Peer-to-Peer Computing for Mobile Networks. Information Discovery and Dissemination

Peer-to-Peer Computing for Mobile Networks Information Discovery and Dissemination

Peer-to-Peer Computing for Mobile Networks Information Discovery and Dissemination by Maria Papadopouli Department of Computer Science University of Crete & Institute of Computer Science Foundation for Research and Technology-Hellas (FORTH) Crete, Greece Henning Schulzrinne Department of Computer Science Columbia University NY, USA 1 3

Authors: Maria Papadopouli Department of Computer Science University of Crete Knossos Ave. P.O. Box 2208 Heraklion, GR-714 09 Crete, Greece & Institute of Computer Science Foundation for Research and Technology-Hellas (FORTH) N. Plastira 100, Vassilika Vouton, Heraklion, GR-700 13, Crete, Greece mgp@ics.forth.gr Henning Schulzrinne Department of Computer Science Columbia University 1214 Amsterdam Avenue New York, NY 10027, USA hgs@cs.columbia.edu Library of Congress Control Number: 2008934312 ISBN-13: 978-0-387-24427-3 e-isbn-13: 978-0-387-68504-5 DOI: 10.1007/b110083 2009 Springer Science+Business Media, LLC. All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper. springer.com

This work is a metaphorical bridge for me, bringing together research results obtained, while I was in the following academic institutions: Computer Science Department of Columbia University, as a Ph.D. student (1996 2002) Department of Computer Science of the University of North Carolina at Chapel Hill (UNC), as an assistant professor (2002 2004) Department of Computer Science of the University of Crete and the Institute of Computer Science of the Foundation for Research and Technology- Hellas (FORTH-ICS), as an assistant professor (2004 ) Parts of this book are based on research results obtained in joint work with Haipeng Shen, Merkourios Karaliopoulos, Félix Hernández-Campos, George Tzagkarakis, Panagiotis Tsakalides, Elias Raftopoulos, Manolis Ploumidis, Manolis Spanakis, Mark Lindsey, Francisco Chinchilla, Thomas Karagiannis, Charalampos Fretzagias, Niko Kotilainen, Lito Kriara, and Konstantinos Vandikas. Haipeng Shen, Merkourios Karaliopoulos, and Félix Hernández- Campos played an important role in the wireless measurement and modeling research. I am grateful to have the opportunity to closely collaborate with them. Thanks go to Jim Gogan, Todd Lane, Kevin Jeffay, Don Smith and their students for helping to setup the monitoring and data collection system for our network measurement research while at UNC. I would really like to thank all my colleagues at UNC for their help in several different ways. I am also grateful to Diane and Mark Pozefsky for their support while at UNC. FORTH- ICS has provided a state-of-the-art infrastructure to continue my research. Thanks to all my colleagues at the University of Crete and FORTH-ICS for their support that made the transition from US to Greece easier. I would like to acknowledge the support of the director of FORTH-ICS, Constantine Stephanidis and of my colleagues Grigoris Antoniou, Angelos Bilas, Vassilis Christophides, Prasun Dewan, Kevin Jeffay, Manolis Katevenis, Jasleen Kaur, Evangelos Markatos, Christos Nikolaou, George Polyzos, K. K. Ramakrishnan, Montek Singh, Don Smith, Panos Tsakalides, Leandra Vicci, and Gregory Welch. I am also grateful to Anthony Ephremides, Anargyros Papageorgiou, Leandros Tassiulas, and Apostolos Traganitis for their mentoring. Frederick Brooks, Jr. has been a source of inspiration. Elias Raftopoulos and Manolis Ploumidis my first graduate students at the University of Crete and FORTH have been enthusiastically participating in the wireless measurement project. Several other students also contributed in the implementation of CLS, 7DS and applications that use the peer-topeer paradigm: Denis Abramov and Stelios Sidiroglou-Douskos at Columbia University; Mark Lindsey, Daniel Plaisted, Julien Jomier at the University of North Carolina at Chapel Hill; Niko Kotilainen at Jyväskylä University; and Kostantinos Vandikas, Lito Kriara and Sofia Nikitaki at the University of Crete. It was a pleasure working with all of them.

VI I am grateful to my editorial assistant Anthony Griffin for reviewing this manuscript several times and providing useful feedback. Several people reviewed the monograph and provided feedback: Thanasis Mouchtaris, Anargyros Papageorgiou, Haipeng Shen, Leandros Tassiulas, Apostolos Traganitis, Panos Tsakalides, and George Tzagkarakis. I would like to acknowledge Antonis Makrogiannakis and Elias Raftopoulos for helping me with Latex, and Mary-Rose James and Vana Manasiadi for additional editorial suggestions. I also wish to thank my literary agents Susan Lagerstrom-Fife and Sharon Palleschi for their patience. Finally, I would like to gratefully acknowledge the support of several agencies: the Greek General Secretariat for Research and Technology (Regional of Crete Crete-Wise and 05NON-EU-238) the European Commission (MIRG-CT-2005-029186) the Department of Computer Science of the University of North Carolina at Chapel Hill for their generous startup fund and the UNC Junior Faculty Development Award the Institute of Computer Science of the Foundation for Research and Technology-Hellas, for their state-of-the-art infrastructure and generous startup fund IBM, for the IBM Faculty Awards in 2003 and 2004 This monograph also marks an academic journey, that started and ended in Crete. A major visit in this journey: Columbia University and New York City, places that offered intellectual stimulations with such generosity. In 2002, I arrived in a very warm and supportive academic family: the Department of Computer Science at the University of North Carolina at Chapel Hill. Being an assistant professor in this institution my first real job was a particularly rewarding experience. My return as a faculty at the Department of Computer Science in the University of Crete where I had completed my undergraduate studies offered a sense of continuity that has a strong impact on me. The following people made an immense impact on enabling this journey: Manolis Maragkakis, a beloved math teacher. Stelios Orphanoudakis Professor of the Department of Computer Science at the University of Crete a charismatic human being that made a large impact on the development of FORTH-ICS and forthnet S.A. George Papadopoulis and Xacousti Papadopouli-Plevraki, my parents, and Eva Papadopouli, my sister, for always being ready to help, advise, care, and love. This monograph is dedicated to them. Maria Papadopouli

