes it necessary to use a technology like the SM-FIT system. It allows for transit law enforcement officials to get real-time answers for their queries about the drivers’ situations, and also about their vehicles. But this system can also be adapted to be used in many other application scenarios, such as to make restaurant reservations, access credit card accounts, and make payments.
REFERENCES Alves, D. (2004). Introdução ao J2ME. Retrieved November 30, 2004, from http://www.conexaojava.com.br/conexaojava04/download/minicursos/Java2.Micro.Edition-Conexao. Java.2004.pdf Borges, R. L. (2004). J2ME na prática. Retrieved November 30, 2004, from http://www.ucb.br/java/JavaDays/J2ME_ RosfranBorges.pdf Easy Process. (n.d.). A software development process. Retrieved from http://dsc.ufcg.edu.br/~yp Gomes, H. M. (2004). Arcabouços de software para desenvolvimento de aplicações embarcadas. Retrieved November 30, 2004, from http://www.dsc.ufcg.edu.br/~hmg/disciplinas/nokia/ASDAE.pdf XP1. (n.d.). A software development process. Retrieved November 30, 2004, from http://www.dsc.ufcg.edu.br/~jacques/ cursos/2002.2/projii/xp1/xp1.html
KEY TERMS Base Station: A centralized repository for the storage and management of information, organized for a particular area. CDC: Connected device configuration. CLDC: Connected limited device configuration. IP Address: An Internet protocol address attributed to a client or a server in the client-server paradigm. J2ME: Java Second Micro Edition. MIDP: Mobile information device profile. Personal Digital Assistant (PDA): A handheld device that combines computing, telephone, Internet, and networking features.
Category: Service Computing 981
Using Service Proxies for Content Provisioning U Panagiotis Kalliaras National Technical University of Athens, Greece Anthanasios-Dimitrios Sotiriou National Technical University of Athens, Greece
INTRODUCTION In modern broadband mesh networks, communication between two end nodes is carried out not directly, but through a number of intermediate nodes. While these nodes’ only function may be to relay information from one point to another, they may also host computational elements which perform some service on behalf of other applications. We deal with the problem of optimally mapping multimedia content transcoding service elements onto network resources. There may be several places in the network where the required compression and decompression services could be performed. We would like to select the best locations that meet the application’s requirements. We propose a new approximation algorithm for constrained path optimization, which provides better scalability and simplicity than previous approaches. This is accomplished basically by partitioning the overall problem into smaller ones.
RELATED WORK The majority of the proposed schemes are focused on solving the similar multi-constrained optimal-path problem (MCOP). This problem aims to find in a network an optimal path that satisfies multiple additive path constraints and has been proven to be of NP-complete complexity, therefore unsolvable in polynomial time. Several algorithms have been proposed for the above problem. For the MCOP problem with two parameters, Jaffe (1984) proposed to use a linear weight combination of the two constraint parameters. Other proposed algorithms include Iwata et al. (1996), SAMCRA (Van Mieghem, De Neve, & Kuipers, 2001), and the ChenNahrstedt algorithm (Chen & Nahrstedt, 1998). All of the above algorithms require a global state to be maintained at every node. Most algorithms transform the routing problem to a shortest path problem and then solve it by Dijkstra’s or the Bellman-Ford algorithm. The concept of service path has been proposed in smaller numbers. In TranSquid (Maheshwari, Sharma, Ramamritham, & Shenoy, 2002), a transcoding and caching proxy for heterogeneous clients is proposed. In the Ninja Project (Gribble et al., 2001), service path is defined as a sequence
of application-level service operators and connectors. In Lienhart, Holliman, Chen, and Yeung (2002), the authors propose the addition of a media support module (MAPS) on top of an existing peer-to-peer service layer, in order to improve multimedia services across heterogeneous computing platforms. This module is responsible for transcoding and route path selection based on the single-pair shortestpath problem and utilizes Dijkstra’s algorithm to provide a solution. Our system is a combination of the above two research areas. It uses the service paths concept and also fully implements the MCOP, rather than the simpler pathfinding solutions.
NETWORK AND SERVICE MODELING We assume that our network topology matches that of a partial mesh network. A mesh network is reliable and offers redundancy. If one node can no longer operate, all the rest can still communicate with each other, directly or through one or more intermediate nodes. Mesh networks work well when nodes are located at scattered points that do not lie near a common line. As a service, we denote any network resource which may include computational elements and performs some online activity on behalf of other applications. A service is an online facility that is always available to all requestors, at a predefined cost and delay. The services may be available on more than one node, either serially or concurrently. Although services in communication networks delivering multimedia content may include conversion processes like media data adaptation, merging of multiple media sources, copyright protection, metadata extraction, enhancements, and recovery, in our modeling we focus mainly on transcoding. Possible forms of transcoding include: lowering the bit rate of a media stream by reducing the image/video resolution, size, and/or frame rate; converting a media stream from one encoding format to another; or a combination of the above. The expected benefits from these adaptations are on one hand to move computation (data transformation) from the client site to the proxy, and on the other hand to reduce the volume of data transferred to the client. In any case, a service in our modeling accepts a media stream which is
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Using Service Proxies for Content Provisioning
Figure 1. Network and service paths
are part of the actual network and services, and requests are part of the overlay network.
PROBLEM ENTITIES Edge e
Vertex v
Service s
Request req
Be: available bandwidth
dv: delay
vs: host vertex
vreq: host vertex
de: delay
rin/rout: input/ output rate
ce: cost per rate
cs: cost
max Dreq
: max delay
Rreq: requested rate
ds: delay
characterized from an input data rate, performs applicationlevel processing on the media data, and forwards the stream at a new output data rate. The connection requests are initiated by clients that wish to acquire specific content from content providers. The procedure that the clients follow to detect content providers with desirable content is out of the scope of this article. The notion of clients, content providers (CP), and service providers (SP) is used here to state the dynamic nature of the network connections. The scope of multimedia service provision is to provide clients with customized and satisfactory QoS, under the constraint of end-to-end resource availability observed by each client. Several requests are initiated, targeting various service entities. The main pattern for the proposed solution is to split the problem into smaller ones, as shown in Figure 1. Each request forms two bids. The first bid concerns the path from the request to the service that has the desired output. This bid includes the network and service paths, and the delay and cost for the path and the usage of the service. The second bid is required if the chosen service does not own the content, and acquisition is needed from somewhere else. In this case, a new request is formed and its result is returned to the initial request. The results from the second bid contain everything after the service, including any subsequent services that may be used.
PROBLEM FORMULATION AND HEURISTICS Let directed, connected graph G(V, E) denote the network topology, where V is the set of vertices of the graph (representing network components (e.g., switches, routers, hosts, aggregated subgraphs) and E the edges (representing communication links). There are four basic entities used in our problem formulation, shown in Table I. Edges and vertices 982
For every request, ensure the end-to-end delay constraint is met:
∑d + ∑d + ∑d v∈P
v
e∈P
e
s∈SP
s
max ≤ Dreq
(1)
while the available bandwidth for all edges in the path is at least the required rate at that point: Rreq ≤ Be , ∀e ∈ P
(2)
and minimize cost: Ctot = ∑ Ce ∗ re + ∑ c s e∈P
s∈SP
(3)
where P is the network path and SP the service path. The problem, as described above, is more accurately described as unicast link-constrained path-constrained path-optimization routing. The link constraint refers to the available bandwidth, the path constraint to the total delay, and the desirable path optimization to the minimization of the total cost.
The Heuristics Solution The main steps for estimating the optimal paths are the following: •
•
Step 1. Topology Filtering: For every request req initiated at host vertex t and inquiring an object with max optimal rate Rreq and maximum delay Dreq , filter all links (and possibly disconnected nodes) that do not satisfy the requested minimum linear QoS constraints, in our case, minimum available bandwidth Be as in equation (2). Step 2. Finding Available Services: Let SR ∈ S be the sum of services that have an output rate Rreqthat is,
Using Service Proxies for Content Provisioning
•
double a=1, b=0, step=0.1; while (delay1 > Dreq1║b!=1) { a=a-step; b=b+step; P+new Dijkstra(a,b); delay1=getDelay(P); }
•
•
O(Nlog N + 2E). Also, for a total number of S services, the algorithm can be run S! times, that is S times for the first time and S-k, for the k-th iteration, when s services will have been added to the service path. However, the algorithm is bounded by the maximum delay Dreq, so ~ given that each service s adds a mean delay d s (service ~ delay + path delay), the iteration happens k=Dreq/ d s times and the MCOP algorithm of bid1 is executed S! / (S-k)! < Sk times. So, the overall complexity of our heuristics is O(S!/(S-k)!(Nlog N + 2E)).
the services that either own or can produce the object are chosen. Step 3. Bid 1: For every service s ∈ SR hosted at vertex ts, solve the partial MCOP problem of (1) and (3) from treq to ts. In our implementation, we use an extended version of Jaffe’s algorithm, combining: w(P) = a * c(P) + b * d(P), in order to obtain the results. The main difference from the original algorithm is the usage of non-static variables a and b, which change according to the pseudocode below:
The logic is simple. First seek to optimize path with respect to cost only and check if the delay constraint is met. If not, increase the delay factor by a step and try again until a solution is found. Step 4. Bid 2: For every node v ∈ V that hosts a service s ∈ SR, if s has an input rate Rin ≠ 0, form a new request for object with optimal rate Rin and maximum delay max Dreq 2 < Dreq .Go to step 1. Results are path2, spath2, cost2, and delay2. Step 5. Negotiation of Bids: The negotiation process deals with the sharing of partial delay constraints Dreq1 and Dreq2 between bids 1 and 2 respectively, so that max delay1+delay2 < Dreq and the overall cost is kept minimal. The negotiation is done according to the pseudocode as follows: PROCEDURE bid() Dreq1 = D max – delay2 req delay = tryBid1(Dreq1) max
Dreq2 = D req – delay1 delay2 = tryBid2 (Dreq2) do if (delay1 < 0 && delay2 < 0) No feasible solution for bid1 and bid2, bid fails if (delay1 < 0 && delay2 > 0) ask trybid2 () fro a better delay, try bids again Dreq2 = delay2 – 1 if (delay1 > 0 && delay2 > 0) valid delays for both bid1, bid2, bid succeeds if (delay1 > 0 && delay2 < 0) ask trybid () for a better delay, try bids again Dreq1 = delay1 – 1 while bid succeeds or bid fails END PROCEDURE
The worst case overall complexity of the algorithm is analogue to the complexity of the Jaffe Algorithm:
SIMULATION RESULTS The algorithm described above was implemented in Java, using the JUNG (http://jung.sourceforge.net/) software package. The simulations were conducted using three different network topologies: •
• •
Grid 16: This network is composed of 16 vertices that form a grid network. Each vertex is connected to its two neighbors to the east and south, for a total of 32 edges. The east-most and south-most nodes at the edges of the square grid also have links that wrap around to the corresponding node at the opposite side, which results in a grid topology. Grid 64: This network has the same structure as the Grid 16 network described above, but has 64 vertices and a total of 128 edges. Random: The network is composed of 40 vertices and 86 edges.
For each of the above topologies, two major sets of simulations were contacted: • •
With CPs but without TSs, simulating the classic QOSR MCOP problem. With a variable number of service proxies. The optimal route selection heuristics described above was used in this case.
The following configuration parameters affect the simulation results: •
• •
Density and Location of Services: The density of servers is defined by the ratio of the number of services to the total number of nodes. The host nodes for the services were randomly selected for Grid64 and Random, while for Grid16, all nodes hosted services. Bandwidth Reserved per Request: The bandwidth that an individual request reserves at each link is set approximately to 4% of the available link bandwidth. The values for the delays in edges and nodes are of the same order and 70% less than the delay caused by 983
U
Using Service Proxies for Content Provisioning
Figure 2. The Grid 16 network V0
V4
V8
V12
V1
V5
V9
V13
V2
V6
V10
V14
V3
V7
V11
V15
services, while the cost of transcoding is generally less than the cost of content provider.
Finding the Optimal Position for the Transcoding Services The algorithm that was described previously was used in order to find the most suitable host nodes for positioning transcoding services. The Grid 16 network, shown in Figure 2, is used as an example. The following parameters hold for nodes, edges, services, and requests: • • •
Every node causes a delay=1. Every edge has cost/rate=1, delay=4 and bandwidth=1500. Two services at V0 are content providers: • S0→10@V0 with delay=10 cost=130. • S0→20@V0 with delay=10 cost=20.
Figure 3. Importance of services
We wanted to find out where to put a number of TS with S10→20 and S20→10 with delay=4 cost=10 that would serve as many requests as possible. Every other node (V1-V15) was a potential host for services, so we put TSs in all of them. Next, we placed two requests for the two rates (10, 20) that were supported in our network on each node other than V0 which hosted the CPs. Their delay constraint was loose and equal to 50. In total, there were 30 requests. As expected, from Figure 3 we see that the most useful services for implementation in our test network are the S20→ 10@V1 and S20→10@V4. These services compensate for the high cost of S0→10@V0. They are preferred by all 15 requests that demand rate=10, while the S0→10@V0 serves zero clients and is actually ignored. The S20→10@V1 and S20→10@V4 services are also used for five requests that demand rate=20, despite the fact that there is already a CP (S0→20@V0) that produces it. These requests come from nodes that are far away from node V0, and it is more convenient for them to choose a service path that would reduce the cost spent at the edges of the network. For example, for node V7, the service path is SP= S0→20@V0 S20→10@V1 S10→20@V7. From Figure 4, we confirm the fact that most network paths contain the nodes that host the most preferred services, particularly E0(V0,V1), E1(V0,V4), and E3(V1, V5). The results from the simulations that we conducted can lead us to certain useful considerations. Using transcoding services is necessary when there is no content provider that can satisfy a client request or its costs are unacceptable. When transcoding results to higher bit rate, the transcoder should be placed as close to the destination as possible. Accordingly, when transcoding results to lower bit rate, the transcoder should be placed as close to the content provider as possible. Transcoding resulting in lower bit rate may be necessary even when it leads to worse performance (i.e., excess delay)
Figure 4. Importance of edges
35
25
30
20
Requests served
20
15
15
Se rvice s
984
E5(V2,V6)
E16(V8,V9)
E9(V4,V5)
E19(V9,V13)
E17(V8,V12)
E13(V6,V10)
E26(V13,V14)
Edges
E20(V10,V11)
E8(V4,V8)
E11(V5,V9)
E12(V6,V7)
E3(V1,V5)
E10(V5,V6)
E1(V0,V4)
0
E6(V3,V0)
0
E4(V2,V3)
5
E2(V1,V2)
5
E0(V0,V1)
10
10
S0 -> S2 20 0- @ > V S2 10 0 0 - @V > S1 10 1 0- @ S1 > 2 V4 0 - 0@ > S1 20 V7 0- @ > V S1 20 10 0 - @V > S1 20 11 0- @ > 2 V1 0@ 3 V1 4
Requests served
25
Using Service Proxies for Content Provisioning
Figure 5. Cost-delay diagram for (a) Grid 16 and (b) Grid 64 networks
Figure 6. Cost-delay diagram for random network
U
(a)
Overall Network Statistics
(b)
when the available bandwidth is low and many requests need to be served. Transcoding has better results when the cost and delay of the transcoding service is relatively low, and also when the host node of the service is included in the optimal path between the content provider and the receiver. In other words, the deviation of the flow for using a service must be minimal. The rate reduction is important and the request has not very tight delay constraints.
In the cost-delay diagrams shown in Figures 5 and 6, we show the improvements on performance by using service proxies. In all cases, only the minimum required content providers were used in the network. In Figure 7(a) for total number of 200 requests, and in Figures 7(b) and 8 for 500 requests, the ratio of served requests is shown for each network with and without transcoding services. For the Grid 16 and Grid 64 networks, the percentages are about 90% and 80% respectively with the usage of services. Again, the benefits of using services are more evident when the network tends to reach congestion. As the optimized network resource usage increases, so does the admission control rate. The random network has an even better comparative performance admission control rate, which is evident from the beginning due to the fact that certain client requests are unfulfilled due to the limited number of content providers.
Figure 7. Admission control diagram for (a) Grid 16 and (b) Grid 64 networks a a b
b
a
b
(a)
a
b
(b) 985
Using Service Proxies for Content Provisioning
Figure 8. Admission control diagram for random network
Lienhart, R., Holliman, M., Chen, Y. K., & Yeung, M. (2002). Improving media services on P2P networks. IEEE Internet Computing.
a
b
a
Jaffe, J. (1984). Algorithm for finding paths with multiple constraints. Networks, 14(1), 95-116.
b
CONCLUSION In this article we have presented a distributed approach for finding the best feasible paths in networks with respect to QoS requirements. We focused on the usage of trascoding services and especially services that alter the bit rate. This approach is applied in the unicast distribution of multimedia streams. Our heuristics involved transformation of the original problem to a series of conventional shortest path QoSR problems that are solved using a new MCOP algorithm. Through our extensive simulations we proved that transcoding proxies can induce positive results in reducing costs, improving load balancing/network congestion, and consequently increasing admission control rates.
Maheshwari, A., Sharma, A., Ramamritham, K., & Shenoy, P. (2002, February). TranSquid: Transcoding and caching proxy for heterogeneous ecommerce environments. Proceedings of the 12th IEEE Workshop on Research Issues in Data Engineering (RIDE’02), San Jose, CA. Van Mieghem, P., De Neve, H., & Kuipers, F. A. (2001). Hop-by-hop quality of service routing. Computer Networks, 37(3-4), 407-423.
KEY TERMS Content Provider: A service that provides multimedia content. Mobile Computing: Ability to use technology untethered, that is not physically connected, or in remote or mobile (nonstatic) environments. Multimedia: Media that uses multiple forms of information content and information processing (e.g., text, audio, graphics, animation, video, interactivity) to inform or entertain the (user) audience. Network: A network of telecommunications links arranged so that data may be passed from one part of the network to another over multiple links.
REFERENCES
Network Topology: The pattern of links connecting pairs of nodes of a network.
Chen, S., & Nahrstedt, K. (1998). On finding multi-constrained paths. Proceedings of ICC’98, New York (pp. 874-879).
Path Optimization: Finding a path between two vertices such that the sum of the weights of its constituent edges is minimized.
Gribble, S. et al. (2001). The Ninja architecture for robust Internet-scale systems and services. Computer Networks, (Special Issue on Pervasive Computing).
QOS Routing: Process of finding a loop-less path between nodes in a network, satisfying a given set of constraints on parameters like bandwidth, delay, etc.
Iwata, A., Izmailov, R., Lee, D.-S., Sengupta, B., Ramamurthy, G., & Suzuki, H. (1996). ATM routing algorithms with multiple QoS requirements for multimedia Internetworking. IEICE Transactions and Communications, E79-B(8), 999-1006.
Service: Any network resource that performs some online activity on behlaf of other applications.
986
Category: Mobile Software Engineering 987
Verifying Mobile Agent Design Patterns with RPOO Elthon Allex da Silva Oliveira Federal University of Alagoas - Campus Arapiraca, Brazil Emerson Ferreira de Araújo Lima Federal University of Campina Grande, Brazil Jorge César Abrantes de Figueiredo Federal University of Campina Grande, Brazil
INTRODUCTION The act of modeling concurrent distributed systems is not a trivial task. Besides, when mobility is added to the scenario, things get worse because of new problems like variations in the communication conditions and remote execution. An important thing to be considered is how to analyze these mobile agent-based systems in order to validate and improve them. An interesting tool to verify such systems is named Petri net. Petri net, or simply PN, is a powerful formal, graphical, and executable tool commonly used for verifying communication protocols and concurrent systems. A variant of Petri net, named RPOO (Guerrero, 2002)—Object Oriented Petri net, brings all the advantages of Petri nets and object-oriented semantics and puts them together. This union produces a new tool, with the formal semantics of Petri net models and the object-oriented semantics of some programming languages, like C++ and Java. This makes the semantics of formal behavior models closer to the semantics of those programming languages. With this, formal models of programs are built easier than with classical Petri nets. Besides, a closer model, speaking about semantics, is used to verify and analyze mobile agent design patterns which gives a more trusted verification. This article presents the formalization and analysis of three migration design patterns—itinerary, star-shaped, and branching—done by using of RPOO. A brief comparison between RPOO models and classical Colored Petri net (Jensen, 1992, 1997) models is also briefly presented.
BACKGROUND Mobile Agent Migration Design Patterns In this article we consider three migration design patterns proposed in Tahara, Ohsuga, and Honiden (1999): itinerary,
star-shaped, and branching patterns. In the sequence we detail each pattern. We used a message sequence diagram to show an overall picture of the design patterns.
Itinerary This pattern provides a way to execute the migration of an agent, which will be responsible for executing a given job in remote hosts. The agent receives an itinerary on the source agency, indicating the sequence of agencies it should visit. Once in an agency, the agent executes its job locally and then continues on its itinerary. After visiting the last agency, the agent returns to its source agency. This pattern is a good solution to agents that need to execute sequential jobs. In Guedes, Machado, and Medeiros (2003) and Medcraft (2003), case studies that apply this pattern are shown. In Figure 1, we present a possible execution sequence for this pattern. We use a notation that is equivalent to the one presented in Klein, Rausch, Sihling, and Wen (2001). In this notation, an object is used to represent an entity that controls agents’ execution in a given agency (creation, destruction, migration) and indicates their location. Migrations are represented by message passing from one agency entity to the other. The message is labeled as MIGRATING AGENT. Before migration, agent execution is interrupted (arrow labeled as destroy()). Execution is continued in the target agency (arrow labeled as initialize()). As we can see, there are three agencies: a SourceAgency and two search agencies (DestinationAgency1 and DestinationAgency2). Following the diagram, we see that there is an agent (ItineraryAgent) that sets its itinerary, moves to the first search agency where it executes its job, then it moves to the second one, executes the job, and returns to the source agency.
Star-Shaped On the star-shaped pattern, the agent receives a list of agencies that it has to migrate to. Initially, the agent migrates to
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
V
Verifying Mobile Agent Design Patterns with RPOO
Figure 1. Message sequence chart for Itinerary DestinationAgency2:
DestinationAgency1:
SourceAgency: :create()
ItineraryAgent: :setItinerary() :nextDestination()
:move() :destroy() :MIGRATING AGENT
:initialize()
ItineraryAgent: doJob :nextDestination()
:move() :destroy() :MIGRATING AGENT :initialize()
ItineraryAgent: doJob :nextDestination()
:move() :destroy()
:MIGRATING AGENT :initialize()
Figure 2. Message sequence chart for Star-Shaped DestinationAgency2:
DestinationAgency1:
SourceAgency: :create()
StarShapedAgent: :setItinerary() :nextDestination()
:move() :destroy() :MIGRATING AGENT :initialize()
StarShapedAgent: doJob
:move() :destroy() :MIGRATING AGENT :initialize() :showResults() :nextDestination() :move() :destroy() :MIGRATING AGENT :initialize()
StarShapedAgent: doJob
:move() :MIGRATING AGENT
:destroy()
:initialize() :showResults()
the first destination agency in the list. After migration is completed, it executes the relevant job and resumes migration going back to the source agency. The agent repeats this cycle until the last agency on its list is visited. The advantage of this pattern is that the agent stores the results of its job in the source agency and does not need to migrate to the others’ agency with them. Depending on the application, the results can be shown to the user as soon as the agent stores them in the source agency. In this way, the user can 988
already know the partial results before the agent finishes its migration through all search agencies. In Figure 2, we can see an execution sequence for the Star-Shaped pattern. In this diagram, we have the same configuration of the sequence diagram shown for the itinerary pattern: three agencies and one agent. Following the diagram, we observe that the agent sets its itinerary and then travels to the first search agency. After executing its job, the agent returns to the source agency, where it stores the job’s result.
Verifying Mobile Agent Design Patterns with RPOO
Figure 3. Message sequence chart for branching DestinationAgency1:
SourceAgency: :create()
V
DestinationAgency2:
BranchingAgent: :setItinerary()
:clone() :create()
BranchingAgent(Clone):
:move() :destroy()
:MIGRATING AGENT :initialize()
BranchingAgent: :initialize()
:move() :destroy()
:MIGRATING AGENT :initialize() :move() :destroy() :MIGRATING AGENT :initialize()
BranchingAgent(Clone): doJob
:move() :MIGRATING AGENT
:destroy()
:initialize()
After that, the agent travels to the second search agency, executes its jobs, and returns to the source agency, storing the results obtained.
Branching In the branching pattern, the agent receives a list of agencies to visit and clones itself according to the numbers of agencies in the defined itinerary. Each clone is assigned an agency from the received list. Each clone has to migrate to its corresponding agency, execute its job, and notify the source agency when the job is completed. The importance of this pattern is that it splits the tasks that can be executed in parallel. The treatment of the final results is an issue not covered by this pattern. For instance, the clones can put the result of the task in a user interface or send it to another agent. Figure 3 shows an execution sequence for this pattern, in a scenario where there are three agencies and one agent. Following this figure, we can see that the agent sets its itinerary and then clones itself. After that, each agent (the original and the clone) migrates to a search agency, where they execute the job and then return to the source agency.
Petri Nets A colored Petri net (Jensen, 1992, 1997), or simply CPN, is a formal method with a mathematical base and a graphical notation for the specification and analysis of systems with characteristics such as concurrent, parallel, distributed, asynchronous, timed, among others. For the mathematical definition, the reader can refer to Jensen (1992, 1997). The graphical notation is a bipartite graph with places, repre-
sented as ellipses, and transitions, represented as rectangles. Transitions represent actions, and the marking of the places represents the state of the model. A marking of a place at a given moment is the token present at that place. A token can be a complex data type in CPN/ML language (Christensen & Haagh, 1996). Each place has an associated color set that represents the kind of tokens the place can have. The transitions can have guards and code associated to it. Guard is a Boolean expression that must be true for the transition to fire. Code can be a function that is executed every time the transition fires. Arcs go from places to transitions and from transitions to places, and never from transition to transition or place to place; they can have complicated expressions and function calls associated to them. For a transition to fire, it is necessary that all input places, that is, places that have arcs that go from the place to the transition, have the number of tokens greater than or equal to the weight of the arc, w(p,t), and the guard of the transition must be true. When these characteristics hold the transition is said to be enabled to fire. An enabled transition can fire at any time and not necessarily immediately. Once a transition fires it removes w(pi,t) from each input place pi, and the output places, that is, the places that have arcs from the transition to the place, receive tokens according to the arc expression from the transition to the place: w(t,p). For a detailed explanation of colored Petri nets, consult Jensen (1992, 1997). Object-oriented Petri net (RPOO) models consist of a set of classes and their corresponding Petri nets. Classes are described and can be related to each other like UML classes. The Petri nets describe the internal behavior of objects. For each object, there is one Petri net that models it. A variety 989
Verifying Mobile Agent Design Patterns with RPOO
of RPOO actions (instantiate objects, call methods, destroy objects, etc.) that can be performed by the objects and actions are depicted by transition inscriptions in the colored Petri nets. An inscription may describe several actions, and all the actions in an inscription are executed atomically. A set of atomically executed actions is called an event. In RPOO, each object is a thread, and interaction between two objects may be asynchronous. This means that when an object a calls a method of (sends a message to) object b in asynchronous mode, the system moves to a state where the data passed as parameter will be pending and may be consumed in a further action by object b. RPOO actions also include synchronous calls, where messages are sent and consumed atomically. A set of interconnected objects and its pending messages form a structure of an object system. Besides this structure, an object system also knows what is called imminent actions, that is, actions that may be executed in the current structure. Briefly, an action will be executed, in case of concurrency. In RPOO, synchronously sent messages are presented by inscriptions like obj.set(data). In other words, the current object is sending a message to obj object. To send a synchronous message, the exclamation point obj!set(data) is used. To fire a transition when a message has been waiting, there must be an inscription like obj?set(data), denoting that a message is waiting. For a more detailed explanation of RPOO, consult Guerrero (2002).
FORMAL MODELS AND ANALYSIS OF PATTERNS Before modeling the patterns, it is necessary to identify all the entities that will be part of the whole model. And it is also necessary to build a diagram class containing these entities and their relationships to each other. Figure 4 shows the class diagram for this system. There are three types of agents, one for each one of the migration patterns presented in this article: ItineraryAgent, StarShapedAgent, and BranchingAgent. An agent has a list of agencies that represents its destination. At the beginning of the model, the current agent belongs to an agency, the source one. All the relationships change during the execution of the model. The class diagram represents the initial state of the scenario.
RPOO Models There is no tool for editing RPOO models, so the famous tool set Design/CPN (CPN Group, 1999) is used to edit and analyze such models. Each class is declared using the CPN/ML inscriptions (Christensen & Haagh, 1996). 990
Figure 4. Class diagram of design patterns
Figure 5 shows the model for the Itinerary class. In this model, we see the behavior of the class and its relationship with agent classes that implement agent interface and with SourceAgency class. Figure 6 shows the modeling of the DestinationAgency class. There is no complex behavior here; just the initialization of the agent is modeled. This initialization also represents the jobs to be executed locally. In Figure 7, the SourceAgency class is modeled. Its actions, based on its relationships with itinerary, network, and BranchingAgent, are presented. Figure 8 shows an abstraction of the network (e.g., Internet). It models the network class. Agents and agencies are transmitted through the network. Each agency receives its corresponding agent. Figure 9 shows the RPOO model for the ItineraryAgent class. It models the Itinerary migration pattern. The agent goes to every destination agency, it does its local job, and it returns to the source agency. The StarShapedAgent class is modeled by the RPOO net presented in the Figure 10. This model interleaves each destination agency with the source agency. With this, the agent always visits its source agency after it makes a job remotely. And Figure 11 presents the RPOO model for the BranchingAgent class. For each different destination agency, the model shown in Figure 7 makes a clone of this agent. Each one of these clones receives a list of destination agencies containing just one agency. Each one of the clone agents goes to the remote agency, does its job locally, and then returns with the results to the source agency.
Analysis For each one of the mobile migration patterns presented here, simulations were done using the classical CPN simulator
Verifying Mobile Agent Design Patterns with RPOO
Figure 5. RPOO model for Itinerary class
V
Figure 6. RPOO model for DestinationAgency class
Design/CPN. All the models behaved as expected. However, not all the possible behaviors were analyzed because of restrictions of existent tools. However, the RPOO models presented in this article showed how objected-oriented and mobile systems can be formally modeled and analyzed. To analyze all possible behavior, it is necessary to generate the state space of the formal model, also known as occurrence graph. With this graph and a model checker that supports RPOO model format, it will be possible to reason about some interesting properties described using some temporal logic.
system. The classical colored Petri nets presented in Lima (2004) and in Lima, Figueiredo, and Guerrero (2004) do not have this feature, which makes the mapping between the Petri net world and the programming language work more difficult. Due to this object-oriented feature, models can express the system behavior in a separate way. As it was said, each class in the class diagram is modeled in a unique RPOO model. This aspect makes all the modeled systems more organized and easier to be analyzed and studied.
BRIEF COMPARISON WITH CLASSICAL CPN MODELS
FUTURE TRENDS
The models presented in this article have an object-oriented feature. This feature makes the modeling task easier because the formalism used is semantically closer to the modeled
One future trend is to use RPOOt (Guerra, 2005; Guerra, Figueiredo, & Guerrero, 2005), or timed RPOO, to analyze the performance of each pattern presented here in some different contexts. With this, it will be possible to say more about them. 991
Verifying Mobile Agent Design Patterns with RPOO
Figure 7. RPOO model for SourceAgency class
Figure 8. RPOO model for Network class
Figure 9. RPOO model for ItineraryAgent class
992
Verifying Mobile Agent Design Patterns with RPOO
Figure 10. RPOO model for StarShapedAgent class
V
Figure 11. RPOO model for BranchingAgent class
Besides, only a few simulations were done with the models presented above. As is known, simulations do not cover all possible behaviors of the model. As mentioned before, to do this it is necessary to generate the occurrence graph. The state space generator for RPOO presented in Silva (2005)
will be used to obtain the occurrence graph. It will check for some of the properties described in CTL (Sifakis, 1990). To reason about these temporal properties, the model checker presented in Rodrigues (2004) will also be used.
993
Verifying Mobile Agent Design Patterns with RPOO
CONCLUSION As it can be seen, modeling a system with an object-oriented Petri net is not a trivial task. However, the benefits it brings to the development process is worth the effort. RPOO has an advantage over classical colored Petri nets: their semantics are closer to the programming language paradigm. But it is clear that the absence of a specific tool for editing RPOO models represents a large disadvantage for its usage for verifying and analyzing tasks. It is clear that RPOO is a good formalism to model mobile systems. Due to its object-oriented feature, the whole model is kept organized. Based on this experiment, RPOO showed to be more interesting in modeling such a mobile system than the classical colored Petri nets. But, a larger experiment is needed to show if this feature is general or if it was just in this specific context.
REFERENCES Christensen, S., & Haagh, T. B. (1996). Design/CPN overview of CPN ml syntax. CPN Group (1999). Design/CPN 4.0. University of Aarhus, Denmark. Retrieved from http://www.daimi.au.dk/ designCPN/ Guedes, F. P., Machado, P. D. L., & Medeiros, V. N. (2003). Developing mobile agent based applications. La Paz: Latin American Center of Studies in Computer Science. Guerra, F. V. de A. (2005). Modelagem de sistemas com restrições temporais em Redes de Petri orientadas a objetos. Master thesis, Curso de Pós-Graduação em Informática, Universidade Federal de Campina Grande, Brasil.
Klein, C., Rausch, A., Sihling, M., & Wen, Z. (2001). Extension of the Unified Modeling Language for mobile agents. In K. Siau & T. Halpin (Eds.), Unified Modeling Language: Systems analysis, design and development issues (pp. 116128). Hershey, PA: Idea Group Publishing. Lima, E. F. de A. (2004). Formalização e análise de padrões de projeto para agentes móveis. Master’s Thesis, Curso de Pós-Graduação em Informática, Universidade Federal de Campina Grande, Brasil. Lima, E. F. de A., Figueiredo, J. C. A. de, & Guerrero, D. S. (2004). Using coloured Petri nets to compare mobile agent design patterns. Electronic Notes on Theorical Computer Science, (95), 287-305. Medcraft, P. S. (2003). Integração de bancos de dados federados na Web usando agentes movies. Master’s Thesis, Universidade Federal de Campina Grande, Brasil. Rodrigues, C. L. (2004). Verificação de modelos RPOO. Master thesis, Curso de Pós-Graduação em Informática, Universidade Federal de Campina Grande, Brasil. Sifakis, J. (Ed.). (1990). Proceedings of the International Workshop on Automatic Verification Methods for Finite State Systems. Berlin: Springer-Verlag (LNCS 407). Silva, T. M. de (2005). Simulação automática e geração de espaço de estados de modelos em Redes de Petri orientadas a objetos. Master thesis, Curso de Pós-Graduação em Informática, Universidade Federal de Campina Grande, Brasil. Tahara, Y., Ohsuga, A., & Honiden, S. (1999). Agent system development method based on agent patterns. Proceedings of the 21st International Conference on Software Engineering (pp. 356-367). Los Angeles, CA.
Guerra, F. V. de A., Figueiredo, J. C. A. de, & Guerrero, D.S. (2005). Protocol performance analysis using a timed extension for an object oriented Petri net language. Electronic Notes on Theorical Computer Science, (130), 187-209.
KEY TERMS
Guerrero, D. D. S. (2002). Redes de Petri orientadas a objetos. PhD Thesis, Curso de Pós-Graduação em Engenharia Elétrica, Universidade Federal da Paraíba–Campus II, Brasil.
Computation Tree Logic (CTL): A type of temporal logic in which different futures are considered.
Jensen, K. (1992). Coloured Petri nets: Basic concepts, analysis, methods and practical use. Berlin: Springer-Verlag. Jensen, K. (1997). Coloured Petri nets: Basic concepts, analysis methods and practical use (vol. 2). Berlin: SpringerVerlag.
994
Agent Migration Pattern: A solution for a given problem inside the context of mobile agents.
Mobile Agent: A piece of computer software that is able to migrate (move) from one computer to another autonomously and continue its execution on the destination computer. Object-Oriented Petri Net: A formalism that puts together the Petri net formalism and the object-oriented paradigm. In Portuguese: Redes de Petri Orientadas a Objetos (RPOO).
Verifying Mobile Agent Design Patterns with RPOO
Occurrence Graph: A graph in which each node represents a single possible state of the model.
RPOOt (Redes de Petri Orientadas a Objetos temporizadas): A timed Object-Oriented Petri Net.
RPOO (Redes de Petri Orientadas a Objetos): See Object-Oriented Petri Net.
Temporal Logic: Term used to describe any system of rules and symbolism for representing, and reasoning about, propositions qualified in terms of time.
995
V
996 Category: Mobile Software Engineering
Virtualization and Mobility in Client and Server Environments Eduardo Correia Christchurch Polytechnic Institute of Technology, New Zealand
INTRODUCTION A great deal of popular software is not designed for mobility (Griffiths, 2004). This is peculiar because many mobile users expect to have easy access to an information infrastructure that links up their mobile phone, laptop, personal digital assistant (PDA), and other devices, while the backend systems of organizations need to be agile, especially as the number, range, and diversity of services and associated technologies grow. Enter virtualization, a technology that has been part of computing for many years, but only fairly recently become mainstream (Intel, 2006; Singh, 2004). It makes use of a virtual machine monitor (VMM), a mechanism that frees up systems from many of the physical constraints of hardware, by adding a software layer that abstracts hardware, so that an entire machine, operating system, applications, and even data can be stored as a set of standard folders and files. While it is well established that this architecture enhances security and reliability (Rosenblum & Garfinkal, 2004), it also enables both users and systems, as this article shows, to be mobile and responsive to change, both in client and server environments. VMware Workstation, Microsoft Virtual PC, and other virtualization software takes the form of a standard application that can be installed on physical computers. As Figure 1 shows, these applications provide a VMM, which Figure 1. Virtualization architecture App 1
App 2
App 1
App 2
Guest Operating System 1 (Virtual Machine )
Guest Operating System 1 (Virtual Machine)
Virtual Hardware
Virtual Hardware
App 1 App 2
Virtual Machine Monitor
Host Operating System (Physical Machine ) Physical Hardware
enables one system (the guest) to run within the context of another system (the host). The VMM presents a complete set of virtual hardware to each guest virtual machine (VM) running within this environment. Just as ordinary computers access physical resources, such as memory, processors, hard disks, and network adapters, so too do each of the virtual or guest systems, only their hardware is an instance of a generic abstraction that the VMM generates for each of them. The VMM then mediates various calls made by VMs to access the physical hardware of the host machine. Whereas the standard computer has a single operating system with applications installed to it, a computer with virtualization software runs, in addition, within the VMM one or more operating systems, each with their own applications installed.
MOBILE VIRTUAL MACHINES While virtual private networks enable users to work from remote locations, as if they are sitting at a machine on the local network to some extent, this approach has certain drawbacks. Users may wish to connect with machines that do not belong to the organization and therefore do not adhere to its policies and standards. Antivirus software may be out of date or updates not installed, for example. One solution is to provide a quarantine area that will allow a machine into the network, but restrict its access to resources until certain criteria have been met. Cisco Systems’ Network Admission Control (NAC) and Microsoft’s Network Access Protection (NAP) are examples of this kind of solution (Conry-Murray, 2005). Alternatively, network administrators can make use of the VMware Assured Computing Environment (ACE) to produce and deploy secure, fully built virtual machines that apply custom policies and adhere to certain specified standards (Burt, 2004). The fact that it is in effect the VM that forms part of the network and not the physical machine means that it does not matter to the network administrators that these particular hosts may not have the latest antivirus signature files or applied recent updates, as this underlying (physical) system does not interact with the network and cannot influence it in any way. Naturally, the user’s physical machine could fail causing the VM itself to fail, but this will still not affect the network, and restoring the client machine is simply a matter
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Virtualization and Mobility in Client and Server Environments
of copying the ACE VM from removable media, such as DVD or a pen drive (VMware, 2006a). In this way anyone needing temporary access to the network can be given it because perhaps the only way of accessing it using their own machine is through a VM compliant with the standards set by the organization, including exactly which resources can be accessed and the length of time the VM can be used.
SYSTEM AND SESSION MIGRATIONS When conventional systems need to be moved from one set of hardware to another, the operating system and server applications first need to be installed and then the data, assuming it is on the same system, moved, either by simply copying it or restoring it from a backup. It is not just that this is a time-consuming exercise, but also it often entails having to navigate the complexities of making a system work with significantly different hardware, and its associated drivers and other software. This can cause lengthy outages and make system migrations complex and arduous. Virtual systems, by contrast, always access the same set of physical resources the VMM presents to it, whatever the differences in the actual underlying physical hardware, making them much more mobile than conventional systems. In fact, moving the entire system is simply a matter of copying data from one (physical) machine to another, making VMs highly portable (Wolf & Halter, 2005, p. 481). With Vmotion, systems administrators can even move virtual machines without interrupting service availability (VMware, 2006b), something that has been used with success in live production environments (Rosenkoetter, 2006). This makes it feasible to apply many changes during business hours that would otherwise have been scheduled for weekends or at least when the network is quiet (Cline, 2006). In such cases virtualization enables easier migration of systems from one set of hardware to another, and significantly reduces the risk associated with such change. This risk is reduced further by the ease with which it is possible to store multiple older versions of virtual systems. Where for instance poor software or problematic devices are installed, virtualization enables administrators to return the machine easily to a selected previous state. This concept can be taken a step further by capturing and virtualizing a client’s entire session with an existing server, then migrating to another system that is the same or different, such as from a desktop machine to a PDA (Baratto, Potter, Su, & Nieh, 2004). In fact virtualization can even be used by mobile users to “decouple” a computer into a “body (display, CPU, RAM, I/0) and a soul (session state, software, data, preferences)” so that a user with a portable device can walk up to any computer and resume a session started on another machine (Cáceres, Carter, Narayanaswami, & Raghunath, 2005, p. 65).
CONCLUSION Virtualization has expanded dramatically in recent years because it is a flexible, scalable technology. It allows powerful hardware to be used more efficiently by distributing the processing, storage, and movement of data among several virtual systems that can still make use of clustering and other conventional forms of load balancing and redundancy. VMware ESX Server and Microsoft Virtual Server for instance make it cost effective to host a single major application per server, so reducing software conflicts and increasing reliability by isolating each of the guest systems, as well as the host from one another while enabling systems administrators to fine tune systems to resident applications. It also makes it easy to retain (copies of) entire systems either for the purposes of disaster recovery or to test future development, perhaps with a view to implementing such changes on production systems. Virtualization can be utilized in a range of situations in both client and server environments, be it as part of a mission-critical system users connect to; a disaster-recovery infrastructure; a gateway for connecting securely to the network; a deployment of secure, fully built clients that comply with specified standards; or a portable learning environment comprising an entire network of virtual machines that may or may not be connected to virtual and even physical switches, but which can be easily moved from one physical machine to another. It is true that virtualization often demands powerful hardware in order for it to cope with the demands of hosting numerous systems; but with the reduction in the cost of hardware, the use of virtual systems becomes a viable proposition for organizations, especially as they are easier to manage and more agile than conventional systems. It is no wonder then that manufacturers are beginning to produce hardware that is designed with virtualization in mind (Intel, 2006), and that VMware and Virtual PC have become such well-known brands in recent times. According to one survey, respondents expected 45% of new servers deployed this year to make use of virtual machines (IDC, 2005), a trend, it appears, that is set to continue.
REFERENCES Baratto, R. A., Potter, S., Su, G., & Nieh, J. (2004, September 26-October 1). MobiDesk: Mobile virtual desktop computing. Proceedings of the 10th Annual International Conference on Mobile Computing and Networking, Philadelphia. Burt, J. (2004, September 20). VMware takes virtual machines mobile. Retrieved April 5, 2006, from http://www. eweek.com/article2/0,1895,1647632,00.asp
997
V
Virtualization and Mobility in Client and Server Environments
Cáceres, R., Carter, C., Narayanaswami, C., & Raghunath, M. (2005, June 6-8). Reincarnating PCs with portable SoulPads. Proceedings of the 3rd International Conference on Mobile Systems, Applications, and Services, Applications, and Services 9pp. 65-78), Seattle, WA. Conry-Murray, A. (2005). Caymas Systems’access gateways. Retrieved May 10, 2006, from http://www.itarchitect.com/ shared/printableArticle.jhtml?articleID=171000823 Cline, K. (2006, March 1). Re: Interesting comment from MS tech specialist about VS & ESX. Posted on VMTN Discussion Forums; retrieved from http://www.vmware.com/community/thread.jspa?threadID=33012&start=30&tstart=0 Griffiths, G. (2004). Help for the mobile worker: Keeping your remote workforce productive and secure. Retrieved May 7, 2006, from http://www.everdream.com/wp/mobile_workforce_whitepaper.asp IDC. (2005). Increasing the load: Virtualization moves beyond proof of concept in the volume server market, according to IDC. Retrieved May 7, 2006, from http://www. idc.com/getdoc.jsp?containerId=prUS00259905 Intel. (2006). Intel virtualization technology: Hardwareassisted virtualization for today’s businesses. Retrieved May 7, 2006, from http://www.intel.com/products/processor/xeon/vt_prodbrief.pdf Rosenblum, R., & Garfinkal, T. (2005). Virtual machine monitors: Current technology and future trends. Computer, 38(5), 39-47. Rosenkoetter, R. (2006, March 1). Re: Interesting comment from MS tech specialist about VS & ESX. Posted on VMTN Discussion Forums; retrieved from http://www.vmware. com/community/thread.jspa?threadID=33012&start=30& tstart=0 Singh, A. (2004). An introduction to virtualization. Retrieved May 9, 2006, from http://www.kernelthread.com/publications/virtualization/ VMware. (2006a). VMware ACE facilitates and streamlines remote working at Siemens Industrial Turbomachinery. Retrieved April 5, 2006, from http://www.vmware.com/customers/stories/siemens_ace.html
998
VMware. (2006b). VMware VMotion: Delivering game changing virtual machine mobility. Retrieved May 9, 2006, from http://www.vmware.com/products/vc/vmotion.html Wolf, C., & Halter, E.M. (2005). Virtualization: From the desktop to the enterprise. Berkeley, CA: Apress.
KEY TERMS Application: A program that performs a specific task or related group of tasks and which requires an operating system to run successfully. Guest: A virtual machine that runs in the context of a virtual machine monitor, which provides for it an abstracted hardware environment. See Virtual Machine. Host: A node on a network that requires an IP address to communicate with other hosts, but in this article refers specifically to a physical computer that runs a virtual machine monitor capable of running one or more virtual machines. Migration: The movement of operating systems, applications, and data from one computer to another; it can now also include the transfer of a session from one device to another. Operating System: A general-purpose program that performs many basic tasks such as accepting input from the keyboard or printing output to a screen, as well as mediating access to software and hardware resources. Virtualization: Partitioning a physical machine into a number of virtual machines, each of which effectively functions the same as a physical machine and can replace a physical machine on a network. Virtual Machine: A machine much like a physical machine, in that it requires an operating system, can have applications installed, can present itself in exactly the same way as a physical machine on the network, but which functions within the specific abstracted hardware environment of a virtual machine monitor. Virtual Machine Monitor: An environment created by an application, such as VMware or Virtual PC, with the purpose of enabling one or more virtual machines to run.
Category: Mobile Phone 999
Voice Recognition Intelligent Agents Technology Călin Gurău Montpellier Business School, France
INTRODUCTION Mobile computing technology is evolving at a rapid pace. Under the pressure of market demands, the format of mobile devices evolves towards a contradictory situation: on one hand, the handset tends to become smaller, but on the other hand, the users demand increased data search, transmission, and saving capabilities. In order to achieve this, the model of interaction between humans and mobile devices has to evolve from the presently prevalent keyboard-screen system for data input and output towards voice-recognition intelligent agents technology. This article attempts to present the rationale and the advantages of this development, and to analyze the possible problems raised by the introduction of this technology. The article starts with a presentation of the existing mobile phone technology, outlining its main limitations in terms of functionality, which are logically determined by the way and the context in which mobile phones are normally used. Based on the analysis of the contradictions between the present model of interaction with mobile phones and the requirements of users, the article presents possible solutions to this problem. The study argues that the introduction of voice recognition intelligent agents can enhance significantly the functionality of mobile phones, representing a true revolution in mobile computing. Various practical applications of this technology are briefly presented, as well as the main problems related with its development and implementation. The article ends with a summary of the arguments discussed and with definitions of the main terms and concepts presented.
BACKGROUND Mobile phone technology was developed with the aim to provide users with a telephone connection anyplace, anytime. The main innovation that allowed the mass adoption and use of mobile phones was the cellular approach in transmitting a radio signal. Traditionally, people that required frequent communications could install in their car a radio telephone, but the small number of radio channels available in one area limited drastically the number of possible users of this technology. By dividing a large area into small cells,
and each of these cells having a low-power transmitter, the number of communication channels increases significantly, since people that are not located in neighboring cells can use the same frequency to communicate (Layton, Brain, & Tyson, 2005). The introduction of digital technology (2G) has increased even further the number of communication channels. Finally, 3G technology represents the latest trend in mobile phones standards, offering increased bandwidth and information transfer rates to accommodate Web-based applications and phone-based audio and video files. However, the use of mobile phones to access Internet applications presents a number of limitations, some of which are related with the specific interface of mobile phones, and others with the existing Web protocols adapted for mobile networks. The screens of mobile phones are small and have a lower resolution in comparison with PC or laptop screens/monitors. On the other hand, the wireless application protocol (WAP) works badly on wireless devices with small screens, and it is dependent on mobile technology’s bandwidth (such as GSM or CDMA) for access to information and services (Yeo, & Huang, 2003). Other problems are connected with the Web navigation and site structure, or with the input methods available for mobile phone users (Buchanan et al., 2001). The future development of mobile services requires a revolution in the technological model applied, which is based on transforming the Web architecture, as well as the usage of mobile devices.
WIRELESS WEB APPLICATIONS FOR MOBILE PHONES At the moment, the Internet is a network of databases supported by applications that allow users to search, retrieve, and use information contained in computers’ memory. However, the rapid increase of online available data makes it more and more difficult for users to find the specific information they need. A trivial search on the Google search engine usually displays a list of a few million more or less relevant Web documents. In this case, the use of a search engine is only the beginning and not the end of searching for particular data,
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
V
Voice Recognition Intelligent Agents Technology
and this process can indeed be very time consuming, without providing any guarantees for a successful result. An alternative is to use customized intelligent agents, which can search the Web for clearly defined data. These intelligent agents are usually adapted for a specific type of Web search—for example, they can provide a list of companies that offer online a specific type of product. Strictly speaking, they are not very intelligent, because traditionally, these applications were not able to learn and improve in time their searching capabilities. However, this can be changed. Using neural network technology, and registering a history of operations realized or a particular user, the advanced intelligent agents can progressively learn the preferences of their customers and provide improved results. Wireless devices such as mobile phones have a number of limitations determined by their specific circumstances of usage. The main advantage of a mobile phone is obviously its mobility, which implies a small size and weight, combined with good usability, in various environment and circumstances. These characteristics limit the size and the resolution of the screen, the size, and the functionality of keypads; the power and memory capacity; as well as the bandwidth. Because of these problems, the mobile phones cannot be used like a PC device, which incorporates autonomous computing capabilities. However, the model of distributed networks and resources can be effectively applied to a mobile phone (Mattern, 2000). In a traditional sense, mobile phones are simple communicating devices, more like interface terminals than personal computers. The solution for their effective use is to create easily accessible networks with distributed resources, and to develop advanced software
applications in order to improve the mobile communication capability (Alesso, & Smith, 2001). Some of these applications that can significantly improve the interface between the user and the distributed network are based on voice-recognition technology. On the other hand, the nature of mobile devices is adapted to a model of discontinuous, time-limited use; therefore, the user does not have the time himself/herself to browse the Internet in search of relevant information. Even given the required time and stability, the user might prefer the use of an Internet-connected PC or terminal, which offers better interaction and visualization capabilities. This problem can be solved by developing intelligent applications that can work automatically and independently of any human supervision, using the instructions given by the user. The solution of improved Web services using mobile devices is the combination of voice-recognition technology with the use of intelligent agents (Lai, Mitchell, Viveros, Wood, & Lee, 2002). The mobile phone user will initiate a command by directly speaking with one or more intelligent agents, which then can search for specific information on the Web and announce the results through an SMS message. This type of interaction is presented in Figure 1. The intelligent agents pass through a succession of phases in order to fulfill their tasks (Rodríguez, Favela, & Muñoz, 2003): Activated: This represents the main state of the agent, which includes a number of specific sub-states: • Learning: Represents the initial sub-state of an agent, in which the agent acquires knowledge
•
Figure 1. The use of voice recognition intelligent agents in mobile phones services
Internet
Mobile phone user select and instruct send an SMS
Intelligent agent
Database
Intelligent agent
Database
Intelligent agent
search and compare
Result
1000
Database
Voice Recognition Intelligent Agents Technology
Figure 2. The use of voice recognition intelligent agents and mobile agents in mobile phones services Mobile phone user instruct send an SMS
Intelligent agent select and instruct
V
Internet
Mobile agent
Database
Mobile agent
Database
Mobile agent
search and compare
Database
Result
• • •
•
about its environment in various ways (through direct instructions from the user, a service, or other agents; from a sensor connected with the working environment; or through autonomous learning). Analyzing: On the basis of the data obtained in the previous sub-state, the agent establishes its goal and the method of fulfilling it. Executing: The agent performs the action plan established in the previous sub-state. Communicating: The agent can communicate with the user or with other agents, in order to collect additional information or to presents the results of its task. Suspended: While it waits for information, the agent can suspend its activity.
However, this model has two major problems: •
•
For the moment, the intelligent agents active on the Web are strictly specialized for specific tasks. The search for a specific agent can be cumbersome for the mobile phone user in terms of time and interface. In fact, in this case, the search for online information is replaced by the search for the best intelligent agent that can perform the automatic search. The size of software applications that correspond to Web intelligent agents is quite significant. If for the classical online navigation this is not a problem, the use of these applications in a mobile phone network—where bandwidth is still limited—can reduce the speed of the entire operation.
A model developed by Kowalczyk et al. (2003) provides a possible solution to these problems. The model includes the use of two different types of agents (see Figure 2): •
•
Intelligent Agents: These are personalized software applications that manage the information search needs of a particular mobile phone user; they are stationary agents that are not able to migrate to other platforms. Mobile Agents: These are of a smaller size, specialized in specific tasks, active on the Web, and have low demands on network connection, quality, and online time, because after their migration to a new platform or device, the online connection must not be maintained while they are working locally.
This solution is particularly adapted to a distributed network architecture, which represents the most suitable infrastructure for mobile devices; for example, the personalized intelligent agent can be located on the home or office-based PC of the user, the mobile phone representing just a communication device between various elements of this process (Lino, Tate, Siebra, & Chen-Burger, 2003). Despite its advantages, the large-scale use of voicerecognition agents for mobile phones is likely to develop a number of problems (Wong & Starner, 2001): 1.
Unreliable Voice Recognition: The voice recognition techniques that are adequate for an office environment might not be feasible in mobile environments, where the voice power is likely to vary and the ambient noises
1001
Voice Recognition Intelligent Agents Technology
2.
3.
4.
might deteriorate the quality of sounds (D’Agostino, 2005). Privacy: The use of voice-recognition agents on mobile devices raises the problem of privacy of the user, since s/he is transmitting the instructions to the intelligent agent with a loud, clear voice. Cognitive Load and Attention: The capacity of the human brain to coordinate multiple tasks is limited, which has to be taken into account considering that often, the users of mobile devices are also engaged in other activities. Security of Mobile Intelligent Agents: This is noted by van Eijk, Hamers, Klos, and Bargh (2002) and can be found in various environments.
The possible applications of intelligent agents in mobile and pervasive computing have determined the development of various platforms for developing agent and multi-agent systems. On the other hand, the problems raised by platform compatibility have encouraged the standardization of different systems. The result of these efforts was concretized in two internationally recognized standards (van Eijk et al., 2002): the Foundation for Intelligent Physical Agents (FIPA) and the Object Management Group (OMG). One of the best agent platforms for mobile devices is the JADE/LEAP platform, which was the result of the integration between a project funded by an EU grant for the development of an agent platform implemented in the JAVA software language (JADE), and a project called the Lightweight and Extensible Agent Platform, initiated by a consortium consisting of Motorola, ADAC, BT, Broadcom Eireann Research, Telecom Italia Lab, University of Parma, and Siemens.
FUTURE TRENDS These models of mobile phone Web services are still in the development stage. However, the missing elements of the process are gradually developed and integrated in systems with multiple applications. At the basis of this model is the concept of the Semantic Web. The Semantic Web is an extension of the current Web in which information is given well-defined meaning, better enabling computers and people to work in cooperation. It is based on the idea of having data on the Web defined and linked such that it can be used for more effective discovery, automation, integration, and reuse across various applications (Hendler, Berners-Lee, & Miller, 2002). Semantic Web services can be used, among others, for intelligent mobile construction collaborations (Aziz, Anumba, Ruikar, Carrillo, & Bouchlaghem, 2004), participation in Web site auctions (Kowalczyk et al., 2003), and banking services or real estate transactions (Clareity, 2004).
1002
CONCLUSION This article has attempted to provide a general overview of the main problems in mobile communication and computing, and has presented a possible solution through the application of voice-recognition intelligent agents to small mobile devices. The future of mobile intelligent agents based on voicerecognition technology is full of opportunities. Various public and private organizations are already developing the various elements of this system. However, the successful implementation of these applications has to take into account the following issues: 1.
2.
3. 4.
The use of mobile intelligent agents requires a completely restructured Internet system, which defines and links various information and databases using the semantic model. The applicability of intelligent agents to small mobile devices depends on the implementation of pervasive computing environments, with networks of distributed resources in terms of memory, databases, services, and applications. The capacity of mobile intelligent agents to migrate among various platform raises issues of compatibility and security. The use of voice-recognition technology also creates challenges at social, ethical, and moral levels (e.g., the use of voice-recognition systems in social environments or personal privacy issues).
These problems require an answer that can integrate the informational, technological, social, and personal levels before the mobile intelligent agents and voice-recognition technology will become widespread commercial applications.
REFERENCES Alesso, H. P., & Smith, C. F. (2001). The intelligent wireless Web. Boston: Addison-Wesley Professional. Aziz, Z., Anumba, C., Ruikar, D., Carrillo, P., & Bouchlaghem, D. (2004) Semantic Web based services for intelligent mobile construction collaboration. ITCon, 9, 367-379. Buchanan, G., Farrant, S., Jones, M., Thimbleby, H., Marsden, G., & Pazzani, M. (2001). Improving mobile Internet usability. Retrieved December 2005 from http://www10. org/cdrom/papers/230/ Clareity. (2004). The perfect computer interface for the real estate industry. Retrieved December 2005 from http://www. callclareity.com/2004-VR.cfm
Voice Recognition Intelligent Agents Technology
D’Agostino, D. (2005). Weak speech recognition leaves customers cold. CIO Insight, (December 29). Retrieved January 2006 from http://www.cioinsight.com/article2/0,1540,1905930,00.asp Layton, J., Brain, M., & Tyson, J. (2005). How cell phones work. Retrieved December 2005 from http://electronics. howstuffworks.com/cell-phone.htm Hendler, J., Berners-Lee, T., & Miller, E. (2002). Integrating applications on the Semantic Web. Institute of Electrical Engineers of Japan, 122(10), 676-680. Kowalczyk, R., Braun, P., Mueller, I., Rossak, W., Franczyk, B., & Speck, A. (2003). Deploying mobile and intelligent agents in interconnected e-marketplaces. Journal of Integrated Design and Process Science, 7(3), 109-123. Lai, J., Mitchell, S., Viveros, M., Wood, D., & Lee, K. M. (2002). Ubiquitous access to unified messaging: A study of usability and the use of pervasive computing. International Journal of Human-Computer Interaction, 14(3/4), 385-404. Lino, N. Q., Tate, A., Siebra, C., & Chen-Burger, Y.-H. (2003, August 11). Delivering intelligent planning information to mobile devices users in collaborative environments. Proceedings of the 18th Joint Conference on Artificial Intelligence, Acapulco, Mexico. Retrieved December 2005 from http://www.dimi.uniud.it/workshop/ai2ia/cameraready/queiroz.pdf Mattern, F. (2000). State of the art and future trends in distributed systems and ubiquitous computing. Retrieved December 2005 from http://www.vs.inf.ethz.ch/publ/papers/DisSysUbiCompReport.pdf Rodríguez, M., Favela, J., & Muñoz, M. A. (2003, August 11). Providing opportunistic access to information sources and services for mobile users. Proceedings of the 18th Joint Conference on Artificial Intelligence, Acapulco, Mexico. Retrieved December 2005 from http://www.dimi.uniud. it/workshop/ai2ia/cameraready/rodriguez.pdf van Eijk, R., Hamers, J., Klos, T., & Bargh, M. S. (2002). Agent technology for designing personalized mobile service brokerage. Retrieved November 2005 from http://www. recursionsw.com/Mobile_Agent_Papers/Gigamobile.pdf Wong, B. A., & Starner, T. E. (2001). Conversational speech recognition for creating intelligent agents on wearables. Retrieved November 2005 from http://www.cc.gatech.edu/ fac/Thad.Starner/p/030_20_DPP/conversational-speechrecognition.pdf
KEY TERMS Bandwidth: The amount of data that can be transmitted in a fixed amount of time. For digital devices, the bandwidth is usually expressed in bits per second or bytes per second. For analog devices, the bandwidth is expressed in cycles per second, or Hertz (Hz). Code-Division Multiple Access (CDMA): A digital cellular technology that uses spread-spectrum techniques. Unlike competing systems, such as GSM, CDMA does not assign a specific frequency to each user; instead, every channel uses the full available spectrum. Distributed Network: A network structure in which the network resources, such as switching equipment and processors, are distributed throughout the geographical area being served. Global System for Mobile Communications (GSM): One of the leading digital cellular systems. GSM uses narrowband Time Division Multiple Access technology, which allows eight simultaneous calls on the same radio frequency. Neural Network: An interconnected collection of simple processing elements, units, or nodes whose functionality is loosely based on the animal brain. The processing ability of the network is stored in the inter-unit connection strengths, or weights, obtained by a process of adaptation to, or learning from, a set of training patterns. Neural nets are used in bioinformatics to map data and make predictions. 3G: Third generation of mobile communications technology. 3G provides increased bandwidth up to 384 Kbps when a device is stationary or moving at pedestrian speed, 128 Kbps in a car, and 2 Mbps in fixed applications. 3G works over wireless air interfaces such as GSM or CDMA. 2G: Second-generation wireless service, also known as personal communications services. Refers to digital voice cell phone systems deployed in the 1990s. Delivers both voice and data transmissions using switched technology where each call requires its own cell channel. Wireless Application Protocol (WAP): A standard protocol for providing cellular phones, pagers, and other handheld devices with secure access to e-mail and textbased Web pages. WAP provides a complete environment for wireless applications that includes a wireless counterpart of TCP/IP and a framework for telephony integration such as call control and phonebook access.
Yeo, J., & Huang, W. (2003). Mobile e-commerce outlook. International Journal of Information Technology & Decision Making, 2(2), 313-332. 1003
V
1004 Category: M-Health
The Wi-INET Model for Achieving M-Health Success Nilmini Wickramasinghe Illinois Institute of Technology, USA Steve Goldberg INET International Inc., Canada
INTRODUCTION Medical science has made revolutionary changes in the past decades. Contemporaneously however, healthcare has made incremental changes at best. The growing discrepancy between the revolutionary changes in medicine and the minimal changes in healthcare processes is leading to inefficient and ineffective healthcare delivery—one, if not the, significant contributor to the exponentially increasing costs plaguing healthcare globally. Healthcare organizations can respond to these challenges by focusing on three key solution strategies: access, quality, and value. These three components are interconnected such that they continually impact on the other, and all are necessary to meet the key challenges facing healthcare organizations today. The application of mobile commerce to healthcare— namely, m-health—appears to offer a way for healthcare delivery to revolutionize itself and simultaneously address the critical areas of access, quality, and value. Integral to such an approach is the need for a robust wireless model. We propose the Wi-INET (wireless Internet, intranet, extranet) model as the way to deliver m-health excellence.
BACKGROUND Currently the healthcare industry in the United States as well as globally is contending with relentless pressures to lower costs while maintaining and increasing the quality of service in a challenging environment. It is useful to think of the major challenges facing today’s healthcare organizations in terms of the categories of demographics, technology, and finance. Demographic challenges are reflected by longer life expectancy and an aging population; technology challenges include incorporating advances that keep people younger and healthier; and finance challenges are exacerbated by the escalating costs of treating everyone with the latest technologies. Healthcare organizations can respond to these challenges by focusing on three key solution strategies: (1) access—caring for anyone, anytime, anywhere; (2) quality—offering world-class care and establishing integrated
information repositories; and (3) value—providing effective and efficient healthcare delivery. These three components are interconnected such that they continually impact on the other and all are necessary to meet the key challenges facing healthcare organizations today. In short then, the healthcare industry is finding itself in a state of turbulence and flux (National Coalition on Healthcare, 2004; Pallarito, 1996; European Institute of Medicine, 2003; WHO, 2000, 2004; Wickramasinghe & Silvers, 2003). Such an environment, is definitely well suited for a paradigm shift with respect to healthcare delivery (von Lubitz & Wickramasinghe, 2005). Many experts within the healthcare field area agree that mhealth appears to offer solutions for healthcare delivery and management that serve to maximize the value proposition for healthcare. However, to date, little if anything has been written regarding how to achieve excellence in m-health, nor does there exist any useful model for framing m-health delivery.
MAIN THRUST: INTEGRATIVE MODEL FOR M-HEALTH Successful m-health projects require a consideration of many components. Figure 1 provides an integrative model for all key factors that we have identified through our research that are necessary in order to achieve m-health excellence (Wickramasinghe et al., 2005; Goldberg et al., 2002a, 2002b, 2002c, 2002d, 2002e; Wickramasinghe & Goldberg, 2004). What makes this model unique and most beneficial is its focus on enabling and supporting all areas necessary for the actualization of information and communication technology initiatives in healthcare. By design, the model identifies the inputs necessary to bring an innovative chronic disease management solution to market. These solutions are developed and implemented through a physician-led mobile e-health project. This project is the heart of the model to bridge the needs and requirements of many different players into a final (output) deliverable, a “Wireless Healthcare Program.” To accomplish this, the model is continually updated to identify, select, and prioritize the ICT project inputs that will:
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
The Wi-INET Model for Achieving M-Health Success
•
• •
Accelerate healthcare system enhancements and achieve rapid healthcare benefits. The model identifies the key healthcare system inputs with the four Ps: people that deliver healthcare, process to define the current healthcare delivery tasks, platform used in the healthcare technology infrastructure, and protection of patient data. Close the timing gaps between information research studies and its application in healthcare operational settings. Shorten the time cycle to fund an ICT project and receiving a return on the investment.
IT Architecture and Standard Mobile Environment
data is stored within this layer, thereby ensuring compliance with international security standards/policies like HIPAA. Tier-2, shown as the HTTP Server, provides the business logic including but not limited to lab, radiology, and clinical transcription applications; messaging of HL7, XML, DICOM, and other data protocols; and interface engines to a hospital information system (HIS), lab information systems (LIS), radiology information systems (RIS), as well as external messaging systems such as Smart Systems for Health (an Ontario healthcare IT infrastructure project). This latter messaging feature may also be included in the third tier, which consists of the back-end database servers like Oracle, MySQL, or Sybase.
Mapping Case Study to Model
By adopting a mobile/wireless healthcare delivery solution, it is possible to achieve rapid healthcare delivery improvements, which impact both the costs and the quality of healthcare delivery. This is achieved by using an e-business acceleration project which provides hospitals a way to achieve desired results within a standardized mobile Internet (wireless) environment. Integral to such an accelerated project is the ability to build on the existing infrastructure of the hospital. This then leads to what we call the three-tier Web-based architecture (see Figure 2). In such an environment, Tier-1 is essentially the presentation layer; which contains the Web browser, but no patient
During the past six years, INET has used an e-business acceleration project to increase information and communication technology (ICT) project successes (Goldberg et al., 2002a, 2002b, 2002c, 2002d, 2002e). Today INET is repurposing the e-business acceleration project into a mobile e-health project to apply, enhance, and validate the mobile e-health project delivery model. Such a model provides a robust structure, and in turn serves to ensure excellence in the m-health initiative. INETs data provides the perfect opportunity to examine the components of our model (Figure 1), as it is both rich and longitudinal in nature. In mapping the data and specific business case, we have drawn upon many well-recognized
Figure 1. Wi-INET business model Medical Infornatics
People
Business Management
E-Health
Information & Comm.. Tech
Patient
Scope
Research Findings
Community/Home Care Primary Care Acute Care
Process
Medical Compliance
Mobile E-Health Project Actualize Physician-led Innovations In Chronic Disease Management
Administration Information Technology Wireless Devices
Platform
Transcoding 3- Tier App Architecture
Localize Wireless Healthcare Program
Field
Back-end Connections
Protection
Privacy
Evaluate
Security
Funding Criteria
Reliability
Pharmaceutical & Device Firms
Consumers
Government or 3rd Party Insurer
Employers
1005
W
The Wi-INET Model for Achieving M-Health Success
Figure 2. Three-tier Web-based architecture Three-Tier Architecture Internet/Intranet
Request
Data
Page
Browser
HTTP Server
qualitative techniques, including conducting both structured and unstructured interviews, in-depth archival analysis, and numerous site visits. Goldberg et al. (2002a, 2002b, 2002c, 2002d, 2002e) capture and substantiate the findings discussed, while Kavale (1996), Boyatzis (1998), and Einhardt (1989) detail the importance and richness of the methodologies we have adopted in presenting the following findings. Key criteria were established from the Standish Group International (1994) and the Institute of Medicine in America (2001). The INET Mobile E-Health project objectives include: 1.
2.
3.
4.
Accelerate consensus building with an e-health solution that is focused on a disease state and driven by the medical model, with the primary objective to streamline communications and information exchange between patients, and providers of community/home care, primary care, and acute care. Acquire commercial funding early with a compelling business case. For instance, enhancing therapeutic compliance can improve patient quality of life with significant healthcare cost savings. It is well documented that in diabetes, this will have immediate and high impact-benefits for healthcare consumers, pharmaceutical firms, governments, insurers, and employers. Avoid risk by reengineering large-scale healthcare delivery processes in small manageable pieces. Today, organizations can harness a rigorous method to incrementally enhance a process one step at time. This is a way to achieve quick wins early and frequently. Rapid development of simple-to-use, low-cost, and private/secure information and communication technology solutions. Achieve these benefits through a wireless application service provider (ASP). In addition to rapid development, a wireless ASP can easily connect and bring together many independent healthcare information systems and technology projects.
To actualize the mobile e-health project, INET is looking to the Wi-INET model as a framework. For INET, this will 1006
Database Server
support an INET Mobile E-Health Project Management Office (PMO) to manage the costs and quality, deliver many small projects, and replicate projects for local and international distribution. As a first case scenario for the model, INET is proposing an INET Wireless Diabetes Program, with leadership from a family physician. The INET PMO is provisioning a project manager to support this physician-led project to meet both research and commercial sponsors’ interests and objectives in diabetes. A detailed description of the key attributes of the INET Wireless Diabetes Program includes:
Problem Statement There are many communication and information exchange bottlenecks between patients and their family physicians that prevent the effective treatment of diabetes. As background, a fundamental problem today is the ability to have a private and secure way to manage, search, and retrieve information at the point-of-care. In diabetes, physicians cannot quickly and easily respond to patients with high glucose levels. They need to wait for people to come to the office, respond to phone calls, reply using traditional mail delivery, or never receive the patient information.
Solution Mandate Implement a diabetes monitoring program to enhance therapeutic compliance, such as release a program to enhance the usage of oral hypoglycemic agents (drugs) and/or the usage of blood sugar monitoring devices. As background, everyone wins when enhancing patients’ ability to follow instructions in taking prescribed medication. The patient’s health, safety, and quality of life improve with significant healthcare cost savings. However, it is well documented that many patients do not stay on treatments prescribed by physicians.1 This is where wireless technology may have the greatest impact to enhance compliance. One solution may be as simple as using a cell phone and installing a secure wireless application for patients to
The Wi-INET Model for Achieving M-Health Success
Table 1. INET wireless diabetes program results INET Wireless Diabetes Program Patient
Change in HAIC Levels
% reduction in HAIC
Pre-Pilot
Post-Pilot
1
0.082
0.069
-16%
2
0.090
0.071
-21%
3
0.108
0.050
-54%
4
0.113
0.084
-26%
monitor glucose levels, and provisioning a physician to use a PDA (connected to a wireless network) to confidentially access, evaluate, and act on the patient’s data.
Business Case In Ontario the cost savings may represent almost $1 billion over three years. INET uses a simple calculation to determine the $1 billion savings. This can be found at www.inet-international.com; please select the INET mobile e-health project section to review the calculations. The business case can be backed with additional data on how the cost of prevention (drugs) is far less than the cost savings associated with reducing the risk of complications associated with diabetes. For instance, the impact of a 1% decrease in A1C is significant. More data is available to support the business case for the prevention of type 2 diabetes such as lowering the incidence of End Stage Renal Disease (ESRD). In summary, there is plenty of data today to quickly build consensus, fund, and implement a national and international wireless diabetes program to enhance patients’ quality of life with significant healthcare cost reductions—that is, meet the objectives of access, quality, and value.
Systems Development Lifecycle Project Delivery Use an INET mobile e-health project to scope, localize, field, and evaluate an INET wireless diabetes program led by a physician. Each project can easily and simply customize a program to quickly meet the unique needs of a rural and urban healthcare delivery setting, age, ethnicity, income, language, and culture. These are small manageable projects. Each project collects data on patient/healthcare provider relationships, wireless medical informatics, therapeutic compliance business case, and ICT usability to accelerate acceptance of a wireless diabetes program using wireless technology. The program may include cellular network and application usage, support, healthcare provider PDA, and
consulting fees for a family physician and other healthcare providers. However, it is expected that the costs may not include items such as consumer cell phone, medication, or blood sugar monitoring devices/supplies. It is recommended that commercial and/or research sponsors pay for an INET project and help subsidize the user costs. In June 2005, INET applied the Wi-INET model to pilot a wireless diabetes program with the objective to decrease diabetes-related complications with better control of glycemic levels, measured by HA1C. The core component of the program is the relationship between family physicians and patients supported by a wireless diabetes management protocol.2 This protocol describes how a patient can enter his or her glucose readings into a cell phone and transmit the results to his or her family physician. The protocol further details how the physician, in turn, is able to monitor any number of patients on his or her PDA, such as a Palm Treo or RIM Blackberry device. A physician, if required, can take immediate action with a message electronically sent to the patient’s cell phone. The program was tested through a pilot project with four patients and led by Dr. Sheldon Silver, and was completed in July 2005. The pilot project lasted approximately three months. The preliminary results are significant as shown in Table 1. In summary, INETs research data indicates that using the Wi-INET model will increase ICT project success in healthcare. To realize and test this, INET continues to map the player’s from an INET wireless diabetes program (use case scenario) to the model. To show how this works, please review the mapping exercise below. The bold text in black is a project player and the color text in [] parenthesis relates to the sections of the model presented in Figure 1. Physician Mobile E-Health Project Lead: [“Mobile E-Health Project” in Figure 1]. Physicians provide the linkage to the medical model to enhance disease management programs to enhance patient care and safety, improve research and education, increase healthcare quality, and reduce healthcare costs. For INET’s use case scenario, the final outcome is a Wireless Diabetes Program [“Wireless Healthcare Program” in Figure 1]. And the mobile e-health projects are led by Dr. Sheldon Silver, MD, Staff Physician, Credit Valley Hospital. Commercial Sponsor(s) [“Funding Criteria” in Figure 1]. The project delivers information and communication solutions for: • • •
Consumers wishing to improve their quality of life with an enhanced relationship with their healthcare provider’s—that is, family physicians. Pharmaceutical firms looking to increase revenues with e-compliance programs. Government/insurers investigating ways to significantly reduce administration and healthcare costs, and shorten healthcare delivery time cycles (wait times.) 1007
W
The Wi-INET Model for Achieving M-Health Success
•
Employers wanting to increase productivity and avoid absenteeism with a healthier workforce.