Contents 1 Introduction............................................... 1 1.1 Wireless data communications............................ 1 1.2 Mobile information access................................ 3 1.2.1 Wireless Internet via APs.......................... 5 1.2.2 Infostations....................................... 6 1.2.3 Peer-to-Peer systems............................... 7 1.3 Target mobile computing environment..................... 10 1.3.1 High spatial locality of information and queries....... 11 1.3.2 Heterogeneity in application requirements............ 11 1.3.3 Enhancement of information access.................. 12 1.4 Resource sharing using 7DS............................... 13 1.5 Overview of this monograph.............................. 14 1.5.1 Outline.......................................... 16 2 7DS architecture for information sharing.................. 19 2.1 Overview of 7DS architecture............................. 19 2.1.1 Communication................................... 22 2.1.2 Cache management................................ 22 2.1.3 Power conservation................................ 23 2.2 Preventing denial-of-service attacks........................ 25 2.3 Encouraging cooperation................................. 26 2.3.1 Micropayment mechanisms......................... 27 2.3.2 Reputation mechanisms............................ 30 2.4 Location-sensing using the peer-to-peer paradigm............ 30 2.4.1 Overview of CLS.................................. 32 2.4.2 Particle filter-based framework...................... 36 2.4.3 Performance of CLS and other related systems........ 37 2.5 Applications using information sharing via 7DS............. 39 2.5.1 Web browsing..................................... 41 2.5.2 Notesharing and whiteboard tool.................... 41 2.5.3 Multimedia traveling journal........................ 43

VIII Contents 2.6 Related mobile peer-to-peer computing systems............. 45 2.7 Conclusions............................................. 46 3 Performance analysis of information discovery and dissemination.............................................. 49 3.1 Information discovery schemes............................ 49 3.2 Simulation assumptions.................................. 52 3.3 Data dissemination benchmarks........................... 54 3.4 Density of dataholders................................... 54 3.5 Average delay........................................... 59 3.6 Scaling properties of data dissemination.................... 64 3.7 Models of information dissemination....................... 66 3.7.1 Simple epidemic model............................. 68 3.7.2 Diffusion-controlled process......................... 69 3.8 Discussion.............................................. 71 4 Empirically-based measurements on wireless demand...... 77 4.1 Introduction............................................ 77 4.2 Campus-wide wireless infrastructure....................... 78 4.3 Monitoring and data acquisition........................... 79 4.3.1 Packet header traces............................... 80 4.3.2 http traces...................................... 81 4.3.3 snmp traces...................................... 81 4.3.4 syslog traces.................................... 82 4.3.5 Privacy assurances................................ 82 4.3.6 Client identification................................ 82 4.4 State, history, visits and sessions.......................... 83 4.5 Wireless traffic demand at APs............................ 84 4.5.1 Data acquisition................................... 84 4.5.2 Comparative analysis of wireless traffic load at APs.... 87 4.6 Application-based characterization of wireless demand....... 88 4.7 Locality of web objects................................... 94 4.7.1 http requests model.............................. 96 4.7.2 Same-client repeated requests....................... 96 4.7.3 Same-AP repeated requests......................... 97 4.7.4 AP-coresident-client repeated requests............... 97 4.7.5 Same-building and campus-wide repeated requests..... 99 4.8 Discussion.............................................. 101 5 Modeling the wireless user demand........................ 105 5.1 Introduction............................................ 105 5.2 Client access patterns.................................... 109 5.2.1 Session duration................................... 111 5.2.2 Transient sessions................................. 114 5.2.3 Revisits.......................................... 115

Contents IX 5.3 Roaming across APs..................................... 118 5.4 Arrivals of wireless clients at APs.......................... 122 5.4.1 Time-varying Poisson process....................... 123 5.4.2 Arrival process of visits at wireless hotspots.......... 125 5.5 Methodology for modeling user demand.................... 129 5.5.1 Sessions and flows................................. 129 5.5.2 Models of user demand............................. 130 5.6 Syntrig: a synthetic traffic generator....................... 134 5.7 Scalability and reusability in user demand models........... 137 5.7.1 Variation of the session arrival rate within a day...... 140 5.7.2 Variation of the session-level flow-related variables..... 140 5.8 Evaluation of user demand models......................... 142 5.8.1 Statistical-based evaluation......................... 142 5.8.2 Systems-based evaluation........................... 144 5.9 Singular spectrum analysis of traffic at APs................. 154 5.10 Related work........................................... 156 5.11 Conclusions............................................. 159 6 Conclusions and future work............................... 161 6.1 Conclusions............................................. 161 6.1.1 Mobile peer-to-peer computing...................... 161 6.1.2 Wireless measurements and modeling................ 163 6.2 Directions for future research............................. 166 6.2.1 Increasing capacity................................ 167 6.2.2 Capacity planning................................. 168 6.2.3 Network interface and channel selection.............. 169 6.2.4 Monitoring....................................... 170 6.3 Bio-inspired computing networks.......................... 172 6.4 New horizons in cross-disciplinary research.................. 174 Appendices.................................................... 175 A Appendix.................................................. 177 B Wireless measurement-based data repositories............. 179 References..................................................... 183 Index.......................................................... 207