Research Sponsor(s) [“Research Findings” in Figure 1]: The project develops intellectual property for researchers in the fields of: • • • •
Patient and Healthcare Provider Relationships Wireless Medical Informatics Therapeutic Compliance Business Case Wireless Information Technology Usability An INET Mobile E-Health Project Delivery Team:
Healthcare Delivery Team [“People” in Figure 1]: For a wireless diabetes program the players may include: • • • •
Healthcare Consumer: People with Diabetes Community Care: Nurse Specializing in Diabetes Primary Care: Family Physician Acute Care: Endocrinologist and Diabetes Education/Management Centers Business Process Analyst [“Process” in Figure 1]
1]: 1]:
Privacy and Security Consultant [“Protection” in Figure Programmer using a wireless ASP [“Platform” in Figure
• • • •
Wireless Network and Devices Device and Application Transcoding Application Service Provider Back-End Connection
In conclusion, INET is looking forward to further advancements in the mobile e-health project delivery model to: • • • •
achieve rapid advancements in healthcare delivery, improve diabetes management, enhance therapeutic compliance, and realize significant healthcare care cost savings.3
INET is planning to continue its role as a source of use case scenarios for the model with the delivery of mobile e-health projects.
FUTURE TRENDS AND CRITICAL SUCCESS FACTORS INET’s experience and preliminary findings are pointing to one critical success factor—the commercialization of physi1008
cians’ innovations in chronic disease management. INET is expecting significant growth in this healthcare trend, with wireless technology as the fundamental enabler. With antidotal comments from patients that participated in the INET pilot project are testimonials to this new era in healthcare delivery. A few of their comments (Canadian Healthcare Technology, 2005) include: “All I had to do was pick up my phone and dial, punch in my ID, put in the number for my blood sugar level, and hit enter. That was it. Did it twice a day,” said patient Jim Pott at the INET conference. “It gave me the feeling of constantly being looked after because, if I forgot to do it, Dr. Silver would be on my case, sending a text message to remind me.” For patient Dave Rowan: “I’m in the computer industry, so using an input device like my Treo phone is something I am very used to. Even so, to have it confirmed that my information was received and have subsequent interaction with Dr. Silver was terrific. And I could do it from all over North America.” For Joy Merritt, the trial has been life transforming. “Dr. Silver saved my life with this system. I felt totally out of control before. But now I feel I can control my life. And because I see my levels every day, and I know the doctor does too, that’s motivated me to exercise more and try to bring them down.” The preceding has served to outline all the critical aspects that must be considered when trying to actualize a mobile health initiative. Clearly, mobile e-health or m-health projects are complex and require much planning and coordination within and between the web of healthcare players. Success is never guaranteed in any large initiative, however in order to realize the four major healthcare deliverables depicted in Figure 1 (enhance patient care and safety, improve research and education, increase healthcare quality, and reduce healthcare costs), it is vital that any m-health initiative focus on the key success factors of people process and technology. Specifically, the technology must be correct and functioning as desired. Further, it must integrate seamlessly with existing ICT infrastructure and enable the processes. The processes must be well defined and at all times ensure that they are of a high quality and error free. The Institute of Medicine in America (2001) identified medical errors as the fourth leading cause of many deaths. In trying to prevent such errors, it has identified six key quality aims: 1. 2.
3.
Healthcare Should be Bafe: Avoiding injuries to patients from the care that is intended to help them. Healthcare Should be Effective: Providing services based on scientific knowledge to all who could benefit, and refraining from providing services to those who will not benefit (i.e., avoiding under use and overuse). Healthcare Should be Patient-Centered: Providing care that is respectful of and responsive to individual
The Wi-INET Model for Achieving M-Health Success
4. 5. 6.
patient preferences, needs, and values, and ensuring that patient values guide all clinical decisions. Healthcare Should be Timely: Reducing waiting and sometimes harmful delays for both those receiving care and those who give care. Healthcare Should be Efficient: Avoiding waste. Healthcare Should be Equitable: Providing care that does not vary in quality based on personal characteristics.
Finally, and arguably the most critical key success factor, is the web of healthcare players. Any m-health project must consider the impact and role on each of these players, the interactions of such an initiative both within one group of the web of players as well as between groups of players. As discussed, and based on the findings from INETs longitudinal studies, it is critical to the ultimate success of these projects and the ability to in fact realize the healthcare deliverables that they are indeed physician led.
DISCUSSION AND CONCLUSION
Boyatzis, R. (1998). Transforming qualitative information thematic analysis and code development. Thousand Oaks, CA: Sage. Canadian Healthcare Technology. (2005). Wireless reporting for diabetes patients offers up dramatic results. Canadian Healthcare Technology, (September), 4. Retrieved from http://www.inet-international.com/INET/Update/PressCoverage2005.htm Eisenhardt, K. (1989). Building theories from case study research. Academy of Management Review, 14, 532-550. European Institute of Medicine. (2003). Health is wealth: Strategic vision for European healthcare at the beginning of the 21st century. Salzburg, Austria: European Academy of Arts and Sciences. Frost & Sullivan Country Industry Forecast. (2004). European Union healthcare industry. Retrieved May 11, 2004, from http://www.news-medical.net/print_article.asp?id=1405 Goldberg, S. et al. (2002a, January). Building the evidence for a standardized mobile Internet (wireless) environment in Ontario, Canada. Internal INET Documentation.
Healthcare in the United States and globally is at the crossroads. It is facing numerous challenges in terms of demographics, technology, and finance. The healthcare industry is responding by trying to address the key areas of access, quality, and value. M-health, or mobile e-health, provides a tremendous opportunity for healthcare to make the necessary evolutionary steps in order to realize its goals and truly achieve its value proposition. What is important is to ensure m-health excellence. This requires not only detailed theoretically studies, but ultimately the need to turn theory into practice. By an in-depth analysis of the rich and longitudinal data of INET, we have developed the Wi-INET model to facilitate the achievement of m-health excellence. Systematic and detailed analysis and integration of all the key drivers and implication of healthcare delivery have enabled the development of the Wi-INET model. Moreover its structure facilitates rapid development and actualization of m-health solutions. To the best of our knowledge, it is the first such model, and while it is certainly not a panacea, it does help to set the stage and outline the key issues that must be addressed for a successful m-health initiative, and it enables healthcare to reap the benefits of wireless. We are confident that through the adoption of the Wi-INET model, healthcare delivery too can make revolutionary changes and we can all enjoy superior healthcare delivery.
Goldberg, S. et al. (2002b). HTA presentational selection and aggregation component summary. Internal Documentation.
REFERENCES
Kyprianou, M. (2005). The new European healthcare agenda. Proceedings of the European Voice Conference: Healthcare: Is Europe Getting Better? Retrieved from http://www.noticias.info/asp/aspcommunicados.asp?nid=45584
Blair, J. (2004). Assessing the value of the Internet in health improvement. Nursing Times, 100, 28-30.
Goldberg, S. et al. (2002c). Wireless POC device component summary. Internal INET Documentation. Goldberg, S. et al. (2002d). HTA presentation rendering component summary. Internal INET Documentation. Goldberg, S. et al. (2002e). HTA quality assurance component summary. Internal INET Documentation. INET Talk. (2004, October 14). Enhance therapeutic compliance using wireless technology. Proceedings of the WNY Technology & Biomedical Informatics Forum, Niagara Falls, NY. Institute of Medicine in America. (2001). Crossing the quality chasm: A new health system for the 21st century. Committee on Quality of Healthcare in America, Institute of Medicine. Washington, DC: National Academy Press. Kavale, S. (1996). An introduction to qualitative research interviewing. Thousand Oaks, CA: Sage. Kulkarni, R, & Nathanson, L.A. (2005). Medical informatics in medicine. Retrieved from http://www.emedicine. com/emerg/topic879.htm
1009
W
The Wi-INET Model for Achieving M-Health Success
Lacroix, A. (1999). International concerted action on collaboration in telemedicine: G8sub-project 4. Sted. Health Technol. Inform, 64, 12-9. Lee, M. Y., Albright, S. A., Alkasab, T., Damassa, D. A., Wang, P. J., & Eaton, E. K. (2003). Tufts Health Sciences Database: Lessons, issues, and opportunities. Academic Medicine, 78, 254-264. National Center for Health Statistics. (2002). Health expenditures 2002. Retrieved from http://www.cdc.gov/nchs/ fastats/hexpense.htm National Coalition on Healthcare. (2004). Building a better health: Specifications for reform. Washington, DC: National Coalition on Healthcare. Organisation for Economic Cooperation and Development. (2004). OECD health data 2004. Retrieved from www.oecd. org/health/healthdata Pallarito, K. (1996). Virtual healthcare. Modern Healthcare, (March), 42-44. Plunkett’s. (2005). Plunkett’s health care industry almanac. Houston: Plunkett Research. Russo, H.E. (2000). The Internet: Building knowledge & offering integrated solutions to health care. Caring 19, 1820, 22-24, 28-31. Standish Group International. (1994). The CHAOS report. Retrieved from http://www.standishgroup.com/sample_research/chaos_1994_1.php von Lubitz, D., & Wickramasinghe, N. (2005). Healthcare and technology: The doctrine of network-centric healthcare. Health Affairs. WHO. (2000). Health systems: Improving performance (pp. 1-215). Geneva: World Health Organization. WHO. (2004). Changing history (pp. 1-167). Geneva: World Health Organization.
Wickramasinghe, N. et al. (2005). Assessing e-health. In T. Spil & R. Schuring (Eds.), E-health systems diffusion and use: The innovation, the user and the user IT model. Hershey, PA: Idea Group Publishing.
KEY TERMS Healthcare Challenge: One of the challenges that can be thought of in terms of demographic issues such as the aging population, technology issues such as the need to embrace technologies into healthcare, and finance issues such as escalating healthcare costs. Key Healthcare System Input: Includes people, process, platform, and protection. M-Health: The application of wireless technology to healthcare delivery. Mobile Healthcare Business Model: A model to identify all the key drivers and actors for a mobile healthcare initiative. Mobile Healthcare Delivery Model: A model that outlines the process and procedures to realize the outlined business model. Superior Healthcare Delivery: A patient-centric healthcare delivery system that tries to maximize access, quality, and value. Three-Tier Web-Based Architecture: ICT backbone of the wireless m-health initiative. Wireless Healthcare Program: The incorporation of wireless technology into a specific healthcare program, for example, the use of wireless technology in a diabetes program.
ENDNOTES 1
Wickramasinghe, N., & Goldberg, S. (2004). How M=EC2 in healthcare. International Journal of Mobile Communications, 2(2), 140-156. Wickramasinghe, N., & Mills, G. (2001). MARS: The electronic medical record system, the core of the Kaiser galaxy. International Journal of Healthcare Technology Management, 3(5/6), 406-423. Wickramasinghe, N., & Silvers, J.B. (2003). IS/IT: The prescription to enable medical group practices to manage managed care. Health Care Management Science, 6, 75-86.
1010
2
3
Fourteen percent to 21% of patients never fill their original prescription, and 10% to 50% of patients ignore or otherwise compromise their medication instructions (http://www.managedhealthcareexecutive. com/mhe/article/articleDetail.jsp?id=105388). Wireless Diabetes Management Protocol ©Dr. Sheldon Silver, MD, 2005. In Ontario, this may save 1 $billion over three years (INET Talk, 2004).
Category: Wireless Networking 1011
Wireless Access Control System Using Bluetooth Juliano Rodrigues Fernandes de Oliveira Federal University of Campina Grande, Brazil Rodrigo Nóbrega Rocha Xavier Federal University of Campina Grande, Brazil Yuri de Carvalho Gomes Federal University of Campina Grande, Brazil Hyggo Almeida Federal University of Campina Grande, Brazil Angelo Perkusich Federal University of Campina Grande, Brazil
INTRODUCTION
BACKGROUND
Security is one of the world’s main challenges. Research and industrial applications related to security include several areas such as personal security, organizational security, and computer security, among others. This article is concerned with secure environments, which is related to the control of people entering an environment, building, rooms, laboratories, and so forth. In this context, access control systems are the main security mechanisms to control the access of authorized people to environments. Nowadays, locks and keys are not enough to keep an environment secure against unwanted or uncontrolled visitors. To have access, mechanical security systems are widely used, however, such systems—purely mechanical—can be easily defrauded. To construct high-security access systems, the embedded electronics have associated to the mechanical security, with the objective of increasing the level of reliability of such systems. Besides, with the increasing use of mobile devices, users are more and more interested in mobile solutions to support several activities, including security-related ones. This article presents an access control system that uses Bluetooth technology (Ericsson Bluetooth, 2006) to allow control of the entrance to environments. By using the proposed system, a person with a smart phone can use it to get access to environments, such as buildings, labs, rooms, and so forth. The remainder of this article is organized as follows. First we present the architectural components of the proposed system and detail their functioning. We then discuss future trends and offer concluding remarks.
Bluetooth The Bluetooth specification was developed by Ericsson (now Sony Ericsson) and later formalized by the Bluetooth Special Interest Group (SIG). The SIG was formally announced on May 20, 1999, and originally founded by Ericsson, IBM, Intel, Nokia, and Toshiba. Bluetooth is an industrial standard for wireless personal area networks (PANs), also known as IEEE 802.15.1 (Bluetooth SIG, 2004). It provides a secure, low-cost way to connect and exchange information between devices, such as personal digital assistants (PDAs), mobile phones, laptops, PCs, printers, and digital cameras, in a globally available short-range radio frequency. This technology eliminates cables and wires between devices, facilitates both data and voice communication, and enables ad-hoc networks between multiple Bluetooth devices (Cardei, 2002). Bluetooth is a radio standard primarily designed for low power consumption, with a short range (power class dependent: 1 meter, 10 meters, 100 meters) and with a low-cost transceiver microchip in each device. It lets these devices communicate with each other when they come in range, even if they are not in the same room, as long as they are within up to 100 meters of each other, depending on the power class of the product (Kardach, 1998).
Microcontrollers A microcontroller (MCU) is a computer-on-a-chip used to control electronic devices. It is a microprocessor emphasiz-
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
W
Wireless Access Control System Using Bluetooth
Figure 1. Wireless access control system architecture
reduces the number of chips and the amount of wiring and space that would be needed to produce equivalent systems using separate chips. Manufacturers and designers have to balance the need to minimize the chip size against additional functionality.
SYSTEM ARCHITECTURE
ing self-sufficiency and cost-effectiveness, in contrast to a general-purpose microprocessor used in a PC. It can be defined as a single integrated circuit with a central processing unit, usually small and simple; input/output interfaces, such as serial ports; peripherals, such as timers and watchdog circuits; RAM for data storage; ROM for program storage; and a clock generator, often an oscillator for a quartz timing crystal, resonator, or RC circuit (Stewart, 1993). In addition to the key features, most microcontrollers today take further advantage of not needing external pins for memory buses. They can afford to use the Harvard architecture: separate memory buses for instructions and data, allowing multiple access to occur concurrently (Cady, 1997). A typical microcontroller contains all memory and interfaces needed for a simple application, whereas a general purpose microprocessor requires additional chips to provide these functions. Microcontrollers also usually have a variety of input/output interfaces. Serial I/O (UART) is very common, and many include analog-to-digital converters, timers, or specialized serial communications interfaces like I2C, serial peripheral interface (SPI), and controller area network (CAN). A microcontroller is also a programmable device that can be destined for several purposes. The firmware recorded in its memory is responsible for the characteristic of its application. Microcontrollers are versatile tools and with low cost for embedded systems design. Originally, microcontrollers were only programmed in assembly language, or later in C code. Recent microcontrollers, integrated with on-chip debug circuitry accessed by in-circuit emulator via JTAG, enable a programmer to debug the software of an embedded system with a debugger (Cady, 1997). Microcontrollers trade speed and flexibility against ease of equipment design and low cost. This integration drastically 1012
The access control system architecture depicted in Figure 1 consists of two modules: mobile and base. A smart phone contains the software responsible for beginning the authentication process, acting as mobile module. The base module is responsible to receive a valid authentication code and to allow the access to the environment by unlocking an electric lock embedded in the environment entrance door. The base module is composed of a Bluetooth module (Wintec BT Module, 2005); a processing kernel, represented by a microcontroller (Microchip PIC18FXX2, 2002); an external data storage unit, represented by an EEPROM memory (Microchip 24LC256, 2002); and an electric lock interface, represented by a driver circuit to unlock the electric lock. In general, the user authentication process consists of sending the user authentication key from the application running in a mobile device to the Bluetooth module, through Bluetooth connection. The Bluetooth module sends such information to the processing kernel, which performs the authentication through comparison of the user key sent with that stored in the external data storage unit. Next, the processing kernel sends the search authentication result to the mobile device and to the electric lock interface. If the user key is valid, the electric lock interface unlocks the environment entrance door. Each architectural component is detailed in what follows.
Mobile Module and Bluetooth Module Mobile module is the application embedded in a mobile device that performs the communication with the Bluetooth module. It has been developed in J2ME language (J2ME, 2006). Such an application is based on the Bluelet open source software (Bluelet, 2006). The entire connection negotiation process has been implemented using protocols of the Bluetooth protocol stack to perform connections via Serial Port Profile (Wintec Bluetooth, 2004). The basic process to connection negotiation consists of three steps: 1. 2.
Search for the Bluetooth module (discovery function), through the name “Wintec Serial Port” or the Bluetooth module address. Authentication, or pairing, using the code sent by the mobile device to the Bluetooth module (bond function).
Wireless Access Control System Using Bluetooth
3.
Connection establishment (connect function) based on serial port profile. A wireless communication is emulated by a connection via serial port with UART protocol (Wintec Bluetooth, 2004).
Regarding the application functioning, first of all an initial connection attempt is performed to connect to the Bluetooth Module directly, through the Bluetooth module address. Such an address is discovered by localization of Bluetooth devices and stored in the device. If the Bluetooth module is inaccessible—distant or turned off—or if it is not found in the direct connection attempt, the software automatically performs two more attempts, notifying the user visually. If no attempt works, then the software will be finished. After the Bluetooth connection establishment, the access control system awaits the sending of the user key. In this case, a string containing the user authentication key is transmitted from the smart phone to the Bluetooth module. This key needs to be registered in the application by the user and then stored in the mobile device. If the operation is successful, the electric lock is unlocked and the user is informed visually that the door has been opened. Afterwards, the connection is closed and the application is finished.
Processing Kernel During the connection establishment process between the smart phone and the Bluetooth module, messages (in ASCII format) are sent from the Bluetooth module to the microcontroller host, which represents the Processing Kernel. The firmware contained in the microcontroller host monitors these messages awaiting the final message of connection closing. Meanwhile, it continues awaiting the user authentication key that must be sent by the mobile device. When it receives such a key, it analyzes it, and if the key is registered in the system database, the host will unlock the electric lock. After the data exchange between the mobile device and the Bluetooth module, the microcontroller sends to the Bluetooth module the escape sequence to close the Bluetooth connection.
External Data Storage Unit The external memory EEPROM is used as a persistent data storage device in this wireless access control system. It is very important due to the reduced capacity of internal memory available in the microcontroller. The user information is stored in the external memory as a login/password table. A pointer to free memory positions is used to indicate which memory spaces are available to register new users. The recording operation of user authentication keys begins when the processing kernel receives a write command.
The user information, login and password, contained in this command are then stored in the external memory unit. The search operation for user authentication keys begins when the processing kernel receives an access command. The user information, login and password, contained in this command are acquired and stored in vectors, for search and comparison with stored data in the memory unit. If user information is valid, the operation will be successful. Otherwise, the operation has failed and the electric lock will not be unlocked.
Electric Lock Interface The electric lock interface is represented by a driver circuit of the electric lock and by a panel composed by LEDs that indicates the current operations during the authentication process. It is responsible for informing of the status of the driver circuit. There are four LEDs. One of them indicates that the circuit is energized and active. Another indicates that a recording operation is currently being performed. Yet another LED indicates that the search operation for user key has been performed successfully and the electric lock was unlocked, allowing access to the environment. The last LED indicates that the search operation has failed (invalid code) and the electric lock has not been unlocked, not allowing the user access to the environment.
FUTURE TRENDS In the context of the proposed system, in order to improve the reliability of the data exchange between the remote device and the access control system, revisions in the firmware contained in the processing kernel are necessary. Some suggestions to increase the reliability are: addition of an administrator user, development of new functions for this administrator to control and supervise the system, and better sub-routines to search for registered users. Regarding future efforts in mobile-related access control technologies, the main trends are concerned with pervasive environments. For example, current access control technology works by keeping the entrance door closed and opening it for authorized persons. But there is another way: the door could be left open and only closes when an unauthorized person tries to enter. In this case, the access control system must be monitoring the environment, and when an unauthorized person comes close, the door is blocked. This example has a pervasive characteristic in which the system is “invisible” for the user. Several access control researches are moving in that direction, to conceive environments that manage and control their security without needing a direct user intervention.
1013
W
Wireless Access Control System Using Bluetooth
CONCLUSION Security is a growing need throughout the world, and lack of security can result in great damage. Many solutions are available for all levels of access control—from highly restricted areas such as laboratories or computer rooms to less restricted areas such as storage rooms. Access control solutions include electronic keys, magnetic stripe cards, proximity cards, and smart cards or biometric devices, including hand and fingerprint readers. More sophisticated access control capabilities, such as auto-unlock/auto-lock functionalities, allow programming an electronic locking system to lock and unlock any door at any time. With the increasing personal use of mobile devices and growing industry investments in this area, it is a trend to use mobility capabilities to support user activities, mainly security-related ones. The proposed access control system using Bluetooth is a good example of how to join mobility and reliability to support user activities in a practical and secure way.
REFERENCES Bluelet. (2006). Bluetooth GUI component. Retrieved from http://benhui.net/ Bluetooth SIG. (2004). Bluetooth Special Interest Group launches Bluetooth Core Specification version 2.0 + Enhanced Data Rate. Cady, F. M. (1997). Microcontrollers and microcomputers. Oxford: Oxford University Press. Cardei, M. (2002). Overview over the Bluetooth technology. Proceedings of the Wireless Networking Seminar, University of Minnesota. Ericsson Bluetooth. (2006). Ericsson Bluetooth. Retrieved from http://www.ericsson.com.br/bluetooth/index.asp J2ME. (2006). Java 2 Micro Edition. Retrieved from http:// java.sun.com/j2me/index.jsp Kardach, J. (1998). Bluetooth architecture overview. Intel. Microchip 24LC256. (2002). Microchip 24LC256 256k I2C CMOS Serial EEPROM datasheet. Microchip Technology. Microchip PIC18FXX2. (2002). Microchip PIC18FXX2 datasheet. Microchip Technology.
1014
Stewart, J.W. (1993). The microcontroller. Englewood Cliffs, NJ: Regents/Prentice-Hall. Wintec Bluetooth. (2004). Wintec Bluetooth SPP command interface quick start guide. Wintec Industries. Wintec BT Module. (2005). WBTV42-D-XXX Bluetooth module rev 0.8 guide. Wintec Industries.
KEY TERMS American Standard Code for Information Interchange (ASCII): A standard for coding text files. Every character has an associated number, and any text can be represented by a sequence of numbers. Electrically Erasable Programmable Read-Only Memory (EEPROM): A non-volatile storage chip used in computers and other devices (such as USB flash drives, in its flash memory version); also called E2PROM. I2C (Inter-IC) Bus: A bi-directional two-wire serial bus that provides a communication link between integrated circuits (ICs). J2ME: Collection of Java APIs for the development of software for resource-constrained devices such as PDAs, cell phones, and other consumer appliances. Serial Port Profile (SPP): Defines how to configure and connect the virtual serial port between two wireless devices supporting Bluetooth technology. Smart Phone: Any electronic handheld device that integrates the functionality of a mobile phone, personal digital assistant (PDA), or other information appliance. This is often achieved by adding telephone functions to an existing PDA or putting “smart” capabilities, such as PDA functions, into a mobile phone. A key feature of a smart phone is that additional applications can be installed on the device. The applications can be developed by the manufacturer of the handheld device, by the operator, or by any other third-party software developer. Symbian OS: An operating system designed for mobile devices, with associated libraries, user interface frameworks, and reference implementations of common tools, produced by Symbian Ltd. UART: Universal asynchronous receiver/transmitter protocol.
Category: Wireless Networking 1015
Wireless Client Server Application Model Using Limited Key Generation Technique Rohit Singh Monash University, Australia Dhilak Damodaran Monash University, Australia Phu Dung Le Monash University, Australia
INTRODUCTION The introduction of personal digital assistants (PDAs) and laptops has brought in mobility for carrying out computing jobs (Pasquale, Hung, Newhouse, Steinberg, & Ramabhadran, 2002; Varshney & Vetter, 2000). Wireless networks in working areas cater the need for installation flexibility, reduced cost-of-ownership, mobility and scalability. Mobile computing is distinguished from classical, fixed-connection computing due to the mobility of nomadic users and their devices (Jing, Helal, & Elmagarmid, 1999) and mobile environment is defined as low bandwidth and high latency networks with devices supporting limited input methods and containing low power processors, small display size and short battery life (Buszko, Lee, & Helal, 2001). Mobile working environments attract the users, employees and students, but presents hardships to the security programmer because the settings and the environment of mobile devices vary significantly from that of wired devices. The ease of network access should be combined with reliable security services. Wireless networks are exposed to the security compromises because it provides hacker easy access to transport media. An attacker can sniff the packets by setting up equipment monitoring at 2.4GHz frequencies and capable of interpreting the packets of 802.11 standard (Borisov, Goldberg, & Wagner, 2001; IEEE Standard 802.11b, 1999; Josang & Sanderud, 2003). This unauthorized access to the network is a matter of great concern for any network security administrator. Cryptography, that is encryption of the data, is one of the best ways to provide security to wireless communication. Even the strongest of the encryption procedures can become vulnerable to possible attacks when the keys of the parties get compromised. Thus the lack of security is predominantly due to poor management of keys rather than the weakness in the encryption algorithm (Josang & Sanderud, 2003). In this article we implement a client server model using limited-used key generation scheme (Kungpisdan, Le,
& Srinivasan, 2004) to generate a set of session keys that are never transmitted, which means that there is no chance for the attacker to sniff the packets and retrieve keys while they are being transmitted. These session keys are used for encrypting and hashing the data to be transmitted from mobile client device to the servers in wired network and vice versa. The updating of the session keys used in this technique does not rely on any long-term shared key, instead the process is based upon the last session key used. This technique of elevating the frequency of the key update to the next possible level makes the system much more secure than the other present techniques. In addition to providing better security, this technique also enhances the performance of a limited resource device by avoiding the repeated generation of keys on it. The rest of the article is organized as follows. The second section gives an overview of the communication between mobile devices and the various servers running in the wired network. The third section describes the proposed technique. The fourth section discusses the technique in applied state. Next, the fifth section discusses about other key management and security issues of the technique. The last section concludes the article.
CLIENT SERVER MODEL The client server model in a wireless environment is depicted in Figure 1. Mobile clients in wireless networks first connect to an access point, which is connected with wired media to the main network infrastructure. After a successful connection with the access point, mobile clients can start communicating with the gateway and servers that are present within the main network. The model described in this article proposes to set up a gateway program which authenticates every mobile client before it connects to any server inside the main network. The authentication is carried on the basis of the technique discussed in the fourth section. The gateway program can be executed at the gateway server of organization.
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
W
Wireless Client Server Application Model Using Limited Key Generation Technique
Figure 1. Remote access model
Gateway
Mobile Devices
The course of action that is carried out in the proposed protocol is as follows: • • • •
C G: Server Connection (Request), Command Execution (Request) G S: Command-Execution (Request) S G: Command-Execution (Response) G C: Server Connection (Response), CommandExecution (Response)
PROPOSED TECHNIQUE Notations • • • • • • • •
1016
{C, G, S}: the set of clients, gateway and server, respectively. {KA, KA-1}: the set of public/ private key for party A. {IDC, IDS, IDG}: the set of identities of client, the server and the gateway respectively. IDGReq: request for IDG CI: Command execution information. CRes: Command response {M} X: the message M symmetrically encrypted with the shared key X. {M} Kx: the message M encrypted with the public key of the party X.
• • •
{M} Kx-1: the message M signed with the private key of the party X. h (M): the one-way hash function (FIPS, 1995) of the message M. MAC (M, K): the message authentication code (MAC) of the message M with the key K.
Initial Assumption and Settings Initial assumptions and settings for the proposed technique are as follows: 1. 2.
3.
Client is considered to be a person with the wireless device and has access to the organization’s network. Client’s employment record (CER), containing employee code and other information about the client, is the long-time shared secret between the server and Client. CER is assumed to be a key that never expires. The distributed key DK is another shared key between client and server. This key is distributed by performing authenticated key exchange (AKE) (Boyd & Park, 1998; Horn & Preneel, 1998; Kungpisdan, Le, & Srinivasan, 2003; Toh, Kungpisdan, & Le, 2004; Wong & Chan, 2001; Zhu, Wong, Chan, & Ye, 2002) protocol between client and server.
Wireless Client Server Application Model Using Limited Key Generation Technique
K1= h (DK, CER), K2=h (DK, K1)……., Km= h (DK, Km-1)
C S: {IDC, DK, n} k S C: {n} k
4.
This key is then further used to generate session key Yi using the technique explained in the third section. There is another distributed key X that is a shared secret between client and gateway. This key is transferred to the gateway by performing AKE protocol. C G: {IDC, X, n} k G C: {n} k
5.
2.
3.
Client generates the distributed key and transfers the same to the server through a secure channel. Client and server generate a set of preference keys Ki, where i=1,2,3,….m:
Client selects two preference keys Kmid1 and Kmid2. Kmid1is the middle key among {K1, K2, K3……, Kw}, where w= r mod m and Kmid2 is the middle key among {K1, K2…. Kmid1}. Using these keys session initialization key (SIK) is calculated. SIK= h (Kmid1, Kmid2)
Key Generation Technique
1.
CER is then removed from the system. This type of recursive hashing is represented by DK*CER in remaining part of paper. Client generates a random number r, which is sent to the server at the start of every session. C S: r
This key X is then further used to generate session key Xi using the technique explained in 3.3.2. h (M, K) stands for the key hashed function for the message M and key K. For higher security reason HMAC is preferred.
Generating Yi: Figure 2 shows the procedure to generate Yi. The steps to generate keys on the communicating terminals are as follows:
W
4.
After generating SIK, Kmid1and Kmid2 are removed from the terminals. As the value of r was transmitted over to the server, the same SIK is generated at the server’s terminal as well. Server and client generate a set of session keys Yi, where i=1, 2…, n. Y1 = h (SIK, DK), Y2 = h (SIK, Y1)... Yn = h (SIK,Yn−1)
These session keys are then used by server and client to encrypt, decrypt and hash the data passed between them.
Figure 2. Generation of Yi
1017
Wireless Client Server Application Model Using Limited Key Generation Technique
Figure 3. Generation of Xi
The application of the session keys in proposed technique is discussed in the fourth section. Generating Xi: Figure 3 shows the procedure to generate Xi. The steps to generate keys on communicating terminals are as follows: 1.
2.
These session keys are then used by gateway and client to encrypt, decrypt and verify the messages sent amongst them.
KK1= h (X, X m-bit-shift), KK2=h (X, KK1)……., KK m=h (X, KKm-1)
Updating Session Keys
Client generates a random number s, which is later on send to gateway.
Client selects two preference keys KKmid1, and KKmid2. KKmid1 is the middle key among {KK1, KK2, KK3……, KKx}, where x= s mod rm and rm stands for remaining number of keys in the set of KKi. The second key KKmid2 is the middle key among {KK1, KK2…. KKmid1}. After this, session initialization key (SIK′) is calculated. SIK′= h (KKmid1, KKmid2)
1018
X1 = h (SIK′, X), X2 = h (SIK′, X1)... Xn = h (SIK′, Xn−1)
Client generates the key and distributes to the gateway through a secure channel. Client and gateway generate a set of preference keys KK i, where i=1, 2, 3….m.
C G: s 3.
4.
After generating SIK′, KKmid1and KKmid2 are removed from the terminals. Gateway and client generate a set of session keys Xi, where i=1, 2…, n.
Figure 4 shows the procedure to update session keys. The technique proposed for updating session keys Yi and Xi is similar, so for this section Yi will be used for representing both. The session key update is performed as follows: •
Step 1: Client and the server have used Yi up to YP, where 1 ≤ p ≤n. Then two preference keys are chosen from remaining Ki: the first key selected is K′Mid1, where K′Mid1 is the middle key among {K1... Kq}, q = p mod rm, rm is the rm is the number of preference keys in the set of Ki. The second preference key chosen is K′Mid2, where K′Mid2 is the middle key among {K1 …K′Mid1}. Then, a new session initialization key (SIK′) is generated as follows:
Wireless Client Server Application Model Using Limited Key Generation Technique
Figure 4. Updating session keys
SIK′ = h (K′Mid1, K′Mid2)
•
After generating SIK′, K′Mid1 and K′Mid2 are removed from both the systems. Step 2: After SIK′ is generated, A new set of session key Y′i is generated where i = 1, 2, 3…n, using the following method. Y′1 = h (SIK, DK), Y′2 = h (SIK, Y′1)... Y′n = h (SIK, Y′n −1)
This results in the generation of a set of new session keys (Y′i) without redistributing or updating the distributed key DK. DK can be used repeatedly until all the preference keys in set of Ki gets used up.
Updating Distributed Key Updating the distributed key leads to generation a new set preference keys K′i. After new DK′ has been generated and distributed, K′i is generated as follows: It should be noted that CER, which was used with DK for the first time to generate Ki, is no more stored on the client terminal. Thus a preference key KV is selected from remaining Ki, where v= s mod rm and s is the index of most recently used session key YS. K′1= h (DK′, Kv), K′2=h (DK′, K′1)……., Km= h (DK′, K′m-1)
W
After having the new set of K′i the same procedure, as described in the third section, can be used to generate the new set of session keys.
Applying the Proposed Technique After generating the session keys Xi and Yi and the random numbers s and r on the client terminal, the client side protocol proceeds in the following way to connect to remote server and to get its command executed. •
Step 1: C G: IDC, s, r, IDGReq G C: {IDG} X i
•
Step 2: C G: {IDS, CI, MAC [(IDC, IDG, CI), Yi]} MAC [(IDC, IDS), Xi+1] Xi,
•
Step 3: G S: {{MAC [(IDC, IDG, CI), Yi], IDG, CI, r, IDC} K S} K G-1
•
Step 4: S G: {{{CRes} Yi, IDC} K G} K S-1
•
Step 5: G C: {{CRes} Yi} Xi+1
Command information (CI) and command response (CRes) can vary according to the server to which the client wants to connect. The technique discussed above has been implemented in a lab environment using HP iPAQ Pocket PC as client device, D-link wireless bridge as access point,
1019
Wireless Client Server Application Model Using Limited Key Generation Technique
and Redhat Linux machine with the gateway program and server applications accessing databases on Windows and Linux machines.
DISCUSSION Key Management For the remaining part of the article we will be discussing about the keys on the server side, but any discussed feature on keys will also apply on the keys with the gateway. The proposed technique requires two set of keys, Ki and Yi, to carry out the secure communication. Both of the keys are generated on the terminals of each party but only Ki is stored on the terminals. The stored set of keys reduces the overhead of generating Ki for every new session thus enhancing the performance of client device.
Security of the Proposed Technique Against Possible Attacks The long-time shared secret (CER) is never stored on the terminal. It is removed as soon as the first Ki is generated from it and never used again. The set of preference keys, which were used to generate SIK, is also removed from the terminal to eliminate any chance of regenerating a previously used SIK. Non-deployment of CER in the proposed technique offers flexibility in making changes to the records without giving much consideration about the network connectivity. The remainder of the section points out the other attacks and security schemes that need to be employed to keep track of the attacks and how to recover from them. Given a situation that the session key Yi gets compromised, even then the value of SIK still remains secure due to the fact that it is totally infeasible to perform the reverse computation on the values generated by one way hash functions. In another situation where an attacker has collected a number of session keys and tries to guess the next session key, the server can keep track of the total number of incorrectly hashed or encrypted messages and can even suspend the client’s account in case the number of incorrect messages goes beyond the precise limit. Taking into account the worst scenario, where an attacker has guessed all the right values of session keys and the server has failed to track the fake messages, the short life span of session keys can embark upon this threat in a comprehensively better way. So the compromise of session keys does not concern the participating parties in a longer run. As session keys are a result of hash function (SIK*DK), there is a possibility that the session keys can be generated if these values are compromised. If an attacker has captured all the session keys and has generated the same set of session 1020
keys, the deceit can still be traced out by the system when it finds two MACs (one from a valid client and one from a fake client) hashed with same session keys. A simple solution to this situation is to update the session keys. The participating parties—who ever discovers the deceit—requests the other party to update their session keys. Thus both the parties generate new set of session keys Yi′ (as described in the third section) by generating SIK′ which makes the compromised Yi and SIK as invalid. A successful attack on this technique means that the attacker should succeed in capturing the entire set of Ki from the device of any of the participating parties and hack the values of r, p and DK during distribution. Given a scenario as mentioned the deceit will still be detected by the system when it receives a MAC with a previously used session key. The proposed technique inherits these security features because of the property that it never re-uses a session key and, moreover, the generation of these session keys are not based on any long term shared key. These keys are generated from randomly chosen set of preference keys Ki, which are deleted after they are used. Thus, more and more usage of the technique for conducting sessions keeps on enhancing the security of the system.
CONCLUSION In this article we presented the wireless client model for remote access to servers by mobile clients based on a simple but secure key generation protocol. We modified and implemented the technique that was originally proposed by Kungpisdan, Le, and Srinivasan (2003) in our protocol to generate session keys for encrypting and hashing the data ought to be transmitted. Then we applied the protocol in our model to demonstrate the security provided to the wireless communications between mobile clients and servers. Our work considerably enhances the systems capability to alleviate attacks based on key compromise.
REFERENCES Borisov, N., Goldberg, I., & Wagner, D. (2001). Intercepting mobile communications: The insecurity of 802.11. Paper presented at the 7th annual International Conference on Mobile Computing and Networking, Rome, Italy. Boyd, C., & Park, D. G. (1998). Public key protocols for wireless communications. In Proceedings of the ICISC 1998 (pp. 47-57). Seoul, Korea. Buszko, D., Lee, W. H., & Helal, A. (2001). Decentralized ad-hoc groupware API and framework for mobile collaboration. Paper presented at the 2001 International ACM SIG-
Wireless Client Server Application Model Using Limited Key Generation Technique
GROUP Conference on Supporting Group Work, Boulder, Colorado, USA. FIPS. (1995). Secure hash standard (SHS). Federal Information Processing Standards Publication (FIPS) PUB 180-1. Horn, G., & Preneel, B. (1998). Authentication and payment in future mobile systems. In Proceedings of 5th European Symposium on Research in Computer Security (pp. 277293). Belgium.
Wong, D. S., & Chan, A. H. (2001). Efficient and mutually authentication key exchange for low power computing devices. LNCS 2248. Zhu, F., Wong, D. S., Chan, A. H., & Ye, R. (2002). Password authenticated key exchange based on RSA for imbalanced wireless networks. LNCS 2433.
KEY TERMS
IEEE. (2000). Supplement to IEEE standard 802.11b-1999, Part 11: Wireless LAN medium access control (MAC) and physical layer (PHY) specifications: Higher-speed physical layer Extension in the 2.4 GHz band. New York. IEEE Publications.
Authenticated Key Exchange (AKE): AKE is block of operation in which two parties generate shared keys secretly. Many information exchanges are authenticated and secured on the basis of these shared keys.
Jing, J., Helal, A., & Elmagarmid, A. (1999). Client-server computing in mobile environments. ACM Computing Surveys (CSUR), 31(2), 117-157.
Hash Function: Hash functions are algorithms that accept messages of any length and compute a fixed length string termed as a digital fingerprint or message digest or hash value.
Josang, A., & Sanderud, G. (2003). Security in mobile communications: Challenges and opportunities. Paper presented at the Australasian Information Security Workshop Conference on ACSW Frontiers 2003, Adelaide, Australia. Kungpisdan, S., Le, P. D., & Srinivasan, B. (2004). A limited-use key generation scheme for Internet transactions. In Proceedings of the 5th International Workshop on Information Security Applications 2004 (WISA2004). Korea. Kungpisdan, L., & Srinivasan, S., Srinivasan, B., & Le, P. D. (2003) Lightweight mobile credit-card payment protocol. In Proceedings of the 4th International Conference on Cryptology (pp. 295-308). Lecture Notes in Computer Science. Pasquale, J., Hung, E., Newhouse, T., Steinberg, J., & Ramabhadran, N. (2002). Improving wireless access to the Internet by extending the client/server model. Paper presented at the European Wireless Conference, Florence, Italy. Toh B. T. S., Kungpisdan, S., & Le, P. D. (2004, December 1-3) KSL protocol: Design and implementation. In Proceedings of the 2004 IEEE Conference on Cybernetics and Intelligent Systems (pp. 544-549). Singapore. Varshney, U., & Vetter, R. (2000). Emerging mobile and wireless networks. Communications of the ACM, 43(6), 73-81.
Long-Term Shared Key: Long-term shared keys are keys that are shared for a longer period of time. This key forms the basis of authentication and security of information exchange between two parties. Message Authentication Code: MAC is defined as checksum for the data transferred through unreliable media like the Internet. This check sum is also computed at the receiving ends and then the similarity of the checksum computed and received proves the authenticity of the data. Mobile Environment: Is a generic term used to describe the environment where portable and computing devices connect wirelessly to utilize the central repository of data or applications. Session Initialization Key (SIK): SIK can be considered as the mother of all session keys. SIK is hashed recursively with long-time shared key to generate session keys. One SIK expires after it has generated a predefined number of session keys. Session Keys: Session keys are short lifespan encryption keys, which are used for encrypting one message or group of messages for a session.
1021
W
1022 Category: Security
Wireless Network Security Kevin Curran University of Ulster, Northern Ireland Elaine Smyth University of Ulster, Northern Ireland
INTRODUCTION Wireless networks have a number of security issues. Signal leakage means that network communications can be picked up outside the physical boundaries of the building in which they are being operated, meaning a hacker can operate from the street outside or discretely from blocks away. In addition to signal leakage, the wired equivalent privacy protocol is inherently weak, and in addition to WEP’s weaknesses, there are various other attacks that can be initiated against WLANs, all with detrimental effects. On the surface WLANs act the same as their wired counterparts, transporting data between network devices. However, there is one fundamental, and quite significant, difference: WLANs are based on radio communications technology as an alternative to structured wiring and cables. Data is transmitted between devices through the air by utilizing the radio waves. Devices that participate in a WLAN must have a network interface card (NIC) with wireless capabilities. This essentially means that the card contains a small radio device that allows it to communicate with other wireless devices within the defined range for that card, for example, the 2.4-2.4853 GHz range. For a device to participate in a wireless network, it must firstly be permitted to communicate with the devices in that network, and secondly it must be within the transmission range of the devices in that network. To communicate, radio-based devices take advantage of electromagnetic waves and their ability to be altered in such a manner that they can carry information, known as modulation (Sundaralingham, 2004). Here we discuss wireless security mechanisms.
are being discovered just as quickly as security measures are being released. Perhaps the issue that has received the most publicity is the major weaknesses in WEP, and more particularly the use of the RC4 algorithm and relatively short initialization vectors (IVs). WLANs suffer from all the security risks associated with their wired counterparts; however, they also introduce some unique risks of their own. The main issue with radio-based wireless networks is signal leakage. Due to the properties of radio transmissions, it is impossible to contain signals within one clearly defined area. In addition, because data is not enclosed within cable, it makes it very easy to intercept without being physically connected to the network (Hardjono & Lakshminath, 2005). This puts it outside the limits of what a user can physically control; signals can be received outside the building and even from streets away. Signal leakage may not be a huge priority when organizations are implementing their WLAN, but it can present a significant security issue, as demonstrated below. The signals that are transmitting data around an organization’s office are the same signals that can also be picked up from streets away by an unknown third party. This is what makes WLANs so vulnerable. Before WLANs became common, someone wishing to gain unauthorized access to a wired network had to physically attach themselves to a cable within the building. This is why wiring closets should be kept locked and secured. Any potential hacker had to take great risks to penetrate a wired network. Today potential hackers do not have to use extreme measures, there’s no need to smuggle equipment on site when it can be done from two streets away. It is not difficult for someone to obtain the necessary equipment; access can be gained in a very discrete manner from a distance.
BACKGROUND Wired networks have always presented their own security issues, but wireless networks introduce a whole new set of rules with their own unique security vulnerabilities. Most wired security measures are just not appropriate for application within a WLAN environment; this is mostly due to the complete change in transmission medium. However, some of the security implementations developed specifically for WLANs are also not terribly strong. Indeed, this aspect could be viewed as a work-in-progress; new vulnerabilities
WIRELESS SECURITY MECHANISMS To go some way towards providing the same level of security the cable provides in wired networks, the wired equivalent protocol (WEP) was developed. WEP was designed to provide the security of a wired LAN by encryption through use of the RC4 (Rivest Code 4) algorithm. Its primary function is to safeguard against eavesdropping (sniffing), by making the data that is transmitted unreadable by a third party who does
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Wireless Network Security
not have the correct WEP key to decrypt the data. RC4 is not specific to WEP, it is a random generator, also known as a key stream generator or a stream cipher, and was developed in RSA Laboratories by Ron Rivest in 1987 (hence the name Rivest Code). It takes a relatively short input and produces a somewhat longer output, called a pseudo-random key stream. This key stream is simply added modulo two that is exclusive ORed (XOR), with the data to be transmitted, to generate what is known as ciphertext (Briere, 2005). WEP is applied to all data above the 802.11b WLAN layers (physical and data link layers, the first two layers of the OSI reference model) to protect traffic such as transmission control protocol/Internet protocol (TCP/IP), Internet packet exchange (IPX), and hyper text transfer protocol (HTTP). It should be noted that only the frame body of data frames are encrypted, and the entire frame of other frame types are transmitted in the clear, unencrypted (Karygiannis & Owens, 2003). To add an additional integrity check, an initialization vector (IV) is used in conjunction with the secret encryption key. The IV is used to avoid encrypting multiple consecutive ciphertexts with the same key, and is usually 24 bits long. The shared key and the IV are fed into the RC4 algorithm to produce the key stream. This is XORed with the data to produce the ciphertext; the IV is then appended to the message. The IV of the incoming message is used to generate the key sequence necessary to decrypt the incoming message. The ciphertext, combined with the proper key sequence, yields the original plaintext and integrity check value (ICV) (Hardjono & Lakshminath, 2005). The decryption is verified by performing the integrity check algorithm on the recovered plaintext and comparing the output ICV to the ICV transmitted with the message. If it is in error, an indication is sent back to the sending station. The IV increases the key size, for example, a 104-bit WEP key with a 24-bit IV becomes a 128-bit RC4 key. In general, increasing the key size increases the security of a cryptographic technique. Research has shown that key sizes of greater than 80 bits make brute force1 code breaking extremely difficult. For an 80-bit key, the number of possible keys10^24, which puts computing power to the test; but this type of computing power is not beyond the reach of most hackers. The standard key in use today is 64 bit. However, research has shown that the WEP approach to privacy is vulnerable to certain attacks regardless of key size (Karygiannes & Owens, 2003). Although the application of WEP may stop casual sniffers, determined hackers can crack WEP keys in a busy network within a relatively short period of time.
WEP’s Weaknesses When WEP is enabled in accordance with the 802.11b standard, the network administrator must personally visit each wireless device in use and manually enter the appropriate WEP key. This may be acceptable at the installation stage of
a WLAN or when a new client joins the network, but if the key becomes compromised and there is a loss of security, the key must be changed. This may not be a huge issue in a small organization with only a few users, but it can be impractical in large corporations, which typically have hundreds of users (Gavrilenko, 2004). As a consequence, potentially hundreds of users and devices could be using the same, identical key for long periods of time. All wireless network traffic from all users will be encrypted using the same key; this makes it a lot easier for someone listening to traffic to crack the key, as there are so many packets being transmitted using the same key. Unfortunately, there were no key management provisions in the original WEP protocol. A 24-bit initialization vector WEP is also appended to the shared key. WEP uses this combined key and IV to generate the RC4 key schedule; it selects a new IV for each packet, so each packet can have a different key (Walker, 2002). Mathematically there are only 16,777,216 possible values for the IV. This may seem like a huge number, but given that it takes so many packets to transmit useful data, 16 million packets can easily go by in hours on a heavily used network. Eventually the RC4 algorithm starts using the same IVs over and over. Thus, someone passively listening to encrypted traffic and picking out the repeating IVs can begin to deduce what the WEP key is. Made easier by the fact that there is a static variable (the shared key), an attacker can eventually crack the WEP key (Nakhjiri, 2005). For example, a busy AP, which constantly sends 1,500 byte packets at 11Mbps, will exhaust the space of IVs after 1,500 x 8/(11 x 10^6) x 2^24 = 18,000 seconds, or 5 hours. (The amount of time may actually be smaller since many packets are less than 1,500 bytes). This allows an attacker to collect two ciphertexts that are encrypted with the same key stream. This reveals information about both messages. By XORing, two ciphertexts that use the same key stream would cause the key stream to be cancelled out and the result would be the XOR of the two plaintexts (Vines, 2002).
War-Driving So called war-driving is a term used to describe a hacker whoarmed with a laptop, a wireless NIC, an antenna, and sometimes a GPS devicetravels, usually by car, scanning or sniffing for WLAN devices, or more specifically unprotected or open and easily accessed networks. The name is thought to have come from another hacking technique called war-dialing, where a hacker programs a system to call hundreds of phone numbers in search of a poorly protected computer dial-up (Nakhjiri, 2005). Due to the increased use of WLANs in recent years, it is quite possible that the number of unsecured devices has also risen in tandem, thus providing potential hackers with more choice. After all that has been written about the insecurities of WLAN, some users/organizations still insist on implementing them with their 1023
W
Wireless Network Security
default settings and no encryption (Ulanoff, 2003). There is a plethora of hacking tools widely available to download from the Internet for any potential war-driver to use. There has been a lot of press globally, and many articles and papers written about wireless networks and their security vulnerabilities. However, despite all the literature, some enterprises still make the mistake of believing that they do not have to worry about wireless security if they are running non-critical systems with non-sensitive information across their WLANs. All information is sensitive information, and what an enterprise may class as being non-sensitive to them may be very useful to a hacker. In addition, most WLANs will connect with the wired enterprise backbone at some point, thus providing hackers with a launch pad to the entire network. The havoc an unwelcome third party could cause from here would be unlimited and very difficult to trace. Aside from the various attacks they could instigate (DoS and viruses), the loss of confidentiality, privacy, and integrity that would occur if someone where able to steal, alter, or delete information on your customer database is damaging enough. Access to sensitive information would be made relatively easy, perhaps even customers’ credit card details. This could have an un-quantifiable effect on business, perhaps resulting in the loss of customers/clients and future revenue (AirDefense, 2003).
Wireless Attack Methods A passive attack is an attack on a system that does not result in a change to the system in any way; the attack is purely to monitor or record data. Passive attacks affect confidentiality, but not necessarily authentication or integrity. Eavesdropping and traffic analysis fall under this category. When an attacker eavesdrops, he or she simply monitors transmissions for message content. It usually takes the form of someone listening into the transmissions on a LAN between stations/devices. Eavesdropping is also known as sniffing or wireless footprinting. There are various tools available for download online which allow the monitoring of networks and their traffic; these are developed by hackers, for hackers. Netstumbler, Kismet, Airsnort, WEPCrack, and Ethereal are all well-known names in wireless hacking circles, and all are designed specifically for use on wireless networks, with the exception of Ethereal, which is a packet analyzer and can also be used on a wired LAN. NetStumbler and Kismet can be used purely for passive eavesdropping; they have no additional active functions, except perhaps their ability to work in conjunction with global positioning systems (GPSs) to map the exact locations of identified wireless LANs. NetStumbler is a Windows-based sniffer, where Kismet is primarily a Linux-based tool. NetStumbler uses an 802.11 Probe Request sent to the broadcast destination address, which causes all APs in the area to issue an 802.11 Probe Response 1024
containing network configuration information, such as their SSID, WEP status, the MAC address of the device, name (if applicable), the channel the device is transmitting on, the vendor and the type, either peer or AP, along with a few other pieces of information. Using the network information and GPS data collected, it is then possible to create maps with tools such as StumbVerter and MS Mappoint. Kismet, although not as graphical or user friendly as NetStumbler, is similar to its Windows counterpart, but it provides superior functionality. While scanning for APs, packets can also be logged for later analysis. Logging features allow for captured packets to be stored in separate categories, depending upon the type of traffic captured. Kismet can even store encrypted packets that use weak keys separately to run them through a WEP key cracker after capture, such as Airsnort or WEPCrack (Sundaralingham, 2005). Wireless network GPS information can be uploaded to a site called Wigle (http://www.wigle.net). Therefore, if Wigle data exists for a particular area, there is no need to drive around that area probing for wireless devices; this information can be obtained in advance from the Wigle Web site. All that remains is to drive to a location where known networks exist to observe traffic. Wigle currently has a few hundred thousand networks on its database. Traffic analysis gains intelligence in a more subtle way by monitoring transmissions for patterns of communication. A considerable amount of information is contained in the flow of messages between communicating parties. Airopeek NX, a commercial 802.11 monitoring and analysis tool for Windows, analyzes transmissions and provides a useful node view, which groups detected stations and devices by their MAC address and will also show IP addresses and protocols observed for each. The Peer Map view, within Airopeek NX, presents a matrix of all hosts discovered on the network by their connections to each other. This can make it very easy to visualize AP and client relationships, which could be useful to hackers in deciding where to try and gain access or target for an attack (McClure, Scambray, & Jurtz, 2003). Some attacks may begin as passive, but then crossover to active as they progress. For example, tools such as Airsnort or WEPCrack may passively monitor transmissions, but their intent is to crack the WEP key used to encrypt data being transmitted. Figure 1 shows a screen shot of Airsnort showing the number of packets gathered and the WEP key17BE311175. Ultimately the reasons for wanting to crack the key are so that an unauthorized individual can access a protected network and then launch an active attack of some form or another. These types of attacks are classed as passive decryption attacks. An active attack, also referred to as a malicious attack, occurs when an unauthorized third party gains access to a network and proceeds to perform denial of service (DoS) attack, to disrupt the proper operation of a network, to intercept network traffic and either modify or delete it, or
Wireless Network Security
Figure 1. Screen shot from Airsnort showing 64-bit key crack
inject extra traffic onto the network. There are many active attacks that can be launched against wireless networks; the following few paragraphs outline almost all of these attacks, how they work, and what effect they have (Karygiannis & Owens, 2003). DoS attacks are easily the most prevalent type of attack against 802.11 networks and can be waged against a single client or an entire WLAN. In this type of attack, the hacker usually does not steal information; he or she simply prevents users from accessing network services, or causes services to be interrupted or delayed. Consequences can range from a measurable reduction in performance to the complete failure of the system. Some common DoS attacks are outlined below. A man-in-the-middle attack is carried out by inserting a malicious station between the victim station and the AP, thus the attacker becomes the man in the middle; the station is tricked into believing that the attacker is the AP, and the AP into believing that the attacker is the legitimate station. To begin the attack, the perpetrator passively monitors the frames sent back and forth between the station and the AP during the initial association process with an 802.11 analyzer. As a result, information is obtained about both the station and the AP, such as the MAC and IP address of both devices, association ID for the station, and SSID of the network. With this information a rogue station/AP can be set up between the two unsuspecting devices. Because the original 802.11 does not provide mutual authentication, a station will happily re-associate with the rogue AP. The rogue AP will then capture traffic from unsuspecting users; this of course can expose information such as user names and passwords (Gavrilenko, 2004). An Association flood is a resource starvation attack. When a station associates with an AP, the AP issues an associate identification (AID) number to the station in the range of 1-2007. This value is used for communicating power management information to a station that has been in a power-save state. This attack works by sending multiple authentication and association requests to the AP, each with a unique source MAC address. The AP is unable to differentiate the authentication requests generated by an attacker and those created by legitimate clients, so it is forced
W
to process each request. Eventually, the AP will run out of AIDs to allocate and will be forced to de-associate stations to reuse previously allocated AIDs. In practice, many APs will restart after a few minutes of authentication flooding, however this attack is effective in bringing down entire networks or network segments; if repeatedly carried out, it can cause a noticeable decrease in network up time (Wright, 2003). The final issue is a threat posed by the simple network management protocol (SNMP).
Attacks that Alter Transmissions The following attacks describe how it is possible for an attacker to modify messages in transit, without detection. Message modification attacks are made relatively trivial if no message encryption exists; however, even if it does, the hacker can still get around it by first cracking the encryption and then carrying out the attack. •
•
Injecting Traffic: If an attacker knows the exact plaintext for one encrypted message, he or she can then use this knowledge to construct more correctly encrypted packets. This procedure involves constructing a new message, calculating the CRC-32 checksum, and performing bit-flips2 on the original encrypted message to change the plaintext to the new message. This packet can now be sent to the AP, and it will be accepted as a valid packet. Because RC4 encrypts data a byte at a time, an attacker can modify one byte of ciphertext and the recipient would not know the data has been changed. RC4 does not detect errors (Borisov, Goldberg, & Wagner, 2003). IP Redirection: By intercepting and modifying the IP address of the destination in a packet, an attacker can effectively re-route messages. This attack can be used where an AP acts as an IP router with Internet connectivity, which is fairly common. The idea is to take an encrypted packet that has been transmitted and modify it so it has a new destination addressone the attacker controls. The AP will then decrypt the packet and send it off to its new destination, where the attacker 1025
Wireless Network Security
•
can read the packet, now in the clear (Borisov et al., 2003). SNMP Attack: The final issue is a threat posed by the simple network management protocol (SNMP). Some APs can be managed via wireless link, usually with a proprietary application, replying on SNMP. Executing these operations can represent a frightening vulnerability for the whole LAN; because eavesdroppers can decipher the password to access read/write mode on the AP using a packet analyzer, this means that they share the same administration privileges with the WLAN administrator and can manage the WLAN in a malicious manner (Me, 2003). The sheer number of attacks, and their effects, would seem to put WLANs at a severe disadvantage over their wired counterparts. However, there are just as many, if not more, security measures that users can utilize to counteract most of the above attacks. Layering one security measure on top of another, to strengthen the overall system to deter any potential attackers or make their task more difficult, is not impossible. However, not all organizations, or indeed individuals, take the time to implement any form of security, or they implement it weakly. The next article discusses, firstly, what route the primary research will take, and secondly, the findings resulting from the primary research.
FUTURE TRENDS IEEE specifies basically two categories of WLAN standardsthose that specify the fundamental protocols for the complete wireless system and those that address specific weaknesses or provide additional functionality (3COM, 2006). Here we mention just three of the latter standards which may have a significant influence in the coming days. 802.11i is a major extension because it was intended to improve WLAN security on 802.11a and 802.11b networks, which was in tatters. It adds two main blocks of improvements: improved security for data in transit, and better control of who can use a network. It covers key management and distribution, encryption, and authentication, the three main components of security (Briere, 2005). The 802.11i specification can be viewed as consisting of three main sections, organized into two layers. On the lower level are improved encryption algorithms in the form of the temporal key integrity protocol (TKIP) and the counter mode with cipher block chaining-message authentication code, CBCMAC protocol (CCMP). Both of these provide enhanced data integrity over WEP, with TKIP being targeted at legacy equipment and CCMP being targeted at future WLAN equipment (Nakhjiri, 2005). The goal of the 802.11k standard is to make measurements from layers one and two of the OSI protocol stackphysical 1026
and data link layersavailable to the upper layers. It is expected that the upper layers will then be able to make decisions about the radio environment. It is called radio resource management. One feature is better traffic distribution. Normally a wireless device will connect to whatever AP gives it the strongest signal. However, this can lead to an overload on some APs and under-load on others, resulting in an overall lowered service level. The 802.11k standard will allow network management software to detect this situation and redirect some of the users to under-utilized APs. 802.11n is a high-performance standard that would boost both 802.11b and 802.11a 11Mbps and 54Mbps, respectively. Proposals say it could go to 108Mbps or beyond, to as much as 320Mpbs. This standard is not expected to be complete until early 2007. Community 802.11b networks will continue to grow as people realize they can share their high-speed, high-cost Internet connections, turning them into high-speed, lowor no-cost connections for a larger group of people (Imai, 2005).
CONCLUSION Wireless networks have a number of security issues. Signal leakage means that network communications can be picked up outside the physical boundaries of the building in which they are being operated, meaning a hacker can operate from the street outside or discretely from blocks away. In addition to signal leakage, wireless networks have various other weaknesses. WEP, the protocol used within WLANs to provide the equivalent security of wired networks, is inherently weak. The use of the RC4 algorithm and weak IVs makes WEP a vulnerable security measure. In addition to WEP’s weaknesses, there are various other attacks that can be initiated against WLANs, all with detrimental effects.
REFERENCES AirDefense. (2003) Wireless LAN security: What hackers know that you don’t. Retrieved from http://ssl.salesforce. com/servlet.Email/AttachmentDownload?q=00m0000000 003Pr00D00000000hiyd00500000005k8d5 Borisov, N. Goldberg, I., & Wagner, D. (2003) Security of the WEP algorithm. Retrieved from http://www.isaac. cs.berkeley.edu/isaac/wep-faq.html Briere, D. (2005, October) Wireless network hacks and mods for dummies (for dummies S.). New York: Hungry Minds. Gavrilenko, K. (2004, June) WI-FOO: The secrets of wireless hacking. Boston: Addison-Wesley.
Wireless Network Security
Hardjono, T., & Lakshminath R. D. (2005, July). Security in wireless LANs and MANs. Norwood, MA: Artech House. Harte, L., Kellog, S., Dreher, R., & Schaffinit, T. (2000) The comprehensive guide to wireless technologies: Cellular, PCS, paging, SMR and satellite. NC: APDG. Imai, H. (Ed.). (2005, December). Wireless communications security. Norwood, MA: Artech House. Karygiannis, T., & Owens, L. (2003). National Institute of Standards and Technology, special publication 800-48, draft. Retrieved from http://csrc.nist.gov/publications/drafts/draftsp800-48.pdf McClure, S., Scambray, J., & Jurtz, G. (2003). Hacking exposed: Network security secrets and solutions (4th ed.). New York: Osbourne McGraw-Hill. Me, G. (2003). A threat posed by SNMP use over WLAN. Retrieved from http://www.wi-fitechnology.com/Wi-Fi_Reports_and_Papers/SNMP_use_over_WLAN.html Nakhjiri, M. (2005, September). AAA and network security for mobile access: Radius, diameter, EAP, PKI and IP mobility. New York: John Wiley & Sons. Sundaralingham, S. (2004, November). Cisco wireless LAN security. San Francisco: Cisco Press. 3COM. (2006). Retrieved from http://www.3com.com/ whitepapers.html Ulanoff, L. (2003). Get free Wi-Fi, while its hot. PC Magazine, (July). Vines, R. D. (2002). Wireless security essentials, defending mobile systems from data piracy. London: John Wiley & Sons. Walker, J. (2002) Unsafe at any key size; an analysis of the WEP encapsulation. Retrieved from http://www.dis. org/wl/pdf/unsafe.pdf Wright, J. (2003). Detecting wireless LAN MAC address spoofing. Retrieved from http://home.jwu.edu/jwright/papers/wlan-mac-spoof.pdf
KEY TERMS Denial of Service: An incident in which a user or organization is deprived of the services of a resource they would normally expect to have. Typically, the loss of service is the inability of a particular network service to be available or the temporary loss of all network connectivity and services.
Direct Sequence Spread Spectrum (DSSS): Combines a data signal with a higher data rate bit sequence, referred to as a chipping code. The data is exclusive ORed (XOR) with a PRS that results in a higher bit rate, This increases the signal’s resistance to interference. Frequency Hopping Spread Spectrum (FHSS): Here the signal hops from frequency to frequency over a wide band of frequencies. The transmitter and receiver change the frequency they operate on in accordance with a pseudo-random sequence (PRS) of numbers. To properly communicate, both devices must be set to the same hopping code. IEEE 802.11 Standard: IEEE has developed several specifications for WLAN technology, the names of which resemble the alphabet. There are basically two categories of standards: those that specify the fundamental protocols for the complete wireless systemthese are called 802.11a, 802.11b, and 802.11g; and those that address specific weaknesses or provide additional functionalitythese are 802.11d, e, f, h, I, j, k, m, and n. Initialization Vector (IV): Used in conjunction with the secret encryption key in WEP (see Wired Equivalent Privacy). The IV is used to avoid encrypting multiple consecutive ciphertexts with the same key, and is usually 24 bits long. War Driving: A term used to describe a hacker whoarmed with a laptop, a wireless NIC, an antenna, and sometimes a GPS devicetravels, usually by car, scanning or sniffing for WLAN devices, or more specifically unprotected or open and easily accessed networks. Wired Equivalent Privacy (WEP): Designed to provide the security of a wired LAN by encryption through use of the RC4 (Rivest Code 4) algorithm. Its primary function is to safeguard against eavesdropping (sniffing), by making the data that is transmitted unreadable by a third party who does not have the correct WEP key to decrypt the data.
ENDNOTES 1
2
A method that relies on sheer computing power to try all possibilities until the solution to a problem is found; usually refers to cracking passwords by trying every possible combination of a particular key space. Bit-flippingchanging one or more bits within a message. For example, change a 0 to a 1, or vice versa.
1027
W
1028 Category: Security
Wireless Security Meletis Belsis Telecron, Greece Alkis Simitsis National Technical University of Athens, Greece Stefanos Gritzalis University of the Aegean, Greece
INTRODUCTION The fast growth of wireless technology has exponentially increased the abilities and possibilities of computing equipment. Corporate users can now move around enterprise buildings with their laptops, PDAs, and WiFi; enable VoIP handsets; and retain communications with their offices. Business users can work from almost anywhere by attaching their laptops to WiFi hotspots and connecting to their corporate network. However, not many enterprises know and understand the potential security vulnerabilities that are introduced by the use of WiFi technologies. Wireless technologies are insecure by their nature. Anyone with the appropriate hardware can steal information transmitted using the airwaves. This article discusses the security vulnerabilities that are inherited in wireless networks. Also, it provides a description of the current security trends and protocols used to secure such WiFi networks along with the problems from their application.
BACKGROUND Currently, several enterprises consider information security as a monolithic architecture, in which simply they install a firewall or an intrusion detection system. Unfortunately security is not a single device or software: In the real world, security involves processes. It involves preventive technologies, but also detection and reaction processes, and an entire forensics system to hunt down and prosecute the guilty. Security is not a product; it itself is a process. (Schneier, 2000) The above definition represents the fact that total protection of corporate networks goes beyond a firewall engine. Each appliance that is added and/or changed into a system should incorporate the re-designing of a system’s overall
security policy and infrastructure. The same principle exists when incorporating wireless devices to extend the overall enterprise architecture. Deploying a wireless network has as a consequence the change of the security risks and needs of the entire network infrastructure. Nowadays, the techniques that are used for the realization of attacks in wireless connected networks resemble those used to target common LANs. In the next paragraphs, we present the major categories of attacks, including techniques that have been successfully used for attacking corporate wireless networks. Denial of Service. In their simplest form, an adversary can continuously transmit association request packets. Such action could render an access point unavailable to authorized users. Adversaries can use a powerful RF transceiver to transmit amplified signals in all frequency band frequencies (channels), creating an interjection that prevents the communication of terminals with the corporate access points (RF Jamming). Such an attack could be easily deployed from the outside premises of an enterprise (e.g., parking). An example appliance that can be used for the concretization of this attack is the Power Signal Generator (PSG -1) by the YDI. Man-in-the-Middle Attacks. Combining an RF Jamming attack with the use of a portable computer and necessary software, an attacker can easily steal or alter corporate information (Akin, 2003). The adversary will use a denialof-service attack to force authorized terminals connected to a corporate access point to identify and roam to an access point with better signal than the one already connected to. Using this predetermined behavior the attacker can masquerade his/her laptop as an access point and force all wireless clients to connect to it. By using this technique an adversary can intercept all wireless communications links and read or alter information on them. Fresnel Zone Sniffing. Stealing information from pointto-point wireless links is difficult. The attacker needs to calculate the link path and identify ways to attach its laptop to the link’s Fresnel Zone. Rogue Wireless Gateways. A rogue wireless gateway is a security vulnerability that is detected in many of today’s
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Wireless Security
enterprise networks. A rogue wireless gateway is an unauthorized access point that is installed on an enterprise network. Such access points are usually installed by corporate users, to assist them in the everyday work (i.e., transfer files/emails from a desktop to a laptop computer). Unfortunately enterprise users do not know and understand the security implications of installing a wireless device on a system. Leaving such devices connected to the corporate network provides an opportunity to adversaries to connect and steal corporate information. Ad Hoc Networks. The 802.11 protocol specification allows wireless terminals to interconnect without the use of an access point. This mode of operation is called ad hoc. Unfortunately many of today’s corporate users enable the ad hoc facility on their laptops and PDAs either accidentally or deliberately in order to exchange files with other users. Enabling the ad hoc mode without deploying the necessary security procedures (i.e., encryption and authentication) could seriously damage corporate security. Adversaries can search for such unprotected ad hoc networks and connect to those. From there adversaries can either read the locally stored corporate information, or if the user’s device is connected to the corporate networks (i.e., LAN, dialup, and VPN), access the corporate resources (Papadimitratos & Haas, 2002). The previous example attacks emphasize the need for security that results from the use of wireless technology. The problem of security becomes more apparent when the technology of wireless networking is applied in governmentowned systems. The need for security in those systems is extensive due to the legislation on personal data protection and the human lives factors involved.
MAIN THRUST OF THE ARTICLE In the last few years, the computing and telecommunications community has realized the necessity of deploying security controls on wireless networks. Unfortunately most of today’s wireless security controls have been proven unsafe or managerial infeasible to maintain. The next few paragraphs describe the most common security protocols and techniques, as well as their vulnerabilities.
Figure 1. A War Driving result in Los Angeles
W
adversary uses a laptop computer, along with appropriate discovery software (i.e., NetStumbler) and a GPS received, to pinpoint the exact location of access points on a map. Today such maps are distributed among the War Driving community. It is not unusual for enterprises to discover their company access points on maps found on War Driving Web sites (see Figure 1). Many enterprise administrators try to hide their wireless networks by activating the close system option found on access point hardware equipment. This option prohibits the access point from transmitting the network’s beacon information that incorporates the network’s service set identifier (SSID). Unfortunately the SSID is incorporated into almost all network management frames. Software packages like Table 1. War Chalking symbols node
symbol SSID
open node bandwidth
Discovering Wireless Networks Many enterprises support their notion of using insecure WiFi networks based on the idea that their small wireless networks are hidden from hackers and adversaries. This notion is called Security through Obscurity, and is something that the IT security community analyzed and abolished long before the appearance of wireless networks. Modern hackers have invented a number of new techniques, collectively known as War Driving or War Chalking, which aim at discovering unprotected wireless networks. An
SSID
closed node
SSID
WEP node
W
access contact
bandwidth
1029
Wireless Security
NetStumbler will force the access points in transmitting the SSID by issuing such management frames (i.e., Reassociation Request). The techniques of War Driving and War Chalking has been used today to an extended degree, and adversaries have developed their own marking symbols (see Table 1) in order to denote the buildings where wireless networks are discovered. Writing these symbols in various buildings of the city, adversaries mark their potential targets.
MAC Access Control Lists To enhance security, many corporations develop media access control (MAC) control lists declaring the MAC addresses of wireless terminals that are authorized to access the wired segment of a corporate network. Unfortunately the deployment of MAC access control lists increases management time and difficulty without offering real protection from experienced hackers. Having discovered a wireless network, an adversary can eavesdrop on the network and detect authorized MAC addresses that connect to an access point. Having a list of such authorized MAC addresses, the adversary can use MAC spoofing attacks and masquerade his laptop as an authorized client (e.g., using the SMAC software, a snapshot of which is depicted in Figure 2).
Wired Equivalent Privacy (WEP) The first security protocol developed for wireless networks is wired equivalent privacy (WEP). WEP uses RC4 PRNG algorithm (LAN MAN, 1999) for the coding of information. The WEP key, with a 24-bit initializing vector (IV), is used for the encryption/decryption of wireless data. The protocol works with keys of 64 or with 128 bit (the actual key lengths are 40 and 104 bit, but are concatenated with the IV during Figure 2. SMAC software screenshot
the encryption phase). In a WEP environment the encryption keys are installed by the administrator of the system in each terminal and access point, and thus the management of the network becomes more complicated. The WEP does not offer user authentication; therefore, discovering the WEP key allows access to a corporate network (Borisov, Goldberg, & Wagner, 2001). The two authentication models provided by WEP are open system and the shared-key authentication (Lambrinoudakis & Gritzalis, 2005). The open system model uses the MAC access control lists discussed in the previous paragraphs. In the shared key authentication, WEP uses the encryption key to implement a Challenge-Response authentication scheme. At the same time WEP uses a 32-bit cycle redundancy check algorithm as integrity check value (ICV) in order to ensure the integrity of data. Currently, the CRC algorithm has already been broken by researchers from the University of Berkley (Tyrrell, 2003). The key recovery process in a system that uses WEP can be actually realized in a few hours. This is due to a vulnerability found in the way WEP uses the RC4 algorithm. The weakness of WEP is based on the fact that the IV is only 24 bit, and thus, in a busy network, the same IV key is used to encrypt different network packets. Having eavesdropped two or more packets encrypted with the same IV, an adversary can apply cryptanalysis techniques and recover the WEP key. Today, a number of freeware software packages that can perform a successful WEP attack are available on the Internet. Examples of such software artifacts include WEPCrack and AIRSnort (see Figure 3). Due to the fact that WEP encryption keys are static, the time between discovering a compromised key and updating the whole wireless network infrastructure with a new key is extended. This leaves even more time for adversaries to access and copy confidential corporate information.
WiFi-Protected Access (WPA) Understanding the problems of WEP, the international community has moved forward in developing a more secure protocol, namely 802.11i (Edney & William, 2003). Due to the delay in the development of the final 802.11i standard, the international community released a pre-802.11i security Figure 3. AirSnort software screenshot
1030
Wireless Security
Figure 4. 802.1x EAP authentication (EAP Authentication, 2005)
•
•
protocol under the name WiFi-protected access (Edney & William, 2003). The WPA uses algorithm RC 4 (Fluhrer, Mantin, & Shamir, 2001) for the encryption of air data incorporating the temporal key integrity protocol (TKIP), in order to use dynamic encryption keys. In order to avoid the security vulnerabilities of CRC-32, WPA utilizes a novel integrity protection algorithm, the Michael Message Integrity Check (MIC) (Cam-Winget, Housley, Wagner, & Walker, 2003), which uses a 64-bit key and partitions data into 32-bit blocks. TKIP uses an IV of 48 bit, offering better security than the 24-bit IV used by WEP. It combines a 128-bit temporary key, which is preinstalled in all wireless terminals with the MAC address of each terminal, and the 48-bit IV in order to create a new encryption key for each terminal. The protocol changes the encryption key every 10,000 packets that are transmitted. Moreover, WPA employs the 802.1x protocol (port-based access control) to deliver authenticated connections. This protocol allows the usage of a number of authentication methods to be used such as passwords and digital certificates. The user or terminal authentication process is performed by the extensible authentication protocol (EAP), which is usually associated with a radius server in order to securely authenticate users or devices on a network. Figure 4 displays an example EAP authentication process. Currently, there exist several EAP implementations: •
EAP-MD 5 (Funk, 2003): This was the first protocol to use user authentication based on the 802.1x scheme. It provides only one-way authentication, ensuring the authenticity of users but not the servers. The protocol is based on the algorithm MD5. However, research has already proven that this protocol is subject to diction-
•
•
ary and man-in-the-middle attacks (Asokan, Niemi, & Nyberg, 2002). CISCO-LEAP: The lightweight EAP (LEAP), created by CISCO, offers bidirectional authentication. The bidirectional authentication makes the protocol immune to man-in-the-middle attacks, but its challenge handshake authentication protocol (MSCHAP version 2) is subject to dictionary attacks. Currently, there exist several tools on the Internet, like the asleep, that can perform successful attacks on LEAP. CISCO tries to tackle this disadvantage, and at this time, the company is developing a new protocol called EAP-FAST. EAP-FAST (Ghosh & Gupta, 2005): EAP-FAST, developed and marketed by CISCO, is thought to be as secure as EAP-PEAP and as easy to deploy as EAPLEAP. The protocol operates similarly to EAP-PEAP. It uses two distinct phases. In phase 1 a secure tunnel is established using a Protected Access Credential (PAC) shared key. PAC is used in order to avoid deploying digital certificates. After the establishment of the secure tunnel, authentication is performed on phase 2 using the MSCHAP v2 protocol. The PAC secret can either be manually shared to all nodes or automated through an optional Diffie-Hellman process. Unfortunately, using the manual shared key distribution process will make the management of the network extremely difficult. On the other hand the anonymous Diffie-Hellman process can make the protocol suspect to man-in-the-middle attacks. Along with this, during the anonymous Diffie-Hellman, the protocol transmits the user name in cleartext (unencrypted), and thus possession of a user name could further lead an attacker to perform social engineering attacks. It is going to be a while before the protocol is thoroughly tested and used by the international community (Lambrinoudakis & Gritzalis, 2005). EAP-TLS (Aboba & Simon, 1999): The EAP-transport layer security (EAP-TLS), developed by Microsoft Corporation, uses the Transport Layer Security (TLS) protocol with digital certificates for both clients and servers in order to provide bidirectional authentication. The protocol transmits the user name in cleartext. A possible information leakage in this form could provide the basis for further attacks (i.e., social engineering). Along with this, the use of both client and server certificates makes the management of this protocol a hassle for large corporate networks. EAP-TTLS (Funk & Blake-Wilson, 2003): The EAPtunneled TLS (EAP-TTLS) protocol, created by the companies Funk and Certicom, is based on the idea of EAP-TLS, but in order to minimize the management process, it uses their digital certificates only for the servers and not for the clients. Clients authenticate servers by using digital certificates; thus, the proto1031
W
Wireless Security
•
col builds an encrypted tunnel. The encrypted tunnel provides a secure medium on which clients can be authenticated using a challenge response mechanism. Although, currently, there are no known attacks, the protocol is suspected to be vulnerable to man-in-themiddle attacks (Asokan et al., 2002). EAP-PEAP (Palekar et al., 2003): The protected EAP (PEAP) protocol is the result of a common effort from different IT companies. PEAP uses digital certificates for servers. Also, clients authenticate servers. After a successful server authentication, the protocol creates an encrypted tunnel between the client and the server. Inside this secure tunnel, the system can use any of the previously described EAP authentication methods in order to enable client authentication. The chosen combination today is to use the EAP-TLS inside the encrypted tunnel in order to provide client authentication (EAP-PEAP/EAP-TLS). Similar to the TTLS protocol, no known attack exists today, but PEAP is suspected to be vulnerable to man-in-the-middle attacks.
802.11I Having discovered the vulnerabilities in WEP, IEEE started producing the specification of a new protocol, IEEE 802.11i. It follows principles similar to WPA, and uses 802.1x and EAP protocols for authentication and key management. 802.11i uses the counter-mode/CBC-MAC protocol (CCMP) protocol with the advance encryption standard (AES) (NIST, 2001) algorithm to provide data encryption and integrity protection. In addition, 802.11i provides the robust security network (RSN) feature. RSN allows the two ends of a communication link to negotiate the encryption algorithms and protocols to be used. This facility enables updating a wireless network with new algorithms and protocols in order to protect it from future vulnerabilities. Still, the 802.11i protocol requires special encryption hardware to run the AES algorithm; due to this fact, additional time is needed for the vendors to change their existing hardware to support the 802.11i protocol. To enable the migration of WEP and WPA systems to 802.11i, the WiFi Alliance has proposed a new security protocolthe WPA2. The new protocol incorporates all 802.11i functionality, but also enables the use of the TKIP protocol to support devices that do not have the necessary hardware to run the AES algorithm.
VPNs To provide a solution to the problem of security, many companies are extending/developing virtual private networks (VPNs) (Karygiannis & Owens, 2002). Maintaining a VPN requires the engagement of specialized personnel or the training of existing personnel; in both cases, the costs associated with deploying a wireless infrastructure are highly increased. Along with the cost associated with the deployment of a VPN, VPNs incorporate a number of operational problems on a system. In networks where the users roam contentiously, a Layer-3 VPN solution will disrupt a user’s connection and may even force the user to re-authenticate. Along with this, applications that run on client terminals and access data stored on the corporate servers may be seriously disrupted from a Layer-3 disconnection. Such disconnections can seriously damage the integrity and availability of corporate information.
CONCLUSION In this article, we have discussed the critical issue of wireless security. We have presented the security vulnerabilities that are frequently inherited in wireless networks. Also, we have described the most common security protocols and techniques used. Moreover, we have provided a description of the current security trends and protocols used to secure such WiFi networks, along with the problems from their application.
REFERENCES Aboba, B., & Simon, D. (1999). PPP EAP TLS authentication protocol. IETF RFC 2716. Akin, D. (2003). Certified Wireless Security Professional (CWSP) official study guide. New York: McGraw-Hill. Asokan, N., Niemi, V., & Nyberg, K. (2002). Man-in-themiddle in tunneled authentication protocols. Cryptology ePrint Archive, Report 2002/163. Borisov, N., Goldberg, I., & Wagner, D. (2001). Intercepting mobile communications: The insecurity of 802.11. Retrieved December 16, 2005, from http://www.isaac.cs.berkeley. edu/isaac/mobicom.pdf Cam-Winget, N., Housley, H., Wagner, D., & Walker, J. (2003). Security flaws in 802.11 data link protocols. Communications of the ACM, 46(5). EAP Authentication. (2005). Retrieved December 13, 2005, from http://www.wi-fiplanet.com
1032
Wireless Security
Edney, J., & William, A. (2003). Real 802.11 security: WiFi protected access and 802.11i. Boston: Addison-Wesley. Fluhrer, S., Mantin, I., & Shamir, A. (2001). Weaknesses in the key scheduling algorithm of RC4. Proceedings of the 8th Annual Workshop on Selected Areas in Cryptography. Berlin: Springer-Verlag (LNCS 2259). Funk, P. (2003). The EAP MD5-Tunneled authentication protocol (EAP-MD5-Tunneled). IETF Internet Draft. Funk, P., & Blake-Wilson, S. (2003). EAP Tunneled TLS authentication protocol (EAP-TTLS). IETF Internet Draft. Ghosh, D., & Gupta, A. (2005). Analysis of EAP-FAST wireless security protocol. Retrieved December 15, 2005, from http://wwwcsif.cs.ucdavis.edu/~guptaa/finalreport.pdf Karygiannis, T., & Owens, L. (2002). Wireless network security. NIST Special Publication 800-48. Lambrinoudakis, C., & Gritzalis, S. (2005). Security in IEEE 802.11 WLANS. CRC Press. LAN MAN. (1999). Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications. IEEE Standard 802.11, 1999 edition. Standards Committee of the IEEE Computer Society. NIST. (2001). Announcing the Advance Encryption Standard (AES). Federal Information Processing Standards Publication 197. Palekar, A., Simon, D., Zorn, G., Salowey, J., Zhou, H., & Josefsson, S. (2003). Protected EAP Protocol (PEAP) version 2. IETF Internet Draft.
KEY TERMS Encrypted Tunnel: An encrypted logical (virtual) connection between two ends. Data traveling inside the tunnel are encrypted with an agreed encryption algorithm. Fresnel Zone: The area around the visual line of sight of a wireless link on which the RF waves are spread. This area must be clear from obstacles, otherwise the RF signal is weakened. Man-in-the-Middle Attack: An attack where the adversary succeeds in locating himself in an established connection between two or more authorized nodes. Data traveling between the nodes are always passing from the adversary. Reassociation Request Frame: A data packet transmitted in a wireless network. The packet enables a client to reconnect to an access point. The packet is transmitted after a client disconnection or when a client roams from one access point to another. Virtual Private Network (VPN): A set of technologies and protocols used to establish encrypted tunnels between one or more network nodes. WiFi Alliance: A nonprofit organization, with more than 200 members, devoted to promoting the use and operation of wireless networks. Products associated by the WiFi Alliance are able to interoperate. Wireless Computer Network: Any computer network that uses wireless technologies based on the IEEE 802.11x standards to transmit and receive data.
Papadimitratos, P., & Haas, Z.J. (2002). Secure routing for mobile ad hoc networks. Working session on security in wireless ad hoc networks, EPFL. Mobile Computing and Communications Review, 6(4). Schneier, B. (2000). Secret and lies (1st ed.). New York: John Wiley & Sons. Tyrrell, K. (2003). An overview of wireless security issues. SANS Information Security Reading Room, SANS Institute.
1033
W
1034 Category: Wireless Networking
Wireless Sensor Networks Antonio G. Ruzzelli University College Dublin, Ireland Richard Tynan University College Dublin, Ireland Michael J. O’Grady University College Dublin, Ireland Gregory M. P. O’Hare University College Dublin, Ireland
INTRODUCTION The origins of networks of sensors can be traced back to the 1980s when DARPA initiated the distributed sensor networks program. However, recent advances in microprocessor fabrication have led to a dramatic reduction in both the physical size and power consumption of such devices. Battery and sensing technology, as well as communications hardware, have also followed a similar miniaturization trend. The aggregation of these advances has led to the development of networked, millimeter-scale sensing devices capable of complex processing tasks. Collectively these form a wireless sensor network (WSN), thus heralding a new era of ubiquitous sensing technology and applications. Large-scale deployments of these networks have been used in many diverse fields such as wildlife habitat monitoring (Mainwaring, Polastre, Szewczyk, Culler, & Anderson, 2003), traffic monitoring (Coleri, Cheung, & Varaiya,, 2004), and lighting control (Sandhu, Agogino, & Agogino, 2004). A number of commercial WSN platforms have been launched in recent years. Examples include the Mica family (Hill, 2003), Smart-Mesh (http://www.dust-inc.com), Ember (http://www.ember.com), iBeans (Rhee, Seetharam, Liu, & Wang, 2003), Soapbox from VTT (http://www.vtt.fi/ele/research/tel/projects/soapbox.html), Smart-Its (http://www. smart-its.org), and the Cube sensor platform (O’Flynn et al., 2005). As the miniaturization of the constituent components of a WSN continues unabated, power consumption likewise diminishes, thus the current generation of sensors can function perfectly for years using standard AA batteries (Polastre, Hill, & Culler, 2004). Alternative solutions may not require any batteries; for example iBeans (Rhee et al., 2003) coupled with an energy harvester can operate by scavenging energy from tiny vibrations that occur naturally. Miniaturized solar panels are another possible solution for outdoor operation. Production costs of single nodes are estimated to be less than a dollar, a significant cost reduction
over the price of older sensor models, thus paving the way for large-scale WSN deployments, possibly consisting of a number of nodes several orders of magnitude greater than that in ad-hoc networks (Akyildiz, Su, Sankarasubramaniam, & Cayirci, 2002).
BACKGROUND The main components of a WSN are gateways and sensor nodes. The sensor nodes can relay their sensed data either directly to the gateway or through each other depending on the scale of the network. In turn the gateway can send commands down to the nodes to, for example, increase their sampling frequency. In some networks, when the gateway is tethered to an adequate power supply, a greater transmission range can be achieved. This gives rise to an asymmetry in the data acquisition and control protocols, where control commands are sent directly to the node but the data sent from the node to the gateway is multi-hopped. Of course multi hopping of the control commands from the gateway can be used also. Multi-hopping, while useful in extending the reach or scale of a WSN and reducing the overall transmission cost with respect to direct communication, does have its limitations. The cost of transmitting a packet can be greatly increased depending on the distance a node is from its gateway. Secondly, since nodes nearest the base station, that is, one hop away, will not only have to send their data but also that of all other nodes greater than a single hop, there will be a greater demand placed on the power supply of these nodes. This means that, in general, a node lifespan is inversely proportional to the number of hops it is away from the base station. To alleviate this problem, multiple gateways can be used, with the nodes only transmitting data to their local station. A second solution creates a hierarchy of nodes with varying power and transmission capabilities. Higher power
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Wireless Sensor Networks
nodes can act as gateways to the gateways for the lower powered sensors.
DEPLOYMENT CONSIDERATIONS There are a number of issues to consider when deploying a WSN. The first and perhaps most significant is the number of nodes required in the deployment. The quantity of nodes required will primarily be governed by the size of the area to be monitored and the frequency of the sampling required. In general, the more nodes there are in a given WSN, the better the quality of data; however there is a corresponding increase in the time required for processing. Given that the choice on node density has been made, another factor still remainsthe node sampling frequency. The choice of this value will depend on what aspect of the environment is being monitored and the power resources available on each node. If the sampling frequency is too high, more power will be consumed than is necessary. However, if it is too low, important events could be missed. A useful strategy might be to alter this value opportunistically so as to deliver optimum performance. Getting the sensed data from the node requires the wireless transmission of a data packet, a process that can consume a significant portion of the available power resources. One transmission can occur per sensed value when a real-time picture of the environment is required. Or multiple readings can be bundled into a single message. Alternatively the node may intelligently decide not to transmit a packet if, for example, no change in the sensed value had occurred. This can dramatically increase the longevity of the node, and if this is a universal policy adopted by all nodes in the WSN, the lifespan of the WSN can be extended. Due to the battery operation of the nodes, power management is critically important to the health of a WSN. Another approach to performing power conservation is to enable redundant sensors to hibernate. The rationale behind this is that a sensor consumes little or no power while asleep, and so the more nodes that are hibernating, the less power is being consumed collectively by the network. A hibernating node cannot forward any sensed data, effectively reducing the spatial sampling frequency. Therefore, caution must be exercised when selecting which nodes to hibernate. Two broad approaches exist for selecting nodes for hibernation. The first is based on defining a sensing radius for each individual sensor. An area is covered if all points with the sensed area lie within the sensing range of at least one sensor. When there are points covered by more than one sensor, it may be possible to hibernate redundant sensors without breaking the coverage constraint. An alternative approach uses the data being received by nodes. It is based on interpolation and assumes that the
required node density exists. Given a collection of nodes, it is possible for them to interpolate the sensed medium at a required point, assuming an interpolation function exists. By interpolating at the point of an individual node, an interpolated value can be obtained. By comparing this to the actual sensed value, an interpolation error can be derived. If this error is less than a particular threshold, then this node is deemed redundant and will hibernate for a predefined period of time before rechecking its redundancy. Another fundamental issue for practical WSNs is that of sensor calibration (Whitehouse & Culler, 2003). When two nodes observe different values in their sensed data, is it because they are seeing different events or because one or both of the sensors has malfunctioned? Of course calibration can be done prior to deployment, but if the malfunction causes its accuracy to degrade over time, then a recalibration must occur on the fly after deployment. This is a significant problem, since the environment in which the nodes are sensing usually cannot be controlled for the calibration to occur. When sampling the environment, it may be a requirement for all the sensors to sample at the same point in time. This requires a clock synchronization technique that will work over the entire network, and this is quite a difficult task to perform on such computationally challenged devices. Multi-hop routing can introduce a considerable lag between the time a message is sent from the node and the time it is received at its destination. When the destination is a gateway, it will in turn send control commands to the network based on the data it receives. These control commands may also be multi-hopped to their destination. The aggregate delay can be unacceptable and is usually symptomatic of an overburdened gateway. The introduction of an additional gateway that efficiently partitions the network would alleviate this issue. A final issue with a practical deployment of a WSN is that of programming and debugging the nodes themselves. A node considered to have one of the richest user interfaces consists of three LEDS, allowing eight program execution states to be displayed at any given point in time. To alleviate this, a methodology has been developed to allow the development of applications off the nodes so that accuracy of the approach can be verified (Tynan, Ruzzelli, & O’Hare, 2005).
APPLICATION CONSIDERATIONS The autonomous nature of the wireless sensor networks makes the technology versatile with respect to applications. Sensors can be effectively deployed for monitoring and detecting of malfunctioning industrial machinery during normal production activity. Moreover, as no infrastructure is needed, their deployment is immediate and highly adaptive to the environment in which they operate. By means of
1035
W
Wireless Sensor Networks
distributed sampling, sensor networks are able to provide a more accurate and in-depth evaluation of the state of the environment at any moment in time. Sensors can also be programmed to make decisions at a local stage or through a centralized approach. In the first case, nodes are organized in clusters in which a sensor head is elected. Collected data is sent to the sensor head, where it is evaluated before an appropriate decision is made. For instance it can actuate supplementary machinery in cases where production overloading is occurring. In cases where a sensor is not endowed with a decision-making capability, or in particular circumstances where it does not have enough information to make an immediate decision, collected data is sent to the closest gateway for further processing, and the resulting decision is returned to the sensor. In an energysaving maneuver, data is sent to neighboring nodes that forward it, via a routing procedure, towards the gateway. Such a scenario is an illustration of a centralized approach of decision making. For more effective management of the production activity, sensors can interact, possibly using either fixed networked or wireless technology, with personal digital assistants (PDAs) or mobile phones. In this way, wireless sensor networks can improve the supervision of an activity and communicate with a user in a more effective and efficient manner. For example, in the case of industrial machine monitoring, should the sensor head sense that a particular machine has excessive vibration or that a temperature exceeds a certain threshold, it may decide to raise an alarm and call a technician for assistance. However, the approach it takes to this may vary. Flagging an alarm in a centralized control system can be implemented quite easily. However, technicians and maintenance staff are generally mobile, as their duties call them to various places in the factory floor. Thus, the sensor head must be able to route a message to them. As technicians are likely to be equipped with PDAs or mobile phones, instant messaging and SMS are two obvious methods of contacting them. Applications of m-commerce through the use of collaborative wireless sensor networks and PDAs or mobile phones are numerous. In the case of sensors deployed in a shopping center, a shopper with a PDA or smart phone can request a particular product with certain requirements, for example, model, price, and so on. By means of sensor collaboration, products in the shopping center that match the user requirements can be identified. The shopper can then decide whether to buy remotely or go and view the items in question before buying them. Alternatively, a reverse auction could be initiated with the shops in the mall all vying for the shopper’s business. While such a scenario illustrates the potential synergic interaction of WSNs and standard mobile devices, a number of issues must be resolved before such a vision can become a reality.
1036
FUTURE TRENDS Though significant research in WSNs and mobile computing continues, issues concerning the enablement of seamless and transparent interaction between each domain need to be resolved. A number of issues are now identified.
Communication Protocol Issues In order for a PDA to communicate with a sensor network, it is necessary that both PDAs and WSNs use the same communication protocol. At present, off-the-shelf PDAs have the Bluetooth protocol for short-range communication provided. Unfortunately, studies of the Bluetooth architecture (Leopold, Dydensborg, & Bonnet, 2003) showed the unsuitability of such a protocol for wireless sensor networks. On the other hand, although recent advances propose a vast number of protocols tailored to WSNs, the communication compatibility between the two technologies is still an open issue.
Ontology Issues Such kinds of issues arise after PDAs and sensors agree which communication protocol to use. In the context of knowledge sharing between PDAs and sensors at the application layer, they should agree with the specification of a conceptualization, also known as an ontology. Although some research proposes the study of semantic techniques for wireless sensor networks (Whitehouse, Liu, & Zhao, 2006), a comprehensive methodology of PDA/sensor interaction is still an open issue to be addressed.
Trust Management Issues Requests of m-commerce-related information from sensors to PDAs and vice versa raise issues of trust management. In fact, sensors should trust the quality of service offered by the PDA protocol. On the other side, PDAs should trust sensors when, for example, product availability or machinery conditions are sent to a PDA. While the latter case can be considered as an instance of Internet trust management, the former case needs to consider the issue of memory capacity constraints of sensors. Procedures for realizing trust management on individual sensors, for example, through intelligent agent technologies, need further research. The big “umbrella” of trust management also includes more specific issues of security. In fact, the multi-hop routing of WSNs together with the relatively simple architecture of sensors pose an inherent risk, as an attacker may only need to compromise one device to compromise the security of the entire network. This concern is amplified in applications like m-commerce where private credentials must be fully safely encoded.
Wireless Sensor Networks
CONCLUSION Significant advances have been made in WSNs over the last decade. Nevertheless, power consumption remains a significant barrier to their widespread deployment. However, significant strides are being made in this area. Though a mature research discipline in its own right, the issues of interaction and interoperability with conventional computing systems, and mobile computing devices in particular, is becoming increasingly important. In extending the reach of computing technologies into what are frequently remote and hostile environments, WSNs will enable an array of new and innovative applications and services for mobile users.
REFERENCES Akyildiz, I. F., Su, W., Sankarasubramaniam, Y., & Cayirci, E. (2002). Wireless sensor networks: A survey. Computer Networks Journal, 38(4), 393-422. Bychkovskiy, V., Megerian, S., Estrin, D., & Potkonjak. M. (2003). A collaborative approach to in-place sensor calibration. Proceedings of the 2nd International Workshop on Information Processing in Sensor Networks (IPSN’03), Palo Alto, CA. Coleri, S., Cheung, Y., & Varaiya, P. (2004). Sensor networks for monitoring traffic. Proceedings of the Allerton Conference on Communication, Control and Computing, Illinois. Hill, J. (2003). System architecture for wireless sensor networks. PhD thesis, University of California–Berkeley, USA. Leopold, M., Dydensborg, M. B., & Bonnet, P. (2003). Bluetooth and sensor networks: A reality check. Proceedings of the 1st International Conference on Embedded Networked Sensor Systems, Los Angeles, CA. Mainwaring, A., Polastre, J., Szewczyk, R., Culler, D., & Anderson, J. (2002). Wireless sensor networks for habitat monitoring. Proceedings of the ACM International Workshop on Wireless Sensor Networks and Applications, Atlanta, GA. O’Flynn, B., Barroso, A., Bellis, S., Benson, J., Roedig, U., Delaney, K., Barton, J., Sreenan, C., & O’Mathuna, C. (2005). The development of a novel miniaturized modular platform for wireless sensor networks. Proceedings of the IPSN Track on Sensor Platform, Tools and Design Methods for Networked Embedded Systems (IPSN2005/SPOTS2005), Los Angeles, CA.
Polastre, J., Hill, J., & Culler, D. (2004). Versatile low power media access for wireless sensor networks. Proceedings of the 2nd International Conference on Embedded Networked Sensor Systems (SenSys ’04) (pp. 95-107). New York. Rhee, S., Seetharam, D., Liu, S., & Wang, N. (2003). iBean network: An ultra low power wireless sensor network. Proceedings of UbiComp, the 5th International Conference on Ubiquitous Computing, Seattle, WA. Sandhu, J., Agogino, A., & Agogino, A. (2004). Wireless sensor networks for commercial lighting control: Decision making with multi-agent systems. Proceedings of the AAAI04 Workshop on Sensor Networks, San Jose, CA. Tynan, R., Ruzzelli, A. G., & O’Hare, G. M. P. (2005). A methodology for the development of multi-agent systems on wireless sensor networks. Proceedings of the 17th International Conference on Software Engineering and Knowledge Engineering (SEKE’05), Taiwan. Whitehouse, K., & and Culler, D. (2003). Macro-calibration in sensor/actuator networks. Mobile Networks and Applications Journal (MONET), (Special Issue on Wireless Sensor Networks). Whitehouse, K., Liu, J., & Zhao, F. (2006). Semantic streams: A framework for composable inference over sensor data. Proceedings of the European Workshop on Wireless Sensor Networks (EWSN), Zurich, Switzerland.
KEY TERMS Gateway: In general, considered a more powerful node that is used to collect information from the networks and for some architecture to synchronize them. A WSN can have few gateways deployed in the network. Sometimes, they are assumed to be interconnected through alternative wireless technology like WLAN and WiMAX. Gateways are often referred to as sinks or inappropriately called base stations. MAS: Multi-agent system. Node Lifespan: Also known as node lifetime, it represents the operational life expectancy of a sensor. Usually, it is calculated for certain network parameters (e.g., network topology, network density) and certain node parameters (e.g., node data rate, duty cycle of a sensor equipped with two AA standard batteries). Personal Digital Assistant (PDA): Also known as a palmtop. WSN: Wireless sensor network.
1037
W
1038 Category: M-Business and M-Commerce
Wireless Technologies for Mobile Computing and Commerce David Wright University of Ottawa, Canada
INTRODUCTION At the time of writing (1Q06) most countries have a small number (2-6) of major cellular operators offering competing 2.5G and 3G cellular services. In addition, there is a much larger number of operators of WiFi networks. In some cases, a major cellular operator, for example, Deutsche Telekomm and British Telecom, also offers a WiFi service. In other cases, WiFi services are provided by a proliferation of smaller network operators, such as restaurants, laundromats, airports, railways, community associations and municipal governments. Many organizations offer WiFi free of charge as a hospitality service, for example, restaurants. Cellular services offer ubiquitous, low data rate communications for mobile computing and commerce, whereas WiFi offers higher data rates, but less ubiquitous coverage, with limitations on mobility due to business as opposed to technology reasons. Emerging networks for mobile computing and commerce include WiMAX and WiMobile (Wright, 2006), which offer higher data rates, lower costs and city-wide coverage with handoff of calls among multiple base stations. These new technologies may be deployed by the organizations that currently deploy cellular and WiFi networks, and also may give rise to a new group of competitive wireless network operators. This article identifies the capabilities needed for mobile computing and commerce and assesses their technology and business implications. It identifies developments in the wireless networks that can be used for mobile computing and commerce, together with the services that can be provided over such networks. It provides a business analysis indicating which network operators can profitably deploy new networks, and which network operators need to establish business and technology links with each other so as to better serve their customers. The resulting range of next generation service, technologies and network operators available for mobile computing and commerce is identified.
The cellular architecture is the most sophisticated in that the core network includes a circuit network (for legacy circuit switched voice calls), a packet network (for data calls) and an IP Multimedia Subsystem, IMS (for migration of all traffic onto the Internet). These three networks essentially allow the cellular operator to maintain control over all calls to and from the mobile device, and hence derive revenue from them. In particular the IMS network contains servers for establishing voice and video calls over IP, authenticating users, maintaining records of the current location of a mobile user, accounting, and security. Cellular operators are migrating traffic from their circuit and packet networks onto the IMS. By contrast, WiFi (IEEE, 1999a, 1999b, 1999c, 2003), WiMAX (IEEE, 2006; Ghosh et al., 2005), and WiMobile (IEEE, 2006; Lawton, 2005) are simply radio access technologies and do not specify a core network. They therefore allow more direct access from a mobile device to the Internet. In particular, the WiMobile specification, which is under development at the time of writing, emphasizes that its design is being optimized for operation with IP. This more open access to the Internet allows a mobile user to set up, for instance, a VoIP call using a third party service without the involvement of the wireless network operator. As the user moves from one access point to another, the call can be maintained using Mobile IP, involving servers maintained by the user’s ISP, not by the wireless network operator. Mobile IP can operate over diverse wireless access technologies as described by Benzaid et al. (2004). If the operator of a WiFi, WiMAX or WiMobile network wishes to maintain more control over the traffic passing through their network and hence participate more in the revenue generated by that traffic, they can build an IMS network. Alternatively if they already operate a cellular network, they can provide access to their existing IMS network, as shown by the dashed lines in Figure 1.
WIRELESS NETWORK ARCHITECTURES
REQUIREMENTS FOR MOBILE COMPUTING AND COMMERCE
Figure 1 illustrates the network architectures for WiFi, Cellular, WiMAX and WiMobile, including the radio access network on the left and the wired core network on the right.
Any wireless transmitter/receiver has a limited range in order to comply with government regulations regarding maximum power output. A mobile user therefore may move out of the
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Wireless Technologies for Mobile Computing and Commerce
Figure 1. Wireless network architectures
W
WiFi Distribution System
Cellular
IMS Pkt. Ntwk
WiMAX
Radio Network Controller
PSTN
Internet
Circuit Ntwk
Internet PSTN
Internet
WiMobile
PSTN
Backhaul Switch
Internet Router
Radio Access Network
range of its current wireless access point, and it is necessary to handoff the communication to another access point using either the same or a different wireless technology. Handing off the communication means that the current IP session is maintained, for example, the user continues to browse a Web site as a registered user, a VoIP call is not interrupted, and an enterprise user with a laptop-based secure VPN to an enterprise network continues to use the same VPN. There are four requirements in order to achieve handoff suited to mobile computing and commerce: 1. 2.
3.
4.
It must be possible to switch the call from one access point to another If the user is receiving quality of service, QoS, for example, a guaranteed low latency, that QoS is maintained after the handoff, and an acceptable number of packets are lost during handoff. If the access points are operated by different network operators, there must be a business arrangement between them regarding mediation of the billing for the call. The organization deploying the wireless access network must be able to make a profit or to have a business model that focuses on hospitality service.
Requirements 1 and 2 are technology related and are discussed next, followed by the business requirements 3 and 4.
Core Network
TECHNOLOGY ISSUES A mobile device that is capable of using multiple wireless access technologies, such as those described above, can continuously scan its radio environment to search for access points that it could potentially use. Some of them may not be available, if, for instance, they are operated by companies with which the user does not have a subscription. In order to choose among the available access points within range the mobile device can apply criteria including: data rate, cost, ability to handoff seamlessly, and QoS; delay (important for voice) and packet loss rate (important for data). For instance, a mobile device with an interactive voice/video call in progress could choose the lowest cost network that provides acceptable delay. A device downloading a large data file could choose the network with the highest data rate given limitations on cost and packet loss rate. Once the network is selected, handoff is initiated. Handoff among WiFi, WiMAX and WiMobile is handled by IEEE (2006). Handoff between cellular and one of these three technologies is complicated by the need to interwork with the cellular circuit, packet and IMS networks. •
In the case of WiFi, this interworking is provided by a specification from the industry consortium UMA, Unlicensed Mobile Access (2006), which is incorporated as part of the GSM cellular network specifications, release 6. 1039
Wireless Technologies for Mobile Computing and Commerce
Table 1. Wireless access network operators revenue sources
• •
Revenue Source
Service provided by: (N) Network Operator (C) Content Provider
Revenue accrues to: (W) Wireless Network Operator (3) Third Party Service Provider (S) Shared with Content Provider
Voice/video calls
N
W3
Audio content
C
S
Video content
C
S
Gateway to PSTN
N
W3
Geographic info (e.g., travel directions, highway safety)
NC
WS
Location enabled advertising
NC
WS
Location enabled buddy lists
N
W
Multimedia Messaging Service
C
W3S
Gaming
C
S
QoS
N
W
VPN
N
W3
In the case of WiMAX, similar issues are involved and are being resolved by the WiMAX Forum (2006). WiMobile is at an early stage of development and interworking with cellular is not a priority at this stage. A specification may be developed later, or alternatively, WiMobile may differentiate itself from the other technologies by becoming a “native-IP” access mechanism, similar to DSL and cable modem in which customers have direct access to the Internet.
This discussion addresses requirements 1, 2 above. We now move on to requirements 3, 4.
BUSINESS ISSUES This section presents business strategies for wireless access network operators that take into account sources of revenue related to mobile computing and commerce, plus the need to compete with other technologies and network operators. Earlier work in this area (du Preez & Pistorius, 2003) dates from a time when 3G and wireless data services were emerging technologies. The present section incorporates developments in technology and services to date. The sources of revenue are given in Table 1 and are classified in two ways: 1.
1040
Whether the service is provided by a content provider or a network operator, which may be the wireless access network operator or another network operator. For instance, a VoIP service could be provided by the
2.
wireless operator or by a third party such as Vonage. Either way it is provided by a network operator. Who receives the revenue for the service: the wireless access network operator, a third party or a sharing arrangement with a content provider.
It can be seen from Table 1 that there is a large number of mobile computing and commerce services that can be provided by a mix of wireless network operators, content providers and third parties. In addition there are non-revenue generating services such as e-mail and Web browsing. A clear business strategy is needed to operate successfully in competition with the other players. Strategies suited to the different types of wireless network operators are given in Table 2. Table 2 divides wireless access network operators into three groups: incumbent cellular operators, hospitality providers such as restaurants and municipalities, and new competitors, who are starting operations based on the availability of new technology. The incumbent cellular operators have complex core networks as shown in Figure 1 and incur costs of operating legacy technologies. They seek to deploy all possible wireless technologies in order to accommodate the needs of all customers. By contrast the new competitors seek to reduce their costs by only operating the most recent technologies. Both these groups are operating commercial services and therefore use licensed spectrum so that their customers do not experience interference from other users. The hospitality providers, however, are providing a free service. Their customers accept that the performance may vary according to the demands of other users and therefore the operators reduce their costs by using unlicensed spectrum.
Wireless Technologies for Mobile Computing and Commerce
Table 2. Strategy for wireless access network operators Cellular Operators
Hospitality providers
Competitive Wireless Network Operators
Technologies
2.5G, 3G, WiFi, WiMax, WiMobile
WiFi, unlicensed WiMAX
WiMAX, WiMobile
Revenue sources
Generate revenue from the full range of services
Provide Internet access for the full range of services.
Generate revenue from the full range of services
IMS strategy
Lock customers into IMS-based services.
Establish partnerships and interfaces to the IMS of other operators
Build IMS. Establish partnerships and interfaces to the IMS of other operators
Buy up competitors.
Avoid competing with other operators by a competitive bid process.
Differentiate from incumbents by offering low cost services, focusing on IP, developing next generation services, for example, presence, location, QoS.
Competitive strategy
Both the incumbents and the new competitors aim to deliver the full range of services listed in Table 1 to their customers, typically from the IMS, so as to maintain control over the revenue. The hospitality providers, however, are typically providing access only, allowing their customers to get services from any third party they wish, since they do not seek to generate revenue from their networks. For location-based services, the hospitality provider can provide the third party with information about the customer’s current location. The cellular incumbents typically already have an IMS in place and aim to lock customers into service provided by that IMS. The new competitors need to build an IMS and then establish partnerships with other wireless operators so that calls originating on one IMS can be handed off to another operator. These partnerships are also important to the hospitality providers since they typically have no interest in developing their own IMS. The competitive strategy of incumbent cellular operators towards WiFi operators historically has been to buy them up, and this strategy is also appropriate for WiMAX and WiMobile operators. The strategy of hospitality operators is to avoid competition, and this is particularly important for municipalities, who should not be seen to use tax dollars to compete against private industry. In order to avoid this perception, they can use a competitive bid process allowing any operator the opportunity to bid on the contract to build and operate their network. The strategy of the new competitors is to compete on three fronts. First, they can offer low cost services, since they do not have the cost of operating legacy networks. Second, they can offer a full range of next generation services, such as presence and location-based services, thus positioning themselves as state-of-the-art suppliers. Third, they can sell QoS guarantees to their customers, since new technologies such as WiMAX and WiMobile are particularly suited to providing such guarantees.
W
CONCLUSION The enabling technologies for mobile computing and commerce are developing rapidly. New wireless technologies such as WiMAX and WiMobile offer extended coverage and improved QoS compared to WiFi; and higher data rates and lower costs compared to 2.5G and 3G cellular. A wide range of services is available over these technologies including services that generate revenue (a) for the wireless operator, such as location-based services, (b) for a third party, such as VoIP and (c) for a content provider, such as entertainment. Wireless network operators, including incumbent cellular operators, hospitality providers and new competitive wireless network operators, need to develop strategies that allow handoff of calls among the different technologies and operators. Strategies include locking customers into an IMS, interworking with other operators’ IMSs, buying out competitors and developing a broad range of state-of-the-art services such as location and presence services. The mobile computing and commerce user can therefore expect a proliferation of services (Table 1), a number of different network operators (Table 2), an array of different wireless technologies, WiFi, 3G, WiMAX and WiMobile, and a mobile device that can make the best choice among these alternatives at any point in time and space.
REFERENCES Benzaid, M., Minet, P., Al Agha, Kh., Adjih, C., & Allard. G. (2004). Integration of mobile-IP for universal mobility. Wireless Networks, 10(4), 377-388. du Preez, G. T., & Pistorius, C. W. I. (2003). Analyzing technological threats and opportunities in wireless data 1041
Wireless Technologies for Mobile Computing and Commerce
services. Technological Forecasting and Social Change, 70(1), 1-20. Ghosh, A., Wolter, D. R., Andrews, J. G., & Chen, R. (2005, February). Broadband wireless access with WiMax/802.16: Current performance benchmarks and future potential. IEEE Communications, 43(2), 129-136. IEEE. (1999a). 802.11 Wireless LAN: Medium access control (MAC) and physical layer (PHY) specifications. New York: IEEE Publications. IEEE. (1999b). 802.11a high-speed physical layer in the 5 GHz band. New York: IEEE Publications. IEEE. (1999c). 802.11b higher-speed physical layer (PHY) extension in the 2.4 GHz band. New York: IEEE Publications. IEEE. (2003). 802.11g further higher-speed physical layer extension in the 2.4 GHz band. New York: IEEE Publications. IEEE. (2006a). 802.16e air interface for fixed and mobile broadband wireless access systems: Amendment for physical and medium access control layers for combined fixed and mobile operation in licensed bands. New York: IEEE Publications. IEEE. (2006b). 802.20 mobile broadband wireless access (In Progress). Retrieved March 2006, from http://grouper. ieee.org/groups/802/20 IEEE. (2006c). 802.21 media independent handover services (In Progress). Retrieved March 2006, from http://grouper. ieee.org/groups/802/21/ Lawton, G. (2005). What lies ahead for cellular technology? IEEE Computer, 38(6), 14-17. UMA. (2006). Unlicensed mobile access. Retrieved from http://www.umatechnology.org/specifications/index.htm WiMAX Forum. (2006). Retrieved March 2006, from www. wimaxforum.org Wright, D. (2006). Wireless technologies for mobile computing and commerce. In D. Taniar (Ed.), Encyclopedia of mobile computing and commerce. Hershey, PA: Idea Group Reference.
1042
KEY TERMS IMS, IP Multimedia Subsystem: Part of the wired core network containing servers for establishing voice and video calls over IP, authenticating users, maintaining records of the current location of a mobile user, accounting, and security. Location-Based Services: Services that take into account the users current geographical location, for example, advertising locally available products and services, providing directions and alerting drivers to traffic congestion and road accidents. Mobile IP: An Internet standard that allows a mobile user to move from one point of attachment of the network to another while maintaining an existing TCP/IP session. Incoming packets to the user are forwarded to a server in the user’s new access IP subnetwork. Presence: The ability of a user device to specify characteristics, such as whether the user is online, whether the user is willing to receive calls, whether the user is willing to receive calls of a given type (e.g., voice, video, data, MMS) from specified other users and what is the user’s current location to a specified degree of accuracy. Quality of Service (QoS): Features related to a communication, such as delay, variability of delay, bit error rate and packet loss rate. Additional parameters may also be included, for example, peak data rate, average data rate, percentage of time that the service is available, mean time to repair faults and how the customer is compensated if QoS guarantees are not met by a service provider. WiFi: A commercial implementation of the IEEE 802.11 standard in which the equipment has been certified by the WiFi Alliance, an industry consortium. WiMAX: A commercial implementation of the IEEE 802.16 standard in which the equipment has been certified by the WiMAX Forum, an industry consortium. WiMobile: Another name for the IEEE 802.20 standard, which is in course of development at the time of writing (1Q06).
Category: Ad-Hoc Network 1043
Workflow Management Systems in MANETs Fabio De Rosa University of Rome “La Sapienza”, Italy Massimiliano de Leoni University of Rome “La Sapienza”, Italy Massimo Mecella University of Rome “La Sapienza”, Italy
INTRODUCTION The widespread availability of network-enabled handheld devices (e.g., PDAs with WiFi) has made pervasive computing environment development an emerging reality. Mobile (or multi-hop) Ad-hoc NETworks (MANETsAgrawal & Zeng, 2003) are mobile device networks communicating via wireless links without relying on an underlying infrastructure. Each device in a MANET acts as an endpoint and as a router forwarding messages to devices within radio range. MANETs are a sound alternative to infrastructure-based networks whenever the infrastructure is lacking or unusable, for example, military applications, disaster/relief, emergency situations, and communication between vehicles. Generally, the use of a MANET requires a strong collaboration among users/devices constituting the network; more often, that collaboration is translated into a list of activities executed in sequence or concurrently by applications running on mobile devices, thus resulting in cooperative works. Therefore, in such situations, MANET users would benefit from software supporting their collaboration. Such a coordination software would enable them to execute their activities through specific applicationsfor example, computer-supported cooperative work tools (Grudin, 1994), workflow management applications (Leymann & Roller, 2000), and so forthrunning on handheld devices, thus enabling cooperative processes to be run. All such applications typically require continuous inter-device connections (e.g., for data/information sharing, activity scheduling and coordination, etc.), but these are not generally guaranteed in MANETs, due to the high mobility of the nodes in the network.
Collaborative Scenarios Consider a scenario of emergency situations, for example, the case of an aftermath of an archeological disaster. A team is equipped with mobile devices (laptops and PDAs) and sent to the affected area to evaluate the condition of archeological sites and buildings, with the goal of drawing a situation
map to schedule rebuilding activities. A typical cooperative process to be enacted by the team would be as shown in Figure 1 (depicted as a UML Activity DiagramDe Rosa, Malizia, & Mecella, 2005): 1.
2.
3.
The team leader has previously stored all area details (not included in the process), including a site map, a list of the most important objects located in the site, and previous reports/materials. The team is considered as an overall MANET, in which the team leader’s device (requiring the most computing power, therefore usually a laptop) coordinates the other team members’ devices, by providing suitable information (e.g., maps, sensitive objects, etc.) and assigning activities/tasks. Team members are equipped with handheld devices (PDAs), which allow them to run some operations but do not have much computing power. Such operations, possibly involving various hardware items (e.g., digital cameras, GPRS connections, computing power for image processing, main storage, etc.) are provided as software services to be coordinated. Team member 1 might fill in some specific questionnaires (after a visual analysis of a building), to be analyzed by the team leader using specific software in order to schedule subsequent activities; team member 3 might take pictures of the damaged buildings, while team member 2 may be responsible for specific processing of previous and recent pictures (e.g., for initial identification of architectural anomalies).
In this case, it might be useful to match new pictures with previously stored images. The device holding the highresolution camera must therefore be connected to the one containing the stored pictures. But in a situation such as the one shown in Figure 2, the movement of the operator/device equipped with the camera may result in its disconnection from the others. A pervasive architecture should be able to predict such a situation, alert the coordination layer, and possibly have a
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
W
Workflow Management Systems in MANETs
Figure 1. Cooperative process
Compile Questionnaire
Data
Select Building
Selected Building
Select Building
Go to Destination
Zoom on Damaged Part Capture Scene Photos
Result
Send Photos
Matching
Compile Report
Team Member 1
Team Leader
Team Member 2
Team Member 3
Figure 2. Critical situation and adaptive management
Movement needed to accomplish the task
Movement needed to maintain the network connectivity; should be adaptively driven by the cooperative application
“bridging” device (team member 4’s device) to follow the operator/device moving out of range, maintaining the connection, and ensuring a path between devices. In this way the coordination layer schedules the execution of new activities based on the prediction of a disconnection, as shown in Figure 3 (note the new activity for team member 4). 1044
The process’s adaptive change is centrally managed by the coordination layer, which has “global” knowledge of the status of all operators/devices and takes into account idle devices, operations that can be safely delayed, and so forth.
Workflow Management Systems in MANETs
Figure 3. Modified process (details)
W
Select Building
Selected Building Go to Destination
Zoom on Damaged Part
Follow Team Member_3
Capture Scene
Photos
Send Photos
Matching
Team Member 2
Team Member 3
Some Considerations Cooperative work may be performed by Workflow Management Systems (WfMSs) (Hollingsworth, 1995; van der Aalst & Van Hee, 2002), a software layer for defining, handling, and carrying out computerized processes (referred to as workflows), through programs whose performing order is leaded by a process logic representation. A workflow is made up of a set of tasks to be performed, a set of state variables handled by tasks, and a set of conditions on such variables which can cause a selection in performing between two or more tasks. Occasionally, some tasks have to be carried out strictly in sequence; other times in a parallel or arbitrary order because there is no causal relationship between them. Actors performing workflow tasks exchange much information needed for their collaboration and coordination. Such an exchange can occur, of course, only if devices/actors are continuously linked to each other. In a MANET scenario, the “bridging” actions can be seen as supporting activities needed for primary ones which, otherwise, could not be carried out, prejudicing the whole process execution. These support activities are added, when needed, concurrently with other ones. Therefore, a WfMS for MANET has to support dynamic adaptation of processes where the definition of process instances changes on altering of execution context. In MANETs, the altering is caused by disconnection of devices.
Team Member 4
The critical issue of current WfMSs is the lack of support of adaptiveness; most WFMSs (both commercial products and academic prototypes) handling adaptiveness assume the presence of an expert who decides which changes have to be applied. In MANETs, the assumption of an expert whose only purpose is disconnection handling is not acceptable, therefore we must investigate how automatically to decide, first, the inadequacy of process instance, and, then which changes on workflow definition must be applied to handle disconnections.
Related and Background Works In recent years, research in the MANET area has mainly focused on the development of appropriate routing protocols, security and reliability of the communications, methods for energy preservation, and other issues on the lower four ISO/OSI layers (Arbaugh, 2004; Vaidya, 2004). Effective routing in ad-hoc networks is still an actively addressed open problem (Beraldi & Baldoni, 2002; Vaidya, 2004), with some interesting proposals presented in the literature (e.g., dynamic source routingDSR, ad-hoc on-demand distance vectorAODV routing, zone routing protocolZ-RP, etc.). Researchers in this area assert that a sound technical basis for MANETs exists and it is thus time to start thinking about how to support applications based on MANETs. In order to enable the development of application layer software (and 1045
Workflow Management Systems in MANETs
thus of any information system for MANETs), abstractions on the specific characteristics of the routing algorithms, and more generally, on the services and data provided by the lower network layers, are required. (De Rosa, Di Martino, Paglione, & Mecella, 2003) proposes a network service interface to be used as the basic layer on which to build application software, starting from the analysis and abstraction of current routing protocols. As far as the issue of adaptive workflow management (Baresi, Casati, Castano, Mirbel, & Pernici, 1999; Voorhoeve & van der Aalst, 1997), this is still an open issue. Relevant work includes the e-Flow system (Casati & Shan, 2001), in which the issue of manually modifying workflow schemas and then automatically migrating active process instances to the new schema is addressed, and AGENTWORK (Müller, Greiner, & Rahm, 2004), which is one of the few examples of a workflow system in which adaption is not manual, but automatic, on the basis of a rule-base approach. But in MANETs the adaption should be carried out in a very frequently changing environment, whereas the previous approaches are targeted to Web-based workflows (indeed workflows composed of different Web services), in which modifications of the schemas are less frequent, but the number of running instances is very high.
THE MOBIDIS PERVASIVE WfMS Our approach, named MOBIDIS1, is based on the following specific assumptions: •
• •
•
Each device includes hardware that lets it know its communication distance from the surrounding devices that are within radio range. This is not a very strong assumption, because specific techniques and methods are easily available—for example, TDOA (time difference of arrival), SNR (signal-to-noise ratio), and the Cricket compass (Priyantha et al., 2001). No device in the MANET has GPS hardware. At start-up, all devices are connected (that is, each device has a path—possibly multi-hop—to any other device). Each device does not have to be within range of any other device; that is, it does not require a tight (one-hop) connection. It requires only a loose connection, guaranteed by appropriate routing protocols. A specific device in the MANET, called the coordinator, centrally predicts disconnections and manages them by rewriting the workflow schema and (if it is required) by reassigning the process tasks among the participants.
The proposed approach combines local connection management among devices with global management of both network topology and task assignment. Local connection 1046
management consists of monitoring and checking one-hop communications between a device and its neighbors. It is realized as special services running on handheld devices that implement techniques for estimating and calculating distances and relative positions (angle and direction of arrival) between a specific device and its direct neighbors. Global management maintains a consistent state of the network and of each peer in the network. It manages the network topology (and its predicted next states) and the tasks each peer is in charge of, as well as services that peers offer (that is, it provides a service registry). On the basis of that information, the coordinator applies algorithms for choosing a bridge and/or executes workflow task reassignment when needed. Figure 4 shows the proposed architecture to support disconnections and to enact cooperative work in MANETs: each device has a wireless stack consisting of a wireless network interface (the wireless channel and LPC MAC modules) and the hardware for calculating distances from neighbors (angle of arrive and ranging modules). On top, a network service interface offers to upper layers the basic services for sending and receiving messages (through multi-hop paths) to and from other devices, by abstracting over the specific routing protocols. Offered services (i.e., specific applications supporting tasks of the devices’ human users) are accessible to other devices and can be coordinated and composed cooperatively. Some of these services are applications that do not require human intervention (for example, an image-processing utility). Others act as proxies for humans (for example, the service for instructing human users to follow a peer is a simple GUI that alerts the user by displaying a pop-up window on his or her device and emitting a signal). In contrast, the coordinator device presents the predictive layer on top of the network service interface, signaling any probable disconnection to the upper coordination layer. The predictive layer implements a probabilistic technique (De Rosa et al., 2005) which can predict if all devices will still be connected in the next instant. At a given time instant ti in which all devices are connected, the coordinator device collects all device distance information and builds a next-connection-graphthat is, the most likely graph at the next time instant ti+1, in which the predicted connected and disconnected devices are highlighted. In the interval [ti, ti+1], the coordinator layer enacts the appropriate actions to enable all devices to be still connected at ti+1. In predicting at ti the next-connection-graph, the technique considers not only the current situation, but also recent situations and predictions (i.e., at ti-1, ti-2, etc.), specifically considering distances calculated in the recent past. Thus, although the pervasive architecture guarantees the connection of all devices, MANET’s evolution is considered as it would be in a “free” scenario (i.e., without remedial actions by the coordination layer) when predicting the future situation. The reasonable assumption is that if two devices have the
Workflow Management Systems in MANETs
Figure 4. MOBIDIS architecture
LPC MAC
W
Angle of Arrive (AOA)
Ranging (distance)
tendency to go out of radio range if left “free,” and are thus connected through the coordinator’s remedial actions, then this influences the subsequent connection probability. The coordination layer manages situations when a peer is going to disconnect, by applying algorithms for choosing a bridge, and by executing workflow schema restructuring and workflow task reassignment when needed (e.g., it assigns the activity “follow peer x” to the selected bridge). The algorithm for choosing a bridge selects the best one by using the following criteria: 1.
2. 3.
The algorithm chooses as bridge the neighbors without performing tasks. Indeed, if the selected one was a neighbor carrying out a task, its performing activity should be rolled-back with a decrease of productivity. If each neighbor is carrying out a task, then the lowest priority task is chosen. By rolling-back task with the lowest priority, inefficiency should be bounded. If two or more actors perform a task with the same lowest priority, the one with the smallest number of neighbors is preferred. The bridge role likely leads to movement of the node and this might cause new disconnections. By selecting a node with the lowest
4.
number of neighbors, the probability of new disconnection is minimized. In the end, the nearest neighbor is preferred.
When the bridge node is selected, the process instance is modified by adding the supporting activity “follow x” concurrently with the activity of node x going to disconnect. Thus the supporting activities have the same priority of supported ones. This is the situation depicted in Figure 5a in which the bridge node is blue colored. In some cases, the node going to disconnect can produce more than two components (Figure 5b) in the network. In this situation, a bridge is needed for each partition the disconnection is going to create. For each partition, the bridge is selected by applying an algorithm instance over the subset of node belonging to such a partition. An important issue of the bridge algorithm is the priority among activities. The priority must reflect the process structure (causal dependency): the purpose is to assign higher priority to those tasks whose executions generate the achievement of enabled state for a greater number of tasks. An enabled task is a not yet running task that is ready to be assigned to an actor, as all the preceding tasks have been terminated. 1047
Workflow Management Systems in MANETs
Figure 5. Creation of two (a) or more (b) partitions
(a) (b)
The algorithm for computing priorities is based on an n-ary tree that can be built iteratively. A well-structured process can be shown as a composition of many sub-processes decomposable in other smaller sub-processes and so on up to elementary processesthat is, activities. Each node of the tree is a process (elementary or not) whose children are nodes representing the sub-processes it can be decomposed in. The weight of leaves is initially posed to 1; the weight of internal nodes is obtained by combination of weights of children nodes, according to the way the process is decomposed. Therefore the restructuring of the process is reduced to a transformation of such an n-ary tree.
FUTURE TRENDS In the context of some Italian and international research projects, we are going to validate our approach in real scenarios. Preliminary experiments on synthetic data shows the feasibility of the approach (De Rosa et al., 2005), but an extensive real validation is needed. Moreover, we will also address the issue of the approach’s fault tolerance. Our approach currently does not cope with sudden downs of devices, which might be frequent in emergency scenarios and are critical if they affect the coordinator node. We also plan to evolve the coordination layer from a centralized to a distributed one (i.e., having a subset of devices act as coordinators). At the moment, the centralized architecture might be a bottleneck, but the current dimensions of a typical MANET for the considered scenarios (tens of devices) do not pose critical scalability issues. In the future, due to the wide diffusion of mobile devices, applications on MANETs will become more and more interesting. As we have shown, such applications should take into account disconnection anomalies. To date, we have investi1048
gated how classical WfMS concepts should be evolved and adapted in order to cope with MANET scenarios, but surely similar issues will also be raised for many other classical technologies, when applied to MANETs.
REFERENCES Agrawal, D. P., & Zeng, Q. A. (2003). Introduction to wireless and mobile systems. Thomson Brooks/Cole. Arbaugh, W. A. (Ed.). (2004). Making wireless work. IEEE Security & Privacy, 2(3), 7-96. Baresi, L., Casati, F., Castano, S., Mirbel, I., & Pernici, B. (1999). WIDE workflow development methodology. Proceedings of the International Joint Conference on Work Activities Coordination and Collaboration (pp. 19-28). Beraldi, R., & Baldoni, R. (2002). Unicast routing techniques for mobile ad hoc networks. The handbook of mobile ad hoc networks. CRC Press. Casati, F., & Shan, M. C. (2001). Dynamic and adaptive composition of e-services. Information Systems, 6(3), 143163. De Rosa, F., Di Martino, V., Paglione, L., & Mecella, M. (2003). Mobile adaptive information systems on MANET: What we need as basic layer? Proceedings of the 1st IEEE Workshop on Multichannel and Mobile Information Systems (MMIS’03), Rome (pp. 260-267). IEEE CS Press. De Rosa, F., Malizia, A., & Mecella, M. (2005). Disconnection prediction in mobile ad hoc networks for supporting cooperative work. IEEE Pervasive Computing, 4(3), 62-70. Grudin, J. (1994). Computer-supported cooperative work: History and focus. IEEE Computer, 27(5), 19-26.
Workflow Management Systems in MANETs
Hollingsworth, D. (1995, January). The workflow reference model. Workflow Management Coalition. Leymann, F., & Roller, D. (2000). Production workflow: Concepts and techniques. Englewood Cliffs, NJ: Prentice Hall. Müller, R., Greiner, U., & Rahm, E. (2004). AgentWork: A workflow-system supporting rule-based workflow adaptation. Data and Knowledge Engineering, 51(2), 223-256. Priyantha, N. B., Miu, A., Balakrishnan, H., & Teller, S. (2001). The Cricket compass for context-aware mobile applications. Proceedings of the 7th Annual International Conference Mobile Computing and Networking (MobiCom01) (pp. 1-14). Boston: ACM Press. Vaidya, N. H. (2004). Mobile ad hoc networks: Routing, MAC and transport issues. Tutorial on Mobile Ad Hoc Networks, University of Illinois at Urbana-Champaign, USA. Retrieved from http://www.crhc.uiuc.edu/nhv van der Aalst, W., & Van Hee, K. (2002). Workflow management: Models, methods and systems. Cambridge, MA: The MIT Press.
Bridge: A MANET device in a cooperative group following another one d in order to avoid d going out from the network, by preserving a multi-hop path from every other node to d. Computer-Supported Cooperative Work (CSCW): A collective name for the methods and techniques of a system that support the cooperative performance of work. MANET (Mobile (or Multi-hop) Ad-hoc NETwork): Mobile network where devices communicate one with another via wireless links without relying on an underlying infrastructure, like an access point. Signal-to-Noise Ratio (SNR): Difference in decibels between the strength of a signal emitted/received from/to a device and the noise emitted/received from/to the same device. In a radio link, the strength of a signal received decays and the noise grows exponentially with the distance from transmitter (so SNR decays). It is possible to deduce distance between a pair of devices by analyzing SNR of receiving signals.
Voorhoeve, M., & van der Aalst, W. (1997). Ad-hoc workflow: Problems and solutions. Proceedings of the 8th DEXA Conference.
Workflow Management System (WfMS): A system that completely defines, manages, and executes processes through the execution of software whose order of execution is driven by a computer representation of the business process logic.
KEY TERMS
ENDNOTE
Adaptive WfMS: A system that provides process support like normal WfMS, but in such a way that it is able to deal with certain changes in execution context (like node disconnections).
1
MOBIDIS: MOBile at Dipartimento di Informatica e Sistemistica (DIS), of the University of Rome “La Sapienza”, Italy.
1049
W
1050 Category: Mobile Software Engineering
XML-Based Languages for Multimodality in Mobile Environments Danilo Avola Istituto di Ricerche sulla Popolazione e le Politiche Sociali, Italy Maria Chiara Caschera Istituto di Ricerche sulla Popolazione e le Politiche Sociali, Italy Fernando Ferri Istituto di Ricerche sulla Popolazione e le Politiche Sociali, Italy Patrizia Grifoni Istituto di Ricerche sulla Popolazione e le Politiche Sociali, Italy
INTRODUCTION The development of multimodal tools and mobile devices in particular is producing great interest, especially for accessing Web information, performing transactions, and use of services in general. This article considers the different markup languages proposed by the working groups of W3C (World Wide Web Consortium) to manage multimodal interaction and perspectives of multimodal applications and services. The trend toward the convergence of various methodologies and technologies has developed new devices providing complex services, contributing to the sharing of experiences, and promoting the inclusion of people as community members (Paternò, 2004). This trend is based on the development of mobile devices and their usability, accessibility, portability, and versatility (Kvale, Warakagoda, & Knudsen, 2003). The usefulness and usability of services, and the ability to access them and information, are the basic elements in the diffusion of Web systems and development of Web multimodal languages. The diffusion and implementation of multimodal services is supported by the activities of the World Wide Web Consortium, aimed at extending interaction modes for different devices and particularly devoted to solving various problems connected with: (1) multimodal Web interaction through the different devices, and (2) practice Web navigation from different devices. Some W3C working groups focus their activities on issues such as independence from devices, multimodal Web access, and types of contents for multimodal messaging. These specifications allow rich multimodal contents to be transmitted, and are based on the power and extensibility of XML (eXtensible Markup Language) (Bray, Paoli, SperbergMcQueen, Maler, & Yergeau, 2004). XML is highly important in a mobile application environment, as many applications have to manage multimedia
contents and need dedicated tools for this. SMIL (Synchronized Multimedia Integration Language) (Solon, McKevitt, & Curran, 2004) was proposed to achieve this goal. In the early years W3C-MMI (W3CMultiModal Interaction) focused on multimodal interaction modes such as speech and pen interaction, and providing users with W3C technologies. W3C develops these technologies by orienting individual interaction modes in order to create mixed-namespace XML documents, such as SVG (scalable vector graphics) (Chatty, Lemort, Sire, & Vinot, 2005) and XHTML (eXtensible HyperText Markup Language) (Musciano & Kennedy, 2003) for visual interaction, and VoiceXML (Voice Extensible Markup Language) (Lucas, 2000) for voice interaction. However, many other XML-derived languages have helped in the development of mobile services. The next target is the consideration of the mobile network as an extension of the global Internet network. This article explains the importance of XML and its dialects in a mobile application environment to enable their use by the “various applications/services” (today available on the Web). In fact, different dialects may be needed for different mobile devices depending on their characteristics (such as memory, CPU speed, integrated software engine, etc.). For example, two SVG profiles are defined for cellular phones and PDAs (personal digital assistants): SVG Tiny (SVGT) is suitable for the next generation of cellular phones especially, while SVG Basic (SVGB) is aimed at high-tech devices such as PDAs or smart phones (Andersson et al., 2003). The pervasive use of mobile devices will be the target for the near future (Branco, 2001), given the trend towards considering the mobile network as an extension of the Internet global network. This scenario promotes the development of new dialects for multimodal interaction through mobile devices. The dialects developed for speech, sketch, and visual
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
XML-Based Languages for Multimodality in Mobile Environments
interaction are discussed next. An area for future development might focus on interaction through gestures. XML (eXtensible Markup Language) is a simple, flexible, and powerful markup language, based on text format that allows the development of a potentially unlimited number of innovative multimodal services and applications. It was derived from the more complex, complete SGML (Standard Generalized Markup Language, ISO 8879) (Chamberlin & Goldfarb, 1987), designed for more general purposes. However, XML language is easier to manage, and is genuinely Web oriented and mobile oriented. In other words, XML language is an optimal subset of SGML, constructed in consideration of the possible Web services and applications. XML can be used to develop several languages taking the specific working context into account. It also plays an important role in the exchange of a wide variety of data, making them available and accessible by Web using computers and mobile devices.
BACKGROUND The World Wide Web is undergoing continuous development. This has enabled a great expansion in a wide variety of applications and services, covering every human activity. In addition, advanced technology mobile systems are becoming ever more complete, complex devices, which can offer a broad range of Web applications and services originally conceived for personal computers. These two factors explain the need to introduce various multimodal systems (services and applications oriented) to interact with the Web using mobile devices. In this context, the use of XML-based technology to create powerful, multi-purpose system interfaces is a winning choice. Meta-language XML allows ad hoc language solutions to be developed according to the specific argument. A multitude of XML “dialects” for multimodal solutions have already been developed, while any others are under current or future development. An exhaustive point of view on XML-based languages is thus not a simple matter. This section provides a panorama of the XML-based languages, considering the different interaction modes, multimediality, and multimodality features. SMIL (Solon et al., 2004) is a basic, developing technology that allows several multimodal environments to be implemented. It is not an “out and out” multimodal language. As in many multimodal environments, it is necessary to interact with several kinds of multimedia content; SMIL works in the background with many different multimodal applications and services. SMIL enables interactive audiovisual presentations to be easily produced. It is typically used to choreograph complex multimedia presentations, where audio and video streaming, images, text, graphics, and other media types
are combined in real time. In other words, SMIL makes it possible to manage the temporal and spatial constraints of multimedia presentations. The current SMIL conception offers modules for animation, content control, layout, linking, media objects, meta information, timing and synchronization, and transition effects. This modular approach allows reuse of SMIL syntax and semantics in other XML-based languages, especially those used for timing and synchronization. These fundamental features play a leading role in multimodal user interaction. SMIL is an easy-to-learn, HTML-like language and the World Wide Web Consortium (W3C) recommendation to achieve synchronized multimedia. In this context, it simplifies the creation of time-based multimodal interfaces with a high portability factor. SMIL is also exploited to aid the construction of powerful mobile-oriented multimodal applications and services. A classic example of SMIL use is to enable authors to specify and control the precise time a sentence is spoken and make it coincide with the display of a given image. This simple “technical pattern” is at the base of several multimodal general systems. To consider the different languages and problems connected with multimodality, we must take into account the different interaction approaches (visual, voice, etc.). For a visual approach, we must consider SVG. The SVG language (Chatty et al., 2005) is another important basic technology that works in background mode to resolve different types of problems in the multimodal interfaces. It describes twodimensional graphics and graphical applications in XML, providing facilities for document structuring, shape definition, painting, clipping and masking, compositing, text manipulation, styling, linking, scripting, animation, interactivity, integration of multimedia content, alpha masks, filter effects, and template objects. It supports object zooming, interaction and manipulation, and scene annotation, among others. SVG element features can be static or dynamic, and each complex element can be used in interactive mode. A strong point of this language is the ability to interact with script languages (such as javaScript, ECMAScript, JScript, etc.), which provide complete access to all elements, attributes, and properties necessary to develop powerful, complete 2D graphical representations. However, this latter point is not a binding factor. In fact, for example, an animation can be defined and triggered either declaratively (i.e., by embedding SVG animation elements in SVG content) or via scripting. Because of SVG’s ability to produce high-quality rich graphical displays, enable the development of highly interactive user interfaces, and manipulate the contents and structure of an SVG document, it is very well suited for the development of interactive multimodal applications and services. In fact, it is Web accessible by both personal computer and innovative mobile-oriented systems, and needs sophisticated multimodal environments to satisfy the user’s 1051
X
XML-Based Languages for Multimodality in Mobile Environments
requests. SVG recently became a standard on the Web and is now also becoming a standard for mobile devices. However, different technologies need different hardware supports, as each mobile device has its own characteristics (memory, CPU speed, integrated software engine, etc.). Two different profiles are defined for different device families. The first, low-level profile, SVG Tiny (SVGT), is especially suitable for next-generation cell phones, while the second, SVG Basic (SVGB), is aimed at high-tech devices (such as PDAs or smart phones). Another markup language, which always works in background mode, is XForms (Extendible Forms) (Bals, 2005). This represents the next generation of forms for the Web and is used for mobile systems. It permits describing general user interfaces, such as Web forms, interactive GUI (Graphical User Interface), special windows, and so forth. It is well structured and can also be used in a standalone manner to describe more complex user interfaces. The real innovation lies in splitting traditional XHTML forms into three parts: model, instance data, and user interface presentation. The construction of a graphical user interface is the most classic example of a multimodal expression. MPML (Multimodal Presentation Markup Language) (Zong, Dohi, & Ishizuka, 2000) is another important language developed to enable the description of complex multimodal presentations based on character agents. This markup language allows a new style of effective presentation of information and a new approach to the production of multimodal information content. These different multimodal scenarios use interactive lifelike agents with expressive capability (e.g., verbal conversations, facial expressions, body gesture explications, behavioral model applications, etc.) that guide users through the multimedia informative content. MPML language is structured on several defined features, which enhance the capability and potentiality of the multimodal applications and services created with it. Most importantly: •
•
• •
1052
It is Independent of the Character Agent System: It is designed such that authors can write multimodal presentation content independent of specific character agents. It is Easy to Describe: It is a simple and suitably designed XML-based elaboration designed for multimodal complete presentation only, providing a minimal set of tags to control the presentation and also allowing complete graphical representations. It Supports Interactive Presentation: It supports a set of mechanisms to guide the presentation interactively. It has Character Control: Character agents can guide the presentation.
•
It Supports Media Synchronization: It is in accordance with SMIL, thus enabling the combination and synchronization of various types of multimedia elements such as images, text, animations, audio/video streams, and so forth.
The huge advantage of this language is that authors can provide various features on a single character agent to enable it to adapt to new situations and interactions. MPML is easy to extend, enabling the expansion of normal application fields and design of multimodal functionality to more complex contexts. For example MPMLVR (Multimodal Presentation Markup Language for Virtual Reality) (Okazaki, Aya, Saeyor, & Ishizuka, 2002), which facilitates multimodal presentation in a 3D virtual space supported from an agent system in VRML (Virtual Reality Modeling Language), was designed by extending MPML. Another versatile extension is MPML-FLASH (Multimodal Presentation Markup Language with Character Agent Control in Flash Medium). Several attempts are ongoing to further widen fields for application of MPML technology. Another important multimodal XML-based language is EMMA (Extensible MultiModal Annotation Markup Language) (Marrin, Myers, & Aktihanoglu, 2005). Its creators (W3C Multimodal Interaction Working Group) aimed to develop specifications to enable access to the Web (or mobile environment) through multimodal interaction. EMMA is intended for use in systems that provide semantic interpretations for a variety of inputs, including but not necessarily limited to speech, natural language text, GUI, and ink input. Classic examples include prearranged keystroke sequence, speech recognition commands, touch-screen recognition commands, handwriting recognition, semantic vocal streaming understanding, semantic gesture streaming understanding, natural language understanding, interactive ink pens, pointing input tools, and so forth. There is a potentially unlimited possibility to induce numerous input signals from different input devices. EMMA is used primarily as a standard format for data interchange among the components of a multimodal system. It is automatically generated by interpretation components to represent the semantics of user inputs, not directly authored by developers. EMMA turns out to be useful in several contexts. It is independent from the specific application field and can be adapted to heterogeneous input types. Different complex devices have their own interaction methods, which can be separately managed to best interact with the user input, and the language provides a set of elements and attributes that focus on accurately representing annotations on the input interpretations. In other words the language focuses on multimodal semantic interpretation of the events that drive user interaction during the composition of one or more inputs.
XML-Based Languages for Multimodality in Mobile Environments
In general an EMMA document can be considered as consisting of three parts: instance data, data model, and metadata. Instance data is an application-specific markup corresponding to input information meaningful to the consumer of an EMMA document. Instances are built by input processors at runtime. Given that utterances may be ambiguous with respect to input values, an EMMA document may hold more than one instance. The data model imposes constraints on the structure and content of an instance. Metadata represent annotations associated with the data contained in the instance. Annotation values are added by input processors at runtime. EMMA has a large potential for integrating different devices in several environments. Furthermore, it will be able to make the Web fully accessible for future applications. There should be no difference in interacting with a user via phone DTMF (dual tone multiple frequency) tones, PDA ink pens, or even voice browsers for users with disabilities. This is an impressive goal, and the MIF (multimodal interaction framework, a complex system to support multimodal development of applications and services) is a next step in achieving it. Another multimodal markup XML-based language is InkML (Ink Markup Language) (Mohamed, Belenkaia, & Ottmann, 2004). It is utilized to build the data format used to represent ink entered with an electronic pen or stylus in a multimodal system. Its main purpose is to transfer digital ink data between devices and software components, and store hand-input traces for handwriting recognition, signature verification, and gesture interpretation. An increasing number of electronic devices with pen interfaces are now available. These are widely used for various applications and services, such as information, interactive services, semantic interpretation of gestures, multimodal “touched” interaction, and so on. It is also necessary to take into account that handwriting is a very familiar input mode for most users (with or without disabilities), who will thus tend to use it for input and control when available. Digital ink information is a delicate, complex, and difficult format to manage. Several studies have attempted its comprehensive and productive format standardization. These elements have caused a restricted growth of key technical factors such as digital ink capture, processing, and transmission of digital ink data across heterogeneous devices; communications among different ink management software programs; interactions between ink software applications and hardware devices; and so forth. InkML is also designed to resolve these problems. In fact, its important feature is to provide a simple, platform-independent data format to promote the interchange of digital ink among different kinds of software and hardware applications and services. An interesting peculiarity of this technology relates to its ability to capture and express the user’s dynamic behavior and relative semanticsthat is, the dynamics of the user’s
behavior during the interaction with, for example, the electronic pen on the tablet (to interpret hand-speed, break point, acceleration paths, etc.). To date, only a low information level has been used to interact with the user. It is to be hoped that future studies will examine this scenario. Clearly, this markup language is also useful in both a Web environment and mobile devices. It is becoming a useful standard (W3C recommendation) to tackle entire problematic situations on ink contexts. The VoiceXML (Lucas, 2000) language provides speech recognition and speech synthesis capabilities for the scripting of voice dialogs and enables integration with other processes through events and IP connectivity. In voice processing, it offers control of voice recognition capabilities through grammars, speech synthesis control, and some basic telephony control. It provides several voice response systems, such as recognition of spoken input and DTMF input. It also allows registration of spoken input and controlling dialogue flow, and can be used to transfer and disconnect telephone calls. VoiceXML also supports the output of synthesized speech and audio files. This language supports audio file formats, speech grammar formats, and URI schemas. In order to permit speech applications, VoiceXML is based on a grammar format, called speech recognition grammar specification (SRGS) (Hunt & McGlashan, 2003). SRGS is used by speech recognizers and allows developers to specify the words and word patterns to be listened for by the speech recognizer. VoiceXML provides semantic interpretations from grammars and makes this information available to the application. It also permits a separation of service logic from interaction behavior. In IP connectivity VoiceXML allows precise identification of which data to submit to the server, and which HTTP method, GET or POST, to use in the submission. SSML (Speech Synthesis Markup Language) (Burnett, Walker, & Hunt, 2004) is an XML-based language that presents elements for controlling the pronunciation, tone, inflection, and other characteristics of spoken words. It captures text speed, volume, inflection, and prosody in order to convert it to acoustic speech through a text-to-speech (TTS) synthesis engine. SSML’s main features are interoperabilityor interaction with other markup languagesand consistency; in fact, it provides control of voice input across platforms and across speech synthesis implementations. SALT (speech application language tags) (Cisco Systems, Comverse, Intel Corporation, Microsoft Corporation, & Philips Electronics, 2002) is an extension of HTML and other markup languages, integrating speech and telephony interfaces with Web applications and services. In multimodal applications it supports speech input and output in visual pages. It is used where a speech interface is available in addition to the visual interface. 1053
X
XML-Based Languages for Multimodality in Mobile Environments
MAIN FOCUS OF THE ARTICLE The increasing diffusion and utility of the global network has enabled the spread of applications and services supporting every human activity. These involve both scientific domains, such as online technical services, and typical entertainment domains. The widespread use of mobile devices has improved the diffusion of multifunctional complex tools satisfying sophisticated user applications and increases the spread of innovative services. In mobile applications, the “audio/video Multilanguage dictionary” enables the user to immediately discover the meaning of idiomatic expressions and difficult words. The multimedia stream allows the meaning of “sentences” in real contexts to be understood, with practical examples and situations. In this section we discuss some of the most well-known and useful XML-derived languages used to develop mobile services, and give a table that describes these languages according to their modes, functions, and applications/services. SMIL is the W3C recommendation to synchronize multimedia on the Web. This XML-based language is used to write interactive multimedia presentations, enabling management of their temporal and spatial constraints. In mobile phones, it enables lightweight multimedia functionality and integrates timing into profiles such as WAP forum’s WML language and XHTML Basic, while MMS is a subset of SMIL for mobile telephone multimedia messaging. The W3C group’s Voice Browser and Multimodal Interaction are working to standardize XML languages. VoiceXML is used to create voice user interfaces with automatic speech recognition (ASR) and text-to-speech synthesis, audio dialogues that feature synthesized speech; digitized audio; recognition of spoken and DTMF key input; recording of spoken input, telephony, and mixed-initiative conversations; interactive voice response (IVR); and so forth. Many applications and mobile services exploit the potential of VoiceXML,
and there are numerous consolidated contributions in this field. In fact, VoiceXML and SALT are valid instruments to support the integration of visual representation enriched systems and visual browsers, with instruments such as XHTML, cascading style sheet (CSS), SMIL, and SVG. Similar areas are handled by SALT. This language is being designed to “extend existing markup languages such as HTML, XHTML, and XML.” An important difference between SALT and VoiceXML is the overall approach used to develop applications. Whereas VoiceXML is essentially declarative, using its extensive set of tags, SALT is very procedural and script oriented, with a very small set of core tags. The factor connecting them is that multimodal access enables users to interact with an application in a variety of different ways. Obviously in this case too, there are a number of practical demonstrations. New trends are developing to allow intermodal interaction by instruments such as InkML, which captures pen movements and enables data exchange by “digital ink.” This mark-up language has an XML data format to describe digital ink data from the pen or stylus in a multimodal system. A pen-based interface captures digital ink and pen movements by a transducer. The digital ink can be analyzed by recognition software to convert pen input into defined computer actions. Alternatively, it can be stored in ink documents, messages, or notes for later retrieval or exchange through telecommunications. EMMA is used to implement semantic interpretations of a great variety of inputs, such as voice, text languages, and digital ink. The Speech Services Control (SpeechSC) working group of the Internet Engineering Task Force (IETF) develops protocols to support distributed speech recognition and synthesis and speaker verification services, and expects to take advantage of W3C’s work on speech recognition grammar specification (SRGS), SSML, and semantic interpretation (SI). Table 1 summarizes the main features of each multimodal XML-based language. The table includes modalities that
Table 1. Main features of multimodal XML-based languages XML-Based Dialects
Modalities
Functions
SMIL
Synchronization, Real-time combining, Audio, Video, Images, Managing temporal and spatial conText, Graphics straints, Streaming, Timing
Animation, Content control, Layout, Linking, Media objects, Meta information, Transition effects, Web-oriented multimodal application Document structuring, Definition of shape, Painting, Clipping, Masking, Compositing, Text manipulation, Styling, Linking, Animation, Alpha mask, Filter effects, Zooming, Scene annotation, Object manipulation and interaction Web and mobile environment applications, Writing multimodal presentation content independent of specific character agents, Providing a minimal set of tags to control presentation, Interactive presentation guidance
SVG (SVG TinySVGTfor nextgeneration cell phones, SVG BasicSVGBfor high-tech devices)
Graphics
Interactivity, Integration of multimedia content, Two-dimensional graphics, Graphics applications in XML
MPML (MPML-VR, MPML-FLASH)
Integration complex multimedia, Streaming management
Description of complex multimodal presentations based on character agents, Media synchronization
1054
Application and Services
XML-Based Languages for Multimodality in Mobile Environments
Table 1. continued
EMMA
Speech, Natural language, Text, GUI, Ink input
Semantic interpretations for a variety of inputs
Prearranged keystroke sequence, Speech recognition commands, Touch-screen recognition commands, Semantic vocal streaming comprehension, Semantic gesture streaming comprehension, Natural language streaming comprehension, interactive ink pens
X-Form
Interactive complex form
Performing several kind of data manipulation tasks,
Web forms, Interactive GUI, Special windows
Graphics
Building data formats used to represent digital ink entered with an electronic pen or stylus; Transferring digital ink data among devices and software components; Storing hand-input traces for handwriting recognition, signature verification, and gesture interpretation
Providing a simple and platform-neutral data format to promote the interchange of digital ink among different kinds of software and hardware applications and services, Information entry and manipulation, Use of applications, Use of interactive services, Semantic interpretation of gestures, Multi-touch interaction
VOICEXML
Voice
Speech recognition, Speech synthesis, Control of voice recognition capabilities through grammars, Speech synthesis control
Recognition of spoken input and DTMF input, recording spoken input, and controlling dialogue flow; Transferring and disconnecting telephone calls; Synthesizing speech and audio files; Identifying exactly which data to submit to the server in IP connectivity
SSML
Voice
Multimedia streaming management
Controlling the pronunciation, tone, inflection, and other characteristics of spoken words; Capturing text speed, volume, inflection, and prosody
SALT
Voice
Integration of speech and video
Integrating speech and telephone interfaces with Web applications and services, Supporting speech input and output in visual pages
InkML
deal with each XML-based dialect, the specific supported functions, and applications and services in Web and mobile environments.
FUTURE TRENDS Control and interaction with the environment providing different services, and the new need of different devices such as personal computers, mobile phones, PDAs, TVs, and so on, will improve development of standard solutions and languages. In particular a greater development of a new generation of mobile devices, with different interaction modes according to their pervasive use and services, will produce an increasing interest about the development of the XML-based languages about different interaction mode and integration modes.
CONCLUSION The emergence of new devices for human-computer interaction and communication characterized by various sizes of displays, voice-based interaction, new devices for handwriting, and character input are all elements that require the user
interface to be portable and to support multimodal interaction. That is, XML portability and adaptation to the different devices provide a solution with the different languages that were developed. This article provides an overview of the evolution of the multimodal XML-based languages used for mobile applications and services. The main focus is on specific applications of each language that are available on the Web environment and mobile devices. The use of XML and XML-based languages allows solving multimodal problems in several application fields. Special attention has to be considered for the activity of the World Wide Web Consortium, which is devoted to developing innovative XML technologies (that are standard de facto) in order to make the Web the focal element for data interchange, interoperability, and accessibility to the information according to the different interaction modes and with the different devices.
REFERENCES Andersson, O., Axelsson, H. et al.(2003). Mobile SVG profiles: SVG Tiny and SVG Basic. W3C Recommendation.
1055
X
XML-Based Languages for Multimodality in Mobile Environments
Bals, K. (2005). Using XSL, XForms and UBL together to create complex forms with visual fidelity. Proceedings of the XML Conference.
Conversational Characters: Applications, Methods, and Research Challenges (in conjunction with HF2002 and OZCHI2002), Melbourne, Australia.
Branco, P. (2001). Challenges for multi-modal interfaces towards anyone anywhere accessibility: A position paper. Proceedings of the Workshop on Universal Accessibility of Ubiquitous Computing: Providing for the Elderly.
Paternò, F. (2004). Multimodality and multi-device interfaces. Proceedings of the W3C MMI Workshop on Multimodal Interaction.
Bray, T., Paoli, J., Sperberg-McQueen, C. M., Maler, E., & Yergeau, F. (2004). Extensible Markup Language (XML) 1.0 (3rd ed.). W3C Recommendation. Burnett, D., Walker, M., & Hunt, A. (2004). Speech Synthesis Markup Language (SSML) version 1.0. W3C Recommendation. Chamberlin, D., & Goldfarb, C. F. (1987). Graphic applications of the Standard Generalized Markup Language (SGML). Computers & Graphics, 11(4), 343-358. Chatty, S., Lemort, A., Sire, S., & Vinot, J.L. (2005). Combining SVG and models of interaction to build highly interactive user interfaces. Proceedings of the 4th Annual Conference on Scalable Vector Graphics (SVG Open 2005), Enschede, The Netherlands. Cisco Systems, Comverse, Intel Corporation, Microsoft Corporation, & Philips Electronics. (2002). SALT.1.0.doc©. SpeechWorks International Inc. Hunt, A., & McGlashan, S. (2003). Speech recognition grammar specification version 1.0. W3C Proposed Recommendation. Kvale, K., Warakagoda, N. D., & Knudsen, J. E. (2003). Speech centric multimodal interfaces for mobile communication systems. Telektronikk. Lucas, B. (2000). VoiceXML for Web-based distributed conversational applications. Communications of the ACM, 43(9), 53-57. Marrin, C., Myers, R., & Aktihanoglu, M. (2005). Emma: An extensible multimedia architecture. Emma White Paper. Retrieved from emma.3d.org Mohamed, K. A., Belenkaia, L., & Ottmann, T. (2004). Post-processing InkML for random-access navigation of voluminous handwritten ink documents. Proceedings of WWW, Alternate Track Papers & Posters (pp. 266-267). Musciano, C., & Kennedy, B. (2003). HTML & XHTML: The definitive guide. Inf. Res., 8(2). Okazaki, N., Aya, S., Saeyor, S., & Ishizuka, M. (2002). A multimodal presentation markup language MPML-VR for a 3D virtual space. Proceedings of the Workshop on Virtual
1056
Solon, A., Mc Kevitt, P., & Curran, K. (2004). Mobile multimodal dynamic output morphing tourist systems. Intelligent Multimedia Research Group, University of Ulster, Magee Campus, Northern Ireland. Zong, Y., Dohi, H., & Ishizuka, M. (2000). MPML 2.0e: Multimodal presentation markup language supporting emotion expression. Department of Information and Communication Engineering, School of Engineering, University of Tokyo, Japan.
KEY TERMS Extendible Form (XForm): A W3C markup language representing the next generation of forms for the Web and mobile systems, used to describe general user interfaces. Extensible Multimodal Annotation Markup Language (EMMA): An XML-based W3C recommendation used for information automatically extracted from the user’s input. It focuses on annotating the interpretation of input information. Multimodal Presentation Markup Language (MPML): A W3C markup language for use in systems that provide semantic interpretations for a variety of inputs. Ink Markup Language (InkML): A W3C language to describe ink data acquired with an electronic pen or similar device. Scalable Vector Graphics (SVG): A W3C markup language describing a two-dimensional vector graphic. It enables shapes, raster graphics images/digital images, and text. Speech Application Language Tag (SALT): A W3C markup language used for adding voice recognition to HTML and XHTML files. Speech Synthesis Markup Language (SSML): An XML-based W3C recommendation for speech synthesis. Synchronized Multimedia Integration Language (SMIL): A W3C-recommended XML-based markup language for multimedia presentations. VoiceXML: The W3C standard XML format for specifying voice dialogues between a human and a computer.
Index
Index
Symbols [mK]M 521 [mKM] 521 2G (see second generation) 3APL-M 244 3G (see third generation) commercial deployment 940–946 mobile technology 632 network 778 3GPP (see 3rd Generation Partnership Program) 3rd Generation Partnership Program (3GPP) 876 4G (see fourth generation) 7DS (see seven degrees of separation) 802.11 WiFi 90 A A-GPS (see assisted-GPS) A/I (see atomicity and isolation) AAA (see authentication authorization accounting) AAC (see augmentative-alternative communication) academia 365 academic PN (AcPN) 1, 7 access 1004 control 227 , 832 discovery 812 layers 624 network 873 point (AP) 172, 195, 201 selection 812 accessibility 9, 27 accounted attributes 334 accoustic data channel 15
ACK (see acknowledgement) acknowledgement (ACK) 395 ACL (see asynchronous connectionless) AcPN (see academic PN) active credibility 25 element (sensor) 868 server page (ASP) 912 registry (AR) 438 Active Bat location system 863 ad-hoc 693 computer network 393, 395, 397 network 397, 1029 on demand distance vector (AODV) 427 way 878 adaptive multi-rate voice codec (AMR) 726 real-time application 24 signal processing 356, 620 adaptivity 243 ADC (see analogue-to-digital converter) address notification 257 addressing 882 adjacent relationship query 791 admission controller 196 ADSL 418 advanced mobile phone system (AMPS) 266 video coding (AVC) 672 adventure game 929 advertising 886 service 72, 75 AeroScout 774
Copyright © 2007, Idea Group Inc., distributing in print or electronic forms without written permission of IGI is prohibited.
Index
agent 218, 386, 434, 831 -based inter-tree handover 544 freeze 431–432 migration pattern 994 of LSIS 790 platform (AP) 243 pre-activation 432 receptionist 431 transport 431, 433 aggregator 503 AGM (see anycast group membership) Agora 436 AHP (see analytic hierarchy process) AKC (see area key controller) AKE (see authenticated key exchange) alert services 713 allowable operations 950 ALM (see application-level multicast) ALN (see application-level networking) ALSAR (see application load sensitive anycast routing method) ambient intelligence 243 Wood project 525 American customer satisfaction model (ACSM) 143 Foundation for the Blind 569 AMR (see adaptive multi-rate voice codec) analogue-to-digital converter (ADC) 19, 669 analysis 791 analytic hierarchy process (AHP) 12, 28, 183, 376 angle of arrival (AOA) 511, 942 animation 672 annotation and metadata 602 anonymity 840 antenna beamforming 731 anthropology 519 anycast group membership (AGM) 54 open shortest path first (AOSPF) 52 any source multicast (ASM) 541, 592, 594 AOA (see angle of arrival) AODV (see ad-hoc on demand distance vector) AoI (see areas of interest) AOSPF (see anycast open shortest path first) AP (see access point) API (see application programming interface) application 459, 998 -level multicast (ALM) 395, 397 network (ALN) 397 networking (ALN) 395 security 801 broker 386
load sensitive anycast routing method (ALSAR) 52 module 342 programming interface (API) 250–251, 344, 514, 754 server 388, 727 service provider (ASP) 1006 system 181 tunneling (see also port forwarding) 537, 540 AR (see augmented reality) architecture 899 -specific pattern 745 area key controller (AKC) 834 areas of interest (AoI) 390 ARPU (see average revenue per user) artificial limb 651 neural network 69 Asia-Pacific 906–911 ASM (see any source multicast) ASP (see application service provider) ASR (see automatic speech recognition) assertion 59 assisted-GPS (A-GPS) 511, 770 assistive technology (AT) 352, 356, 403, 616, 620 associate identification (AID) 1025 associating 632 asymptotic cost 738 asynchronous connectionless (ACL) 344 transfer mode (ATM) 142 AT (see assistive technology) ATM (see asynchronous transfer mode) atomicity and isolation (A/I) 949 attributes 60 audio memo 757 augmentative and alternative communication (AAC) 352, 356, 574, 616, 620 augmented reality (AR) 207, 212 authenticated key exchange (AKE) 1021 authentication 581, 708, 840–841, 970 authorization accounting (AAA) 802 header 255 auto-configuration 253 automated speech recognition (ASR) 646 automatic parking equipment 403 profile comparison 251 speech recognition (ASR) 677, 1054 autonomic computing system 64 manager 64 autonomy 243 availability, 622 analysis 238 AVATON 387
Index
average perception 415 revenue per user (ARPU) 138 B backward handoff 969 secrecy 232 , 837 Bakhtin 519 bandwidth 124, 268, 459, 489, 627, 805 probing 788 reserved per request 983 utilization 788 banner advertising 636 barcode 898 mobile coupon 643, 888 baseline profile (BP) 592 base station (BS) 172, 739, 743, 810, 950, 980 basic object adapter (BOA) 162 sign-off 235, 951 batching 369 baton handover 944 Bayesian network (BN) 865–868 BB (see broadband) belief-desire-intention (BDI) 244, 315 BGCF (see breakout gateway control function) bi-directional tunneling 543 binary -phase-shift keying (BPSK) 273 runtime environment for wireless (BREW) 309, 503 binding 256, 259 acknowledgement 259 cache 257 request 258 update (BU) 257–259 message 779 biometric security 561 user authentication 841 Blackberry 572 blended learning 905 blind 655 block-based method 890 blog 901 blogging 905 Bluetooth, 118, 249–250, 272, 341, 394, 488, 507 749, 804–809, 820, 825, 839, 1011 device address 825 link 344 Network Encapsulation Protocol (BNEP) 805 BN (see Bayesian network) BOA (see basic object adapter) board game 929
body motion 406 control system 403 bookmarking 79 bootstrap agent transport 433 border node 328 BP (see baseline profile) BP-Services 84 BPSK (see binary-phase-shift keying) Braille 570 brain computer interfacing 68–70 branching 989 break 603 breakout gateway control function (BGCF) 874 BREW (see binary runtime environment for wireless) bricks-and-mortar store 312 bridging 1044 broadband (BB) 458 broadband wireless access (BWA) 921 broadcast 402 communication 399 environment 159 broadcasting 400 broker pattern 745 browser 78–83 -based architecture 646 phone 476 software 34 browsing 346, 488 behaviour 603 BS (see base station) buffering 970 Build-a-PC 323 built-in wireless technology 178 business -to-business (B2B) commerce 33 -to-consumer (B2C) commerce 33 environment 334 logic layer 262 BWA (see broadband wireless access) C CABAC (see context-adaptive binary arithmetic coding) CAC (see call admission controller) cache coherence (see also invalidation strategy) invalidation strategy 104, 107 management strategy 107, 159, 854 replacement 750–751 caching 107, 154, 159, 173–177, 369, 854 call admission controller (CAC) 197, 201
Index
-down system 394 session control function (CSCF) 874 caller/cellular authentication and voice encryption (CAVE) 691 CAMP (see core-assisted mesh protocol) capacity 357 CAR (see correspondent anycast resopnder) card emulation 707 care-of-address (CoA) 53, 255 cargo tracking 660, 907–908 Carnival 516–517, 519 carrier sense (CS) 329 medium access with collision avoidance (CSMA-CA) 273, 276 sensor multiple access (CSMA) 329 cartographic data 390, 392 cascading style sheet (CSS) 377, 1054 CAVE (see caller/cellular authentication and voice encryption) CAVLC (see context-adaptive variable-length coding) CBSD (see component-based software development) CDC (see connected device configuration) CDMA (see code decision multiple access) CDR (see common data representation) cell 194, 739, 743 -ID 770 of origin (COO) 511 cellular architecture 1038 phone 895 cloning 688, 691 tower 691 technology 415 cenceptualization 284 centralized P2P 89 service provision 880 central processing unit (CPU) 318, 585 central scheduling configuration message (CSCF) 927 message (CSCH) 927 certification authority (CA) 582 channel allocation 732 state information (CSI) 166 character agent system 1052 control 1052 chat 778 check-out mode 234, 951 childcare 660 cHTML (see compact HTML) civic structure 935–936
classification 377 CLDC (see connected limited device configuration) CLIC 522 client 219, 388 -server application 260 network 954 site 790 /server model 397 -side handheld computing 302 programming 309 clone 59 cluster 7 gateway 738 clustered model 948 clustering (grouping) 203, 205, 347 CN (see correspondent node) CoA (see care-of-address) code decision multiple access 329 division multiple access (CDMA) 96, 456, 503, 946 mobility 576, 578, 580 Code Division Multiple Access 2000 (CDMA 2000) 535 cognitive dimensions of notations 14 load 1002 collaborative learning 901 signal processing 204–205 collection layer 864–865 collision avoidance 329 scheme 924 combinatorial optimization 683 common spatial information services interfaces (CSISI) 790 comma separated value (CSV) 366 commerce 38 common data representation (CDR) 160 communicating sequential process (CSP) 580 for Java (JCSP) 576 communication 901 availability 234 cost 230, 238 distance 1046 efficiency 835 network 369 community of practice (CoP) 529, 532 support 714 Compact Flash (CF) card 770 compact HTML (cHTML) 376
Index
compatibility 413–414, 418 test 721 compensating mode 948 complements 12 complexity 413 analysis 738 component -based software development (CBSD) 58, 61 configurator 747 model 61 composite attack 829 computation cost 231 efficiency 836 tree logic (CTL) 994 computational resources 877 computed radiographic (CR) 533 tomographic (CT) 533 computer -based cognitive tool 318, 327 -supported collaborative work (CSCW) 634, 934 computing system 734 concept match 377 concurrency control 234 conference communication 589 signaling 590 conferencing service provider 152 confidence score 337 confident 573 confidentiality 581, 840 congestion control 812 connected device configuration (CDC) 305, 980 limited device configuration (CLDC) 305, 793, 980 constrained optimization 683 construction 519 rebroadcast 751 consumer 9, 312 -centric approach 461 demand 78 consuming services 72, 75 content -based taxonomy 928 -sharing software 960 adaptation 75 management 118 optimization 788 page 366 piracy 534 provider (CP) 392, 982, 986 server 388 transformation 123
context management 800 context 7, 118, 398, 402, 716, 722 -adaptation 124 -adaptive binary arithmetic coding (CABAC) 592 variable-length coding (CAVLC) 592 -aware application 137 computing 510, 769 network system 138 security 800 -enabled adaptive service 140 -sensitive profile 251 service logic 141 awareness 124, 130, 711 information 116 metadata 116–118 contiguity 517 continuity 57 continuous query 665 table (CQT) 662 contrast sensitivity functions (CSF) 760 control 882 list 1030 module 354 , 618, 630 plane 873 controlling PoC function 727 convenience 284, 290, 398 convergence technology 149–153 COO (see cell of origin) cookies 79 cooperative caching 154, 159 data dissemination 750 CoP (see community of practice) CORBA® 436 core-assisted mesh protocol (CAMP) 395 corporate social responsibility (CSR) 469 correspondent anycast responder (CAR) 54 node (CN) 256–257 cost concerns 181 coverage analysis 791 CP (see content provider) CPU (see central processing unit) CQT (see continuous query table) CR (see computed radiographic) credibility 25 credit card use 416 Cricket 774 Location Support System 863 crops management 763
Index
cross-layer protocol engineering 166 crossing 739 crossword puzzle 323 cryptography 831, 979 CS (see carrier sense) CSCF (see call session control function) CSCH (see central scheduling) CSCW (see computer-supported collaborative work) CSISI (see common spatial information services) CSMA (see carrier sensor multiple access) CSMA-CA (see carrier-sense medium access with collision avoidance) CSP (see collaborative signal processing; client/server paradigm) CSP (see communicating sequential processes) CSS (see cascading style sheet) CSV (see comma separated value) CT (see computed tomographic) culture 313 customer priorities 500 profiling 99 service 442, 473 support 500 customization 463, 652 cyclic redundancy code (CRC) 840 D DAA (see data access agent) DAB (see digital audio broadcast) DAD (see duplicate address detection) DARPA (see Defense Advanced Projects Agency) data -transfer rate 534 access 906 agent (DAA) 947 and information readiness 484 attack 828 caching 172–177 collecting infrastructure 203–205 collection 563 compression 535 dupression 121 fusion 209, 676, 868 integrity 840 protection protocol 827 layer 262 management 536 sharing 181, 219 sources description information (DSDI) 214 usefulness 180–181 database management system (DBMS) 303, 974 query system 369
DBMS (see database management system) DBN (see dynamic Bayesian network) DB partition 951 DCA (see dynamic channel allocation) DCF (see distributed coordination function) decision support 12, 28 decomposition 791 Defense Advanced Projects Agency (DARPA) 427 degrees-of-freedom (DOF) 212, 656 delay jitter 24 delivery context 14, 26, 30, 380 demodulation 17 denial of service (DoS) 534, 839, 1024, 1028 Department of Defense 32 dependability 233, 800 description logics (DL) 44, 50, 377 Design for All 861 desktop PC 3 destination options header 255, 258 detected attribute 334 detection 204 developer 503 device heterogeneity 75 dexterous hand 653–654 DFSK (see differential FSK) DHCP (see dynamic host configuration protocol) DHT (see distributed hash table) diagonal replication grid (DRG) 234 dial-up 1023 diameter 802 DICOM (see digital imaging and communications in medicine) different handoff scheme 624 differential FSK (DFSK) 16 phase-shift keying (DPSK) 272 diffusion 413 digital assistant 246 audio broadcast (DAB) 455 battlefield 660 cellular phone 475 certificate 434, 827, 831 extended broadcast 613 imaging 318 and communication in medicine (DICOM) 533, 540 ink 679 literacy 568 radiographic (DR) 533 technique 942 technology-mediated communication 562 technology (2G) 999 video 671 encoding 889
Index
DII (see dynamic invocation interface) direct migration technique 120 sequence spread spectrum (DSSS) 273, 276 dirt 519 disability-centered organization 569 discovery 882 process 209 discrete cosine transform (DCT) 594, 629 display 536 modification 79 module 354, 618 distance learning (d-learning) 423 distraction 517 distributed client architecture 646 computing 202, 205 system (DCS) 436 coordination function (DCF) 195 databank 624 detection and estimation 204–205 hash table (DHT) 494, 954, 960 provision 881 query and search 203–205 scheduling message (DSCH) 927 speech recognition (DSR) 645 target tracking 204–205 tracking environment 209 DL (see description logics) DMS (see database management system) dots per inch (dpi) 670 DOF (see degrees-of-freedom) dot.com bust 35 download 488–489, 707 DPSK (see differential phase-shift keying) DR (see digital radiographic) DRG (see diagonal replication grid) DSCH (see distributed scheduling) DSDI (see data sources description information) DSI (see dynamic skeleton interface) DSR (see dynamic source routing) DSSS (see sequence spread spectrum) DTMF (see dual tone multiple frequency) dual-slot mobile phone technology 475 dual tone multiple frequency (DTMF) 1053 duplicate address detection (DAD) 259 dynamic applications suitability 57 Bayesian network (DBN) 863, 869 channel allocation (DCA) 944 content adaptation 624 environment 877 home agent address discovery 53 host configuration protocol (DHCP) 882
invocation interface (DII) 161 routing protocol 425 skeleton interface (DSI) 162 source routing (DSR) 427 trust relations 798 weighing 651 dynamically discoverable 878 E e
-commerce (see electronic commerce) -coupon 299 -government 581 -learning advancement 419 E-OTD (see enhanced observed time difference) ease of service provisioning 140 use 414, 418 eavesdropping 1024 EDGE (see enhanced data rates for global evolution) education 633 EFR (see enhanced full rate) EG (see event grammars) EGPRS (see enhanced GPRS) electric lock interface 1013 electronic commerce (e-commerce) 32, 34, 36, 38, 41, 96, 108, 283, 311, 339, 435, 831, 974 learning (e-learning) 423, 525 product code (EPC) 183, 819 serial number (ESN) 691 service guide 613 wallet 636–637 embedded system 260 EMMA (see Extensible MultiModal Annotation Markup Language) employee readiness 485 enabled task 1047 encrypted tunnel 1032 encryption 435 end-to-end QoS 802 end user 7 energy model 333 enhanced data for global evolution (EDGE) 535 full rate (EFR) 726 GPRS (EGPRS) 726 observed time difference (E-OTD) 511 enhancement 483 enterprise readiness for mobile ICT 486 transformation 486 entertainment 633 entity 118
Index
entry permit 431 environment -specific inter-ORB protocols (ESIOPs) 160 properties 10 EPC (see electronic product code) equalization 165 escape from formalism 11 ESIOP (see environment-specific inter-ORB protocol) ESN (see electronic serial number) event 604 control action pattern 747 grammars (EG) 208 eventing 882 evictor pattern 746 execution sequence 988 experienced credibility 25 experimental verification 787 expertise 13, 25 explicit 378 exploratory factor analysis (EFA) 299 exposed-terminal problem 168 extended profile (XP) 592 extendible form (XForm) 1056 eXtensible HyperText Markup Language (XHTML) 376, 1050 Markup Language (XML) 89, 142, 376, 456, 675, 912, 1050 MultiModal Annotation Markup Language (EMMA) 1052, 1056 extensible authentication protocol (EAP) 1031 extension header 254 eye tracker (ET) 677 F facade pattern 746 FAR (see furthest away replacement) fast fourier transform (FFT) 16, 19 learning 526 fault tolerance (FT) 537, 622, 925 Federal Communications Commission (FCC) 769 FCFS (see first come first serve) FDD (see frequency division duplex) FDDI (see fiber distributed data interface) FDMA (see frequency division multiple access) Federal Communications Commission (FCC) 569 federated identity management 876 FFD (see full function device) FFT (see fast fourier transform) FHSS (see frequency-hopping spread spectrum) fiber distributed data interface (FDDI) 142 FIFO (see first in first out)
fighting game 928 financial security 285, 290 first come first served (FCFS) 447 generation (1G) game 187 in first out (FIFO) 447 -person shooting (FPS) 929 fission 645 fixed infrastructure 738 price charging 549, 552 FL (see foreign link) fleet management 660 flexibility 149, 435, 831 FlexiSPY 557 flooding 395 foreign device 7 link (FL) 256 formal learning 532 forward forward handoff 969 secrecy 232, 837 forwarder-receiver pattern 745 forwarding zone 395 fourth generation (4G) 733 game 187 mobile system 682 fragment header 255 frame rate 24 framework model 370 FreeTV 611 frequency -hopping spread spectrum (FHSS) 272, 276 division duplex (FDD) 940 multiple access (FDMA) 329 resolution 19 response 19 shift keying (FSK) 16–19 Fresnel Zone 1028 FSK (see frequency shift keying) FT (see fault tolerance) full function device (FFD) 273 functional module 303 furthest away replacement (FAR) 740 fusion 645 layer 864–865 future prosecution 978 fuzzy expert system 763 logic 763
Index
G GALILEO 770 GAMA (see generic adaptive mobile agent) gambling 488, 554 gamer 928 gaming 294, 914 gaming GAP (see generic access profile) gate of the logical network 53 gateway GPRS support node (GGSN) 874 Gaussian frequency shift keying (GFSK) 272 gaze tracking (GT) 677 GDBS (see global database system) GDM (see generative domain model) gender 296–301 general design pattern 745 inter-ORB protocol (GIOP) 160, 164 message format 430 packet radio service (GPRS) 142, 194, 503, 505, 510, 515, 726 generative domain model (GDM) 208 generic access profile (GAP) 344 adaptive mobile agent (GAMA) 717 log adapter (GLA) 64 genetic algorithm (GA) 339, 349 geo-referenced information (GRI) 213 geocast 393, 395, 397 geocasting-limited flooding 395–396 geographical content database 388 database 385 geographic information system (GIS) 129, 137, 219, 856, 907 Geography Markup Language (GML) 134, 137 geolocation information 773 geometric models 857 gesture recognition (GR) 677 GFSK (see Gaussian frequency shift keying) GGSN (see gateway GPRS support node) GIOP (see general inter-ORB protocol) GIS (see geographic information system) GKS (see group key server) global cache hit 155 miss 155 computing environment 57 database system (GDBS) 948 mobile system (GSM) 504–509 optimization 683 positioning
satellite 327, 739 system (GPS) 125, 130, 137, 203, 394– 395, 462, 511, 637, 662, 769, 856, 885, 898, 907, 1024 spatial information services (GSIS) 790 system for mobile communication (GSM) 194, 356, 456, 47 6, 515, 620, 726 transaction (GT) 948 coordinator (GTC) 949 manager (GTM) 949 GOEXP (see generic object exchange profile) GOS (see grade of service) GOVOREC 573 GPRS (see general packet radio service) GPS (see global positioning system) grade of service (GOS) 446 graphic interchange format (GIF) 670 graphics 671 GRI (see geo-referenced information) Grid index information service (GIIS) 437 information resource service (GRIS) 437 grid service 77 group communication 227 key management algorithm 228, 833–834, 834 server (GKS) 228 GSM (see global system for mobile communication) GSIS (see global spatial information services) 790 GT (see global transaction) GTC (see global transaction coordinator) GTM (see global transaction manager) guest 998 H HA (see home address) HA (see home agent) hacker 561, 1023, 1029 hand controller 654 handheld 611 computing 309 device 908 firewall 840 terminal (HHT) 894 handoff 230, 621, 834, 837 user list 230 handover 967 hands-free 675 accessory 571 handwriting recognition (HR) 646, 677, 755
10 Index
harmonic distortion 19 hash function 1021 attack 688 hashing 961, 965 HCI (see human-computer interaction) head controller 656 headhunter (HH) 438 health promotion 634 healthcare 504 challenge 1010 organization 1004 hearing accessories 571 helper application 625 heterogeneous device 387 multimedia network 796 wireless technologies 796 HHT (see handheld terminal) hidden-terminal problem 168 hierarchical structure (see also tree structure) 832 high -end consumer device 978 device 305 -featured mobile phone device 4 available server cluster 537 histogram 890 HL (see home link) HLS (see homeland security) HN (see home network) home address (HA) 53, 256–257 agent (HA) 53, 256 link (HL) 256 location agent 969 network (HN) 53 PoC server 727 subscriber server (HSS) 874 homeland security (HLS) 393 hop-by-hop options header 254 horizontal handover 967 interaction 165 mobile business 442 host 998 auto-configuration 53 hotspot 44 HSS (see home subscriber server) HTML 671 HTTP (see hypertext transfer protocol) human -centered LBS 856 -computer interaction (HCI) 68, 676, 935
interface 313, 403 -machine interaction 651–659 behavior 652 interfacing 804 properties 9 visual system (HVS) 760 hybrid coupling 810 data dissemination 749 intelligent system (HIS) 763 model 857 mobile device 138 P2P 89 Hyper-Text Markup Language (HTML) 456 HyperLan 805 hypertext transfer protocol (HTTP) 376, 1023 I i-menu 296 i-mode 643, 871, 888, ICT (see information and communication technology) ID (see identification number) card 477 identification 430 number (ID) 826 identity management 798 vector 967 IDL (see Interface Definition Language) IDT (see innovation diffusion theory) IGO (interactive graphing object) 322 iGrocer 312 image 414, 418 layer 389 resolution 535 transmission 535 imaging 670 IMC (see integrated marketign communications) IMEI (see international mobile equipment identifier) 692 immutability 522 implementation 829 neutrality 878 improved mobile telephone system (IMTS) 266 IMS (see interactive multimedia system) IMS (see IP multimedia subsystem) incentive-based marketing 934 inconsistency ratio (IR) 180 incremental evaluation 665 individual view 935 indoor navigation ontology (INO) 858 positioning 859, 861 spatial model 859 informal learning 532
Index 11
information and communication technology (ICT) 33, 319, 466, 4 81, 516, 519, 563, 568 availability requirement 191 security requirement 191 services 789, 795 technology (IT) 283, 466 infotainment 593 infrastructural pattern 745 infrastructure element 869 sensor 865 initial link 87 initialization vector (IV) 1022 initializing vector (IV) 1030 Ink Markup Language (InkML) 1053, 1056 InkML (see Ink Markup Language) innovation-diffusion theory 33 innovation diffusion theory (IDT) 33, 894 innovativeness 413 INO (see indoor navigation ontology) 858 input module 353 inspection 12, 28 input module 617 instantaneity 517 instant messaging (IM) 488, 778, 967 Institute of Electrical and Electronics Engineers (IEEE) 252 intangible 327 integrated marketing communications (IMC) 885 service digital network (ISDN) 142, 504, 533 integration manager 645 integrity 581 intelligent agent 561, 1000 decision support system (IDSS) 386 LBS 856 sensory system 653 inter -provider relationship 798 -symbol interference (ISI) 940 interaction manager 644 services 789, 795 trajectory 935–936 interactive 398 catalog 345 graphing object (IGO) 322 multimedia system (IMS) 341–344 presentation 1052 TV (iTV) 633 voice response (IVR) 1054
interactivity 928 interface category 498 repository (IR) 162 Interface Definition Language (IDL) 160, 164 international law 393 mobile equipment identifier (IMEI) 692 International Telecommunication Union—Telecommunication Standardization Sector (ITU-T) 594 Internet 78, 298, 489, 507, 553, 612 -based learning 419 -enabled mobile handset 639 inter-ORB protocol (IIOP) 164 packet exchange (IPX) 1023 protocol (IP) 51, 504, 537, 700 -based network 20 wireless communication system 730 -multicast 394 multimedia subsystem (IMS) 138, 724, 801–802 service provider (ISP) 35 technology 416 usage 78–79 interoperability 149, 463 interoperable object reference (IOR) 160 interpersonal communication 291 intuitively-correct 857 invalidation report (IR) 107 strategy (see also cache coherence) 849 inventory 908 inverse replication (IR) 964 involvement 284, 290 IOR (see interoperable object reference) IP (see Internet protocol) address 253, 980 mobility 253 iPod 365, 489 IPv4 253 IPv6 253 IR (see inconsistency ratio) IR (see interface repository) IR (see invalidation report) IR (see inverse replication) ISDN (see integrated services digital network) ISI (see inter-symbol interference) itinerary attack 828 ITU-T (see International Telecommunication Union— Telecommunication Standardization Sector) 594 iTunes 78 iTV (see interactive TV) IVR (see interactive voice response)
12 Index
J J2ME (see Java 2 Platform, Micro Edition) Java media framework (JMF) 915 virtual machine (JVM) 914 Java 2 Platform, Micro Edition (J2ME) 193–194, 269, 309, 367–368, 503, 667, 668, 795 security model in its standard edition (J2SE) 979 JCSP (see communicating sequential processes for Java) JD (see joint detection) jini 77, 882 Joey transaction (JT) 948 joint detection (JD) 940 Joint Video Team (JVT) 594 JPEG 670 JPEG2000 670 JT (see Joey transaction) 948 just-in-time learning 526, 905 JVT (see Joint Video Team) K k
-nearest neighbor (k-NN) 660–665 -NN (see k-nearest neighbor) 660–661 -shortest paths problem 861 searching algorithm 859 Kalman filter 212 kangaroo transaction (KT) 947 KDC (see key distributor center) Keitai 296 KEK (see key encyption key) kernel 342 key attack 828 distributor center (KDC) 832 encryption key (KEK) 227, 232, 832–833, 838 healthcare system input 1010 management 1020 algorithm 232, 838 mapping 961 seed negotiation protocol 826 keyword-based language 370 killer application 856 Kismet 1024 KMS (see knowledge management system) knowledge management system (KMS) 376 readiness 485 representation 380 worker 773 Knowledge Query and Manipulation Language (KQML) 429
KQML (see Knowledge Query and Manipulation Language) KSACI 245 KT (see kangaroo transaction) L L2CAP (see logical link control and adaptation protocol) lab information system (LIS) 1005 LAN (see local area network) laptop 3, 476, 576, 754, 757, 1011 large mobile host (LMH) 950 lateration 394 lazy acquisition pattern 746 LBM (see location-based multicast) LBS (see location-based service) LCR (see low chip rate) LDD (see location-dependent data) LDIS (see location-dependent information services) LDQ (see location-dependent query) leadership readiness 485 LEAP 244 learning 318, 900 management system (LMS) 318, 423 support 900 leasing pattern 746 least frequently used (LFU) 746 recently used (LRU) 135, 740, 746 risk path 857 leaving 835 legal context 26 legality 27 liberty alliance 803 LIGLO (see location-independent global names lookup) limited connectivity 713 ergonomics 713 resources 399, 713 usability 399 link manager protocol (LMP) 272 server 476 Linux 366–368, 895 liquid crystal display (LCD) 533, 536 LIS (see lab information systems) literacy 568 live streaming 778 LKH (see logical key hierarchy) LMCDS (see location-based multimedia content delivery system) LMDS (see local multipoint distribution system) LMH (see large mobile host) LMP (see link manager protocol)
Index 13
LMS (see learning management system) load balancing 53, 925 local area network (LAN) 78, 172, 253, 533, 627 cache hit 155 information services 53 multipoint distribution system (LMDS) 455 spatial information service (LSIS) 790 spread inverse replication (LSIR) 964 replication (LSR) 964 transaction (LT) 948 wireless interface (LWI) 15 Locales Foundation 935 Framework 934 localization 203, 205, 525 location 118, 863 -aware application 863 content provision 856 multicast 395 system 526 -based game service 660 multicast (LBM) 395 multimedia content delivery system (LMCDS) 381, 386 service (LBS) 129, 137, 392, 393, 396, 402, 461, 510, 515, 660, 665, 773– 777, 789, 795, 885, 1042 -dependent cache invalidation 105–107 data data (LDD) 393, 739 information service (LDIS) 743 query (LDQ) 739, 743, 854 processing 394 -independent global names lookup (LIGLO) 85 -tracking application 561 estimation system 864 inference queries 867 of service 983 operating reference model (LORE) 394 server 388–389 logical key hierarchy (LKH) 227, 232, 835–838 structure 228, 833 link control and adaptation protocol (L2CAP) 272 logistics 895, 898 long-term shared key 1021 longevity 522 lookup
pattern 746 server 882 loose coupling 359, 810, 878 LORE (see location operating reference model) low -end consumer device 978 chip rate (LCR) 940 loyalty 906 LRU (see least recently used) LSIR (see local spread inverse replication) 964 LSIS (see location spatial information service) 790 LSR (see local spread replication) 964 LT (see local transaction) 948 LWI (see local wireless interface) M m -advertising (see mobile advertising) -commerce 345–351 device (MCD) 38 -commerce (see mobile commerce) -health 1010 -learning 318 -payment (see mobile payment) 714 m[KM] 521 MAC (see medium access control) MAC (see mobile agent communication) Macromedia Flash 672 MADGIS (see mobile agent-based distributed geographic information system) MAE (see mobile agent environment) magnetic resonance (MR) 533 MAI (see motion adaptive indexing) MAI (see multiple access interference) main profile (MP) 592 malevolent hackin 534 malicious software 581 MAM (see mobile agent manager) man -in-the-middle attack 1025, 1028, 1031 -machine interface 403–412 MAN (see mobile agent naming) management message modified 196 manager 644–645 MANET (see mobile ad-hoc network) maneuverability 41 manufacturing logistics 896 mapping 723 candidate 721 MAR (see mobile AR) MARI (see multi-attribute resource intermediary) marketing 99–100 literature 934 MAS (see mobile agent security)
14 Index
massively multi-player online game (MMOG) 929 MAT (see mobile agent transportation) maturity 41 maximum transmission unit (MTU) 805 MBR (see minimum bounding rectangle) MC (see mobile client) MCOP (see multi-constrained optimal-path problem) MCU (see multipoint control unit) mean square error (MSE) 758 MEC (see mobile electronic commerce) media access controller (MAC) 201 gateway (MGW) 874 gateway control function (MGCF) 874 object 390 player 757 resource function controller (MRFC) 874 processor (MRFP) 874 processing 386 support module 981 synchronization 1052 mediator 312 medical science 1004 medical sensor 504 medium access control (MAC) 150, 272, 276, 329, 921 member join 228, 834 leaving 834 memory -limited mobile device 260–264 card 839 mental context 116 mesh base station (mesh BS) 921 BS (see mesh base station) 921 network 981 message authentication code (MAC) 840, 1021 latency 192 Message F (Free) 635, 638 meta-repository (MR) 438 metrics 12, 28 metropolitan area network (MAN) 839 MGCF (see media gateway control function) MGW (see media gateway) 874 MH (see mobile host) micro -contractor 519 -payment 33, 460 microcontroller (MCU) 1011 microelectronics 877 Microsoft mobile Internet toolkit (MMIT) 912 microwave multipoint distribution system (MMDS) 455
MIDI (see musical instrument digital interface) middleware pattern 745 MIDlets 979 MIDP (see mobile information device profile) migration 119, 988, 998 design pattern 987 migration MIMO (see multiple-input multiple-output) MIN (see mobile identification number) MIND (see mobile IP-based network development) minimum bounding rectangle (MBR) 740 MIS (see mobile information system) MLBG (see mobile location-based game) mLMS (see mobile learning management system) MMA (see mobile mote agent) MMD (see multimedia domain) MMDBS (see mobile multi-database system) MMDS (see microwave multipoint distribution system) MN (see mobile node) MNC (see multinational corporation) MNO (see mobile network operator) MobiAgent 245 MOBIDIS1 1046 Mobile Entertainment Forum (MEF) 487 Magnifier 570 mobile access 25 accessibility 14 ad-hoc network (MANET) 395, 399, 402, 424, 427, 492, 734, 749–753, 803, 925, 950 advertising (m-advertising) 398, 402 (see also wireless advertising, mobile marketing) agent (MA) 218–219, 439, 580, 717, 723, 994, 1001 -based distributed geographic information system (MADGIS) 213, 219 migration design pattern 987 communication (MAC) 216 environment (MAE) 214, 219 manager (MAM) 216 naming (MAN) 216 security (MAS) 216 transportation (MAT) 216 and wireless terminals (MWTs) 711 application 9, 375, 744 AR (MAR) 207 authorization 15 banner advertising 638 billing 547, 552 broadband 870 wireless access (Mobile-Fi) 733 business (m-business) 442 calendar 442 channel 577, 580
Index 15
chat 251 client (MC) 190, 194, 745–746 commerce (m-commerce) 15, 32–37, 38–42, 96, 283, 290, 311, 429, 435, 455, 461–465, 472, 480, 547, 552, 933 communication 393, 589, 700 computing 61, 78, 93, 169, 266, 291, 854, 986 and commerce 466–471 environment 849 conferencing support 591 content provider 552 controls 912–920 cooperative caching 750 convergence 138 credibility 26 data communication 399 terminal (MDT) 907 device 71, 77, 1043 e-mail 442 electronic commerce (MEC) 414, 974 entertainment 487–491, 669–674 environment 107, 159, 850, 1021 game 185–189, 928–932 development 497 industry 930 gaming 497 GIS 134 health (m-health) 504 healthcare business model 1010 delivery model 1010 delivery system (MHDS) 505 host (MH) 102, 749 hunter 510, 515 ICT 486 identification number (MIN) 691–692 information device profile (MIDP) 305, 420, 793, 980 system (MIS) 190, 194 Internet 213, 296–301, 460 access 785 IP 1042 -based network development (MIND) 733 knowledge management (mKM) 520–524 learning (m-learning) 318, 327, 423, 525–527, 528, 899 management system (mLMS) 420 location-based game (MLBG) 510, 515 lotteries 553 mail advertising 636 marketing 96–101 (see also mobile advertising, wireless advertising)
marketplace (m-marketplace) 50 module 1012 mote agent (MMA) 328, 330, 333 multi-database system (MMDBS) 949 multicast 541–546, 592 network node (MN) 800 operator (MNO) 547, 714 technology 497 virtual operator (MVNO) 876 node (MN) 53, 255–257 object table (MOT) 663 payment (m-payment) 548, 552, 714, 716 service 706–710 phone 557, 569–575, 644, 999–1000, 1011 gambling 553–556 human-interface system 617 texting 568 positioning 515 process 580 query processing 855 resource 14 robotic system 403–412 service 716 provider 143–148 station (MS) 150 support station (MSS) 190, 194, 749 system 102 system 369 telephone 318 television (mobileTV) 611–615 terminal 399, 456, 789, 795 ticketing (m-ticketing) 714 transaction 948 manager (MTM) 948 user 393 virtual community (MVC) 632, 634 network operator (MVNO) 548 Web 9 engineering 14, 30 MOBIlearn 528, 532 Mobile Speak 570 mobility 73, 243, 318, 461, 525, 576, 632, 644– 650, 769, 906 management 426, 590, 731 services 789, 795 mobilization 482 Mobitip 311 modern digital communications technology 562 modification score 336 modular system 196 modulation 276
16 Index
monitoring 660 and discovery service (MDS) 437 Morse code 352, 356, 616, 620 MOSPF (see multicast open shortest path first) MOT (see mobile object table) motion adaptive indexing (MAI) 661 sensitive bounding boxes (MSB) 661 Moving Picture Experts Group (MPEG) 594 moving object 665 MP (see main profile) MP3 player 4, 365 MPEG (see Moving Picture Experts Group) MPML-FLASH (see Multimodal Presentation Markup Language with Character Agent Control in Flash Medium) MPML-VR (see Multimodal Presentation Markup Language for Virtual Reality) MPML (see Multimodal Presentation Markup Language) MRFC (see media resource function conroller) MRFP (see media resource function processor) MR (see magnetic resonance) MSB (see motion sensitive bounding boxes) MSS (see mobile support station) MTM (see mobile transaction manager) MUD (see multi-user joint detection) multi -agent system (MAS) 383, 386 -attribute resource intermediary (MARI) 335 -constrained optimal-path problem (MCOP) 981 -database system 947 -hop routing 1035 -layer -based taxonomy 929 mobility 966–973 -objective optimization 683 -user joint detection (MUD) 940, 943–946 -media messaging (MMS) 116 multicast 227, 232, 393–397, 397, 541–546, 838 listener 542 open shortest path first (MOSPF) 395 routing protocol 397 technology 393 multifunctional handheld device 497 multihop communication 427 multimedia 392, 757, 870, 986 -capable 398 content database 388 server 390 data 24 domain (MMD) 873, 876 info presentation 389 message service (MMS) 252, 669, 756–757, 840
multimodal service 1050 user interface 644–650 presentation 389 multimediality 675 Multimodal Presentation Markup Language 1052, 1056 for Virtual Reality (MPML-VR) 1052 with Character Agent Control in Flash Medium (MPML-FLASH) 1052 multimodal 604 access 602 multimodality 675–681 multinational companies database 886 corporation 886 corporation (MNC) 934 multiple -input multiple-output (MIMO) 701 access interference (MAI) 940, 946 multipoint control unit (MCU) 590 Mummy 520 musical instrument digital interface (MIDI) 15, 19 music download 488 mutuality 935–936 MVC (see mobile virtual community) MVNO (see mobile network virtual operator) MWT (see mobile and wireless terminal) MySpace 960 N naming pattern 746 NAL (see network abstraction layer) narratives 928 narrowband (NB) 458 national law 393 navigation 861 algorithm 857 context 857 preferences 858 service 859 system 125 NB (see narrowband) NCFG (see network configuration) NCT (see network connectivity table) near -far problem 168 -real-time balance management 874 Near Field Communication (NFC) 706 negative update 665 negotiation of bid 983 neighbor discovery and management 203, 206 link 87
Index 17
unreachability detection (NUD) 259 neighboring cell 743 NENT (see network entry) network 392, 981, 986 -centric design viewpoint 873 model 874 -level multicast (NLM) 394 abstraction layer (NAL) 592–595 configuration message (NCFG) 927 connectivity table (NCT) 662 entry message (NENT) 927 frame 927 infrastructure 6, 387 security requirement 191 topology 986 survivability requirement 191 networkability 186 networking 207 neutral data format 75 new learning 319, 327 next generation 472 network (NGN) 142, 796, 803 NGN (see next generation network) Ninja secure service discovery service (SSDS) 436 NLM (see network-level multicast) NLOS (see non-line-of-sight) node 330, 580 mobility 172 noise 348 threshold 348–349 nomadic user 576, 589, 591, 595 non -deterministic polynomial time (NP) 733 -formal learning 532 -line-of-sight (NLOS) 921 -linear navigation 599 -quantifiable attributes 334 -renewable resource 763 -vital 951 notification service (NS) 711, 716 NS (see notification service) NUD (see neighbor unreachability detection) O OASIS (see Organization for Advancement of Structured Information Standards) Object Management Group (OMG) 1002 object 604 adapter 164 name service (ONS) 178, 183 -oriented Petri net 994 request broker (ORB) 160, 164
oblette 532 OBU (see on board unit) occurrence graph 995 OFDM (see orthogonal frequency division multiplexing) off -line customer 35 -the-shelf (OTS) 61 offset-quadrature phase-shift keying (OQPSK) 273 old learning 319 on -demand board unit (OBU) 894, 898 environment 159 learning 525 online community 632, 634 ontology 723, 862 session 727 one-time two-factor authentication 533, 540 online betting 553 gambling 553 merchant 311 shopping 283, 290 ONS (see object name service) ontology 50, 339, 380 universally unique identifier (OUUID) 44 open coupling 358 system interconnect (OSI) 165 system model 1030 Open Mobile Alliance (OMA) 463, 613 operational system (OS) 344 operating system (OS) 765, 895, 898, 998 opportunistic communication 165 exploration 345 scheduling (OS) 165–171 opt -in 643 -out 643 optical character recognition (OCR) 667–668 optimization 791 optimized lifetime 725 options model 39 OQPSK (see offset-quadrature phase-shift keying) ORB (see object request broker) organizational security 1011 Organization for the Advancement of Structured Information Standards (OASIS) 876 orthogonal frequency division multiplexing (OFDM) 682, 701
18 Index
OS (see operating system) OS (see operational system) ontology-based context model 718 ontology repository 860 OTS (see off-the-shelf) OTTFA (see one-time two-factor authentication) OUUID (see ontology universally unique identifier) overlay network 738, 965 OWL (see Web Ontology Language) P P-CSCF discovery 779 P-PAN (see private personal area network) P2P (see peer-to-peer) packet data protocol (PDP) 813 loss rate 22–24 radio network 427 routing 257 transmission 198 Palm OS 304, 309 Cobalt 304 Garnet 304 palm pad computer 476 palmtop 757 PAN (see personal area network) paradigm shift 860 parameterized query 602 Parlay X 876 partial global indexing (PGI) 742–743 serialization graph (PGSG) 949 participant 899 passive attack 1024 credibility 25 distributed indexing 492 password 841 path -searching algorithm 857 -selection rules 858 optimization 986 pattern 744–748 extraction 673 recognition 394 PayTV 611 PC-based online survey 640 PC (see personal computer) PC (see pervasive computing) 956 PCI (see perceived characteristics of innovating) PDA (see personal digital assistant) PDF (see policy decision function) 874
PDU (see protocol description unit) peak signal to noise ratio (PSNR) 758 pedagogy 532 peer-to-peer (P2P) 2, 84, 89, 159, 395, 397 architecture 108 communication 165 computing 233, 492, 693, 734 financial transaction 108 networking 357 pen -based interface 754 computing 757 penetration rates 398 PEP (see performance enhancing proxy) perceived knowledge 25 quality of service (PQoS) 758 switching cost 145 perceptual navigation rules 858 performance 27 comparison 239 discussion 230 enhancing proxy (PEP) 785, 788 permission-based marketing 934 personal area network 507, 1011 area network (PAN) 1, 7, 252 computer (PC) 79, 141 context 711 device 7 device assistant (PDA) 265 digital assistant (PDA) 10, 38, 79, 138, 154, 172, 2 50, 260, 317, 368, 375, 443, 461, 505, 511, 5 25, 581, 711, 757, 907, 974, 980, 996, 1011, 1015, 1050 phone 894 information management (PIM) 249–252, 757 network (PN) 1, 8 trusted device (PTD) 581 virtual environment (PVLE) 530 personality 243 personalization 31, 380, 398, 463, 525 pervasive computing (PC) 57, 61, 71, 276, 320, 621, 863, 869, 877–878, 884, 956 location-aware computing environments (PLACE) 661 pest activity 763 control 763 management 763 Petri Net 987 PGI (see partial global indexing) 743 PGSG (see partial global serialization graph) 949 physical
Index 19
context 116 navigation rules 858 network 738 space 774 Physical Markup Language (PML) 183, 819 PIA-SM (see protocol independent anycast - sparse mode) piconet 50, 804, 825 picture download 489 pilot aliasing 692 attack 689 study 641 PIM (see personal information management) phone 894 PIM-SM (see protocol independent multicast-sparse mode) PIN 841 -code 477 pixel-based method 890 PKI SIM card 840 PLACE (see pervasive location-aware computing environments) planar structure 822, 825 planned disconnection mode 951 platform-based taxonomy 929 play 604 PML (see Physical Markup Language) PMP (see existing point-to-point multipoint) PN (see personal network) POA (see portable object adapter) PoC session identifier 727 Pocket PC 304 PoI (see points of interest) point -tomultipoint (PMP) 921 point IP traffic 576 of sale (POS) 706, 907 points of interest (PoI) 390 policy decision function (PDF) 874 porosity 517 portability 27, 186 portable /mobile device 626 computing 195 media center 305 network graphics (PNG) 670 object adapter (POA) 162 sensor 865 port forwarding 540 (see also application tunneling) position-aware mobile device 561 positioning 869 method 386
positive update 665 power control 731 management 1035 pre -committed transaction 696 -generation (Pre-G) 187 -processing of raw data 865 -roaming notification 432 -serialization transaction management model 949 premium fee charging 549, 552 presence 1042 presentation layer 262 presenting 882 presumed credibility 25 preview clip 604 price tolerance 145–146 primary notation 11 printer 4 privacy 75, 117, 399, 462–463, 477, 773–777, 1002 (see also security) concerns 713 private personal area network (P-PAN) 7 space 557 pro-motion model 949 proactivity 243 process readiness 484 processing kernel 1013 procurement activity 896 producer 9 product brokering 340 catalog 345 manager 393 productivity 443, 473, 839, 906 profile information 251 profiling 99 programming languages 915 propitient multi-agent system 957 protocol 435, 831 data unit (PDU) 584 description unit (PDU) 352 independent anycast - sparse mode (PIA-SM) 52 multicast-sparse mode (PIM-SM) 395 of use 59 proximity 117, 394 proxy 77, 235, 626 pseudo-transaction 236 PSTN (see public switched telephone network) public 702 -switched telephone network (PSTN) 139, 142
20 Index
key infrastructure (PKI) 581–588 interface SIM (PKI SIM) 840 regulation 886 space 557 publisher 503 -subscriber pattern 745 pull 400, 402, 879 -based environment 107, 749 service provision 880 -type advertising 635 model 98 pure P2P 89 push 400, 402, 879 -based environment 107, 749 service provision 879 -to-talk over cellular (PoC) 724 -type advertising 635 message 716 messaging service 643, 888 model 98 PVLE (see personal virtual environment) Q QFD (see quality function deployment) QI (see query interface) QoS (see quality of service) quadriplegic 403 quality 14, 31, 758–762, 1004 attributes 11, 27 function deployment (QFD) 12, 28, 376 of service (QoS) 1, 140, 149–150, 195, 201, 207, 37 9, 425, 621, 624, 627, 701, 729, 746, 778, 80 3, 804–809, 906, 923, 982, 1042 routing 986 score 860 quantifiable attribute 334 queries generated summary 604 query decomposer and coordinator 790 interface (QI) 214 language 369 lifetime 738 processing 851 queue manager 197 queuing 739 R R&D (see research and development) RA (see router advertisement) RADAR 774 radio
access 701 network (RAN) 361 technology (RAT) 149, 810 frequency (RF) 942 identification (RFID) 178, 183, 507, 819, 856, 864 network design 946 resource management (RRM) 425, 729 radiology information systems (RIS) 1005 RAID (see redundant array of inexpensive disks) RAM (see random access memory) random access memory (RAM) 163 network 983 range -incline finder 653 -inclination tracer 406 query 665 rapid development 102 prototyping 291 raster data processing engine 389 RDF (see resource descriptiopn framework) reachability 398 reactivity 243 read one write all (ROWA) 234 real-time protocol (RTP) 726 requirement 589 strategy (RTS) 186 tracing 791 traffic 626 management 622 transfer protocol (RTP) 263 transport protocol (RTP) 595 Real Audio 265 reasoning 862 received signal strength (RSS) 863 recognition algorithm 17 recognizer parameters 18 recommender system 345 recommending agent 109 recycling logistics 896 redefinition 483 redirection 1025 reduced function device (RFD) 273 redundant array of inexpensive (or identical) disks (RAID) 540 array of inexpensive disks (RAID) 533 recoding 11 reflection 60, 164 rehabilitation robotics 653 rekey 834 relative advantage 413–414, 418 relaxed check-out mode 951
Index 21
reliability 27 remote authentication dial-in user service (RADIUS) 803 control object 882 subscription 543 replication 965 reputed credibility 25 research and development (R&D) 700 reshapement 483 resource description framework (RDF) 377 management scheme 812 readiness 485 response rate 640 results demonstrability 414, 418 retransmission timeout (RTO) 785, 788 returning mobile agent 579 RF (see radio frequency) RFD (see reduced function device) RFID (see radio frequency identification) ring structure 822, 825 RIS (see radiology information systems) risk 109, 857 set 111 riskiness value 110 roaming permit 431 robust security network 1032 robustness 27 rogue wireless gateway 1028 role playing game 186, 928 round-trip time (RTT) 785, 788 route analysis 791 optimization 253–259 router advertisement (RA) 256–257 discovery 255, 257 solicitation 255, 257 routing 394, 807, 965 algorithm 859 header 254 identification 807 mechanism 732 metrics 925 overhead 925 ROWA (see read one write all) RRM (see radio resource management) RSS (see received signal strength) RTO (see retransmission timeout) RTP (see real-time protocol) RTP (see real-time transfer protocol) RTP (see real-time transport protocol) RTT (see round-trip time) running task 1047
S S-MAC (see sensor MAC) SA (see signal attenuation) SAFE (see secure roaming agent for e-commerce) sales logistics 896 SALT (see speech application language tags) 1053 SAML (see Security Assertion Markup Language) sampling frequency 19 satellite 508 -based augmentation systems (SBAS) 771 communication 907 scalability 925 scalable vector graphics (SVG) 376, 671, 1050, 1056 video coding (SVC) 591–592, 595 scatternet 276, 825 Scavenger Hunt (SH) 292 scene analysis 394 schedule frame 927 scheduler 197 SCO (see synchronous connection oriented) SCP (see servers/client paradigm) SDAP (see service discovery application profile) SDDB (see service discovery database) SDK (see software development kit) SDK (see software development toolkits) SDMA (space division multiple access) 944 SDP (see service delivery platform) 870 SDP (see service discovery protocol) SDP (see session description protocol) SDP (see sophisticated service discovery protocol) secondary notation 11 second generation (2G) 456 mobile telephony 503 wireless system 96 secure authorization 800 multimedia service access 799 roaming agent for e-commerce (SAFE) 429 service discovery 799 shell (SSH) 842 socket layer (SSL) 383 security 6, 75, 109, 192, 313, 426, 435, 462, 477, 505, 624, 633, 831, 839–848, 968, 1002, 1011, 1022–1027 (see also privacy) analysis 828 policy 1028 Security Assertion Markup Language (SAML) 378, 872, 876 segment 604 segmentation and reassembly (SAR) 805 self-independence 569
22 Index
semantic caching 850, 855 distance 43 interpretation (SI) 1054 knowledge engineering 860 layer 43 location-based services 857 matchmaking 50 service discovery 44 Semantic Web 14, 31, 375, 380, 856, 862 Rule Language (SWRL) 858 semantics 377 semiotic level 11, 27 tool 327 semiotics 14, 31, 380 sending window 788 sensing layer 865 sensor 125, 1035 data 126 fusion 126 information fusion 863 layer 864 MAC (S-MAC) 328 model 1034 network 276 sensory system 651, 653 serializability 233 sequence diagram 716 serial port profile (SPP) 344, 1014 Series 60 249 server 219 /client paradigm (SCP) 954 -side handheld computing 302 programming 309 connection 251 site 790 service 774, 884, 986 -oriented approach (SOA) 529 architecture (SOA) 71, 874, 878 computing (SOC) 71, 626, 877–878, 884 paradigm 621, 883 pervasive computing 883 abstraction layer pattern 746 access control 800 advertisement 884 aggregator 7 client 878, 884 configurator 747 creation environment 873 delivery platform (SDP) 870 discovery 53, 89
application profile (SDAP) 344 database (SDDB) 882 protocol (SDP) 50, 272 level agreement (SLA) 383 location protocol (SLP) 43 manager (SM) 799, 803 plane 873 provider (SP) 714, 878, 884, 982 redundancy 53 registry 878, 884 relationship management (SRM) 873 services composer 86 deployer 86 discovery engine 86 session description protocol (SDP) 590 initialization key (SIK) 1021 protocol 590 initiation protocol (SIP) 140–142, 595, 803, 873, 966 key 1021 mobility 967 SGML (see Standard Generalized Markup Language) shared key authentication 1030 shooting Game 929 short message service (SMS) 32, 194, 252, 456, 526, 568 , 643, 666, 668, 726, 755, 757, 840, 885, 933 sound file 670 shot boundary detection 889–893, 890 SI (see semantic interpretation) sign-off/check-out mode 234 signal -to-noise ratio (SNR) 16, 595 attenuation (SA) 511 signature attack 828 scanner 651 SIK (see session initialization key) SIM card 585 simple message service (SMS) 352, 356, 616, 620 network management protocol (SNMP) 1026 object access protocol (SOAP) 376, 795 simplest path algorithm 857 simulation -based learning 901 game 928 simultaneous PoC session 727 single-user game 929 SIP (see session initiation protocol) site -specific service 580
Index 23
design 285, 290 transaction manager (STM) 949 SLA (see service level agreement) SLP (see service location protocol) SM (see service manager) SMA (see stationary monitoring agent) small mobile host (SMH) screen devices 123 smart antenna 942, 946 card 472, 475, 480, 708 cellular phone 308 phone 327, 344, 757, 894, 898, 905, 1014 space 71, 124 smartphone 304 SME (see station management entity) transaction manager (STM) 949 SMH (see small mobile host) SMIL (see Synchronized Multimedia Integration Language) SMS (see short message service) SMS (see simple message service) snap-on/wireless keyboard 571 sniffing 1022 SNR (see signal-to-noise ratio) SOA (see service-oriented approach) SOA (see service-oriented architecture) SOAP (see simple object access protocol) SOC (see service-oriented computing) social context 116 research 557 system 632 software agent 340 component 61 development toolkit (SDK) 12, 394 solidring structure 822, 825 sophisticated service discovery protocols (SDPs) 43 sound 669–674 source -specific multicast (SSM) 541, 592, 595 system 947 SP (see service provider) space division multiple access (SDMA) 944 SPAM 888 spamming 534 spatial diversity 165 models and ontologies 857 services 792 speaker verification (SV) 677 spectrum management 730
speech 645 application language tag (SALT) 647, 1053, 1056 recognition 675 grammar specification (SRGS) 1053–1054 Speech Synthesis Markup Language (SSML) 1056 SpeechPAK TALKS 570 SPIM 888 split mode 948 sponsoring 552 spontaneous service emergence paradigm 957 sports game 929 SPP (see serial port profile) spyware 558, 561 SQL (see Structural Query Language) SQL (see Structured Query Language) SRGS (see speech recognition grammar specification) SRM (see service relationship management) SSL (see secure socket layer) SSM (see source specific multicast) standard transmission protocol 160 type phone bill 552 Standard Generalized Markup Language (SGML) 142 stateful address autoconfiguration 258 stateless address autoconfiguration 258 static access frequency (SAF) 751 agent 218 stationary monitoring agent (SMA) 328, 330, 333 station management entity (SME) 196 STM (see site transaction manager) storage 536 cost 231 efficiency 231, 836 inefficiency 227 protection 840 stream control transmission protocol 813 streaming 77, 266 Structured Query Language (SQL) 370, 533 structured non-super peer 736 super peer 736 student support 901 subscriber identity module (SIM) 840 card 839 summarization technique 121 super frame 927 peers 734 superior healthcare delivery 1010 supervised agent transport 430 sustainable agriculture 763 development 763
24 Index
SVC (see scalable video coding) SVG (see scalable vector graphics) SVG Basic (SVGB) 1050 SVGB (see SVG Basic) SVG Tiny (SVGT) 1050 SVGT (see SVG Tiny) switching cost 145 SWRL (see Semantic Web Rule Language) Symbian 342, 503 operating system (OS) 249, 304, 308, 895 symbolic model 857 synchronization 203, 206, 757 Synchronized Multimedia Integration Language (SMIL) 376, 1056 synchronous connection oriented (SCO) 344 syntactic translation 123 system architecture 1012 capacity 946 design 352, 616 module 342 T t-test 897 tablet PC 755, 757 tacitness 523 tagged image file format (TIFF) 670 talk burst control 727 Talks 570 TAM (see technology acceptance model) tangible score 336 target respondents 640 task-orientated agent 578 Tatoes authoring 321 taxonomy 723 TCP-friendly rate control (TFRC) 20 TD-SCDMA (TD-SCDMA) 946 TDD (see time division duplex) TDMA (see time division multiple access) TDOA (see time difference of arrival) teacher support 901 team (or guild) game 929 technical heterogeneity 399 technobabble 568 technology -conditioned approach to language change and use (TeLCU) 562, 568 acceptance model (TAM) 38, 296–297, 894 readiness 484 TeLCU (see technology-conditioned approach to language change and use) tele-worker 634 telemedicine 504 telephone interview 887
teleradiology 540 temporal logic 995 temporary disconnection 75 terminal equipment 7 testing 12, 28 texting 757 text messaging 899 text to speech (TTS) 645–646, 1053 TFRC (see TCP-friendly rate control) TFT (see thin film transistor) TFTP (see trivial file transport protocol) thematic spatial information services interfaces (TSISI) 790 thin film transistor (TFT) 536 third generation (3G) 475, 503, 533, 611, 933 game 187 mobile network (3G) 733 system 682 network 778 wireless system 96 third generation (3G) three -dimensional (3D) animation 672 -tier Web-based architecture 1010 thumb board text interface 755 tight coupling 359, 810 time and distance sensitive (TDS) 751 division duplex (TDD) 272, 940, 946 -synchronous code division multi-access (TD-SCDMA) 946 difference of arrival (TDOA) 511 division multiple access (TDMA) 329, 456, 535 of arrival (TOA) 511 timestamp 740 TOA (see time of arrival) Toku number 636 Tokusuru Information Board 638 Menu 638 tools 12, 28 tourism 365, 392 tourist guide 713 services 660 TPI (see transport information items) tracking 212, 660 system 207, 907 traditional shortest path algorithm 857 traffic analysis 1024 control 660, 730 transcoding 628–629, 984
Index 25
transmission error 192 transparency 27 transport information item (TPI) 786 layer mobility 968 management 660 traveling mobile agent 579 tree -based routing 395 structure 821, 825 (see also hierarchical structure) trialability 413–414, 418 triangle routing 259 triangulation 394 trilateration 770 triple play 803 trivial file transport protocol (TFTP) 263 trusted checkout 312 trusting agent 110 trustworthiness 25 TSISI (see thematic spatial information services interfaces) TTS (see text-to-speech) tunnel entry point 255 exit point 255 tunneled IPv6 packet 255 tunneling 255, 543 tuple forming 865 U u-commerce 310–316 ubiquitous computing (UC) 62, 561, 589, 775, 954 tracking system 208 ubiquity 525 UC (see ubiquitous computing) UCA (see uniform circle array) UDDI (see universal description, discovery and integration) UE (see user equipment) UI (see user interface) UID (see user interface design) ultra wide band (UWB) 427 ULA (see uniform linear array) ultrasonic signal 864 UMA (see unlicensed mobile access) UMTS (see universal mobile telecommunication system) unaccounted attribute 334 Unicode Standard 376 Unified Modeling Language (UML) 716 unified identity 140 uniform circle array (UCA) 942
linear array (ULA) 942 resource identifier (URI) 376 uniframe 212 resource discovery system (URDS) 212 universal description, discovery and integration (UDDI) 89 mobile telecommunication (UMTS) 535 mobile telecommunications system (UMTS) 20, 194, 511 plug and play (UPnP) 882 serial bus (USB) 368 SIM (USIM) 840 universally unique identifier (UUID) 44 unlicensed mobile access (UMA) 142, 360–361, 701 UNO (see user navigation ontology) unsolicited message 399 unstructured super peer 735 unsupervised agent transport 432 update broadcast 751 updating distributed key 1019 session keys 1018 uplink synchronization 944 UPnP (see universal plug and play) URDS (see uniframe resource discovery system) URI (see uniform resource identifier) usability 27, 118 USB (see universal serial bus port) user -centered operability 140 -oriented rekeying 834 attention model 599, 604 context 26, 856 control 800 datagram protocol (UDP) 583 equipment equipment (UE) 63, 727, 782, 942 identification 775 interface 303 interface (UI) 74, 342–344 design (UID) 319, 644 mobility 967 navigation ontology (UNO) 858 preference 602 profile 31, 126, 380, 626, 629–630 terminal (UT) 810 utility function 44, 168 UUID (see universally unique identifier) UWB (see ultra wide band) V VAG (see virtual anycast group) validation authority (VA) 582 value 1004
26 Index
-added mobile service 497 chain model 497 values and goals readiness 485 VANET (see vehicular ad hoc network) VC (see virtual community) VCEG (see Video Coding Experts Group) VCoIP (see videoconferencing over IP) vector space model (VSM) 87 vehicular ad hoc network (VANET) 424, 427 verification 430 Verisign 35 vertical handoff 151 handover 967 management 812 mobile business 442 very large scale integration (VLSI) 427 video -on-demand 890 coding 590 sequence 889–893 stream 124 transcoding 627–631 Video Coding Experts Group (VCEG) 595 videoconference 758 videoconferencing 589 over IP (VCoIP) 589 video telephony 589 viral marketing 98, 402 virtual anycast group (VAG) 54 cluster 333 community (VC) 632, 634 component pattern 746 machine (VM) 996, 998 monitor (VMM) 996, 998 memo 714 private network (VPN) 533, 842, 1032 proxy pattern 746 reality (VR) 672 Virtual Reality Modeling Language (VRML) 672, 1052 virtualization 998 virus 1024 attack 534 visibility 413–414, 418 visual query language 370 vitals 951 VLSI (see very large scale integration) VM (see virtual machine) VMM (see virtual machine monitor) voice -activated communication 571 -over Internet protocol (VoIP) 138, 142, 269, 700, 735, 898
recognition 999–1003 Voice Extensible Markup Language (VoiceXML) 1050 VoiceXML (see Voice Extensible Markup Language) 1050, 1056 VoIP (see voice-over Internet protocol) VPN (see virtual private network) VRML (see Virtual Reality Modeling Language) VSM (see vector space model) W W3C (see World Wide Web Consortium) W3C-MMI (see W3C--MultiModal Interaction) W3C-MultiModal Interaction (W3C-MMI) WAN (see wide area network) WAP (see wireless application protocol) WAP identity module (WIM) 585 war-driving 1023 water-filling principle 169 WAVE file 669 WB (see wideband) WCDMA (see Wideband Code Division Multiple Access) Web feature service (WFS) 133, 137 GIS 219 map service (WMS) 133, 137 Ontology Language (OWL) 377 page 877 service 77, 84, 89,795 -based LBS (WS-LBS) 789 Services Description Language (WSDL) 89, 795 system 1050 technology 536 Webtop client 80 WFS (see Web feature service) Wi-Fi (see wireless fidelity) access point 701 wide area 207 network (WAN) 253, 533 wideband (WB) 458 code division multiple access (WCDMA) 534 widgets 81 WiMAX (see worldwide interoperability for microwave access) WiMobile 1042 WinCE (see Windows Consumer Electronics or Windows CE) Windows -based smartphone 305 Consumer Electronics or Windows CE (WinCE) 793, 795 Mobile 304, 309 wired equivalent privacy 1030 Internet 298, 300
Index 27
Wireless Application Protocol 2.0 (WAP 2.0) 423 Markup Language (WML) 191, 194, 376, 456, 476, 583 wireless 283, 285, 327, 1000 access network 391 advertising (see also mobile advertising, mobile marketing) application protocol (WAP) 190, 194, 252, 376, 476, 480, 503, 913 communication 744, 947 connectivity 757 communication network 195 datagram protocol (WDP) 584 data network 165–171 device 62 fidelity (Wi-Fi) 427, 507, 733 headset 4 healthcare 1010 hotspots 207 identity module (WIM) 480 Internet 298, 472 connectivity 576 local area network (WLAN) 702, 733, 839, 856, 906 local area networking (WLAN) 62, 139, 394, 455, 50 5, 507, 511 mesh network (WMN) 921 multimedia application 195 network 62, 456, 832, 906, 1022, 1029 architecture 1038 personal area network (WPAN) 272, 276 public key infrastructure (WPKI) 585 routing algorithm 925 scheduling mechanism 924 security 1028–1033 sensor 4 network (WSN) 202, 206, 328, 424, 427, 921, 1034–1037 session protocol (WSP) 190 technologies 253 transaction protocol (WTP) 583–584 transmission 627 transport layer security (WTLS) 584 user behavior 476 Web service 476, 480 WISDOM 573 WLAN (see wireless local area network) WML (see Wireless Markup Language) WMLSCrypt 585 WMN (see wireless mesh network) WMS (see Web map service) word of mouth (WOM) 300 workflow 1045 management 1043–1049
workload 626 World Wide Web Consortium (W3C) 379, 463, 671, 675, 1050 worldwide interoperability for microwave access (WiMAX) 733 WPAN (see wireless personal area network) WPP (see WAP peer protocol) WS-LBS (see Web service-based LBS) WSDL (see Web Services Description Language) WSN (see wireless sensor network) X XDMC (see XML document management client) XHTML (see eXtensible HyperText Markup Language) -MP (see XHTML Mobile Profile) Basic 376 Mobile Profile (XHTML-MP) 423 XLink (see XML Linking Language) XML (see eXtensible Markup Language) Linking Language (XLink) 377 XP (see extended profile) XSLT (see eXtensible Stylesheet Language Transformations) Xtensible Stylesheet Language Transformations (XSLT) 377 Z ZigBee 273, 428, 507, 805 Zire 489 Zone Cooperative (ZC) cache 173–174