Wireless IP and Building the Mobile Interne 2003 pps

The Artech House Universal Personal Communications Series Ramjee Prasad, Series Editor CDMA for Wireless Personal Communications, Ramjee Prasad IP/ATM Mobile Satellite Networks, John Far

TE AM

Team-Fly®

Wireless IP and Building the Mobile Internet

For a listing of recent titles in the Artech House Mobile Communication Series,

please turn to the back of this book.

The Artech House Universal Personal

Communications Series

Ramjee Prasad, Series Editor

CDMA for Wireless Personal Communications, Ramjee Prasad

IP/ATM Mobile Satellite Networks, John Farserotu and Ramjee Prasad

OFDM for Wireless Multimedia Communications, Richard van Nee and Ramjee Prasad Radio over Fiber Technologies for Mobile Communications Networks,

Hamed Al-Raweshidy and Shozo Komaki, editors

Simulation and Software Radio for Mobile Communications, Hiroshi Harada and

Ramjee Prasad

TDD-CDMA for Wireless Communications, Riaz Esmailzadeh and Masao Nakagawa Third Generation Mobile Communication Systems, Ramjee Prasad, Werner Mohr, and

Walter Konhäuser, editors

Towards a Global 3G System: Advanced Mobile Communications in Europe, Volume 1,

Ramjee Prasad, editor

Towards a Global 3G System: Advanced Mobile Communications in Europe, Volume 2,

Ramjee Prasad, editor

Universal Wireless Personal Communications, Ramjee Prasad

WCDMA: Towards IP Mobility and Mobile Internet, Tero Ojanperä and

Ramjee Prasad, editors

Wideband CDMA for Third Generation Mobile Communications, Tero Ojanperä and

Ramjee Prasad, editors

Wireless IP and Building the Mobile Internet, Sudhir Dixit and Ramjee Prasad, editors WLAN Systems and Wireless IP for Next Generation Communications, Neeli Prasad and

Anand Prasad, editors

Wireless IP and Building the Mobile Internet

Library of Congress Cataloging-in-Publication Data

Wireless IP and building the mobile Internet / Sudhir Dixit, Ramjee Prasad, editors.

p cm — (Artech House universal personal communications series)

Includes bibliographical references and index.

ISBN 1-58053-354-X (alk paper)

1 Wireless Internet 2 TCP/IP (Computer network protocol) 3 Wireless communication systems I Dixit, Sudhir II Prasad, Ramjee III Series TK5103.4885 W572 2002

621.382’12—dc21

2002027778

British Library Cataloguing in Publication Data

Wireless IP and building the mobile Internet — (Artech House universal personal communications series)

1 Wireless Internet—Congresses 2 Mobile communication systems—Congresses

3 TCP/IP (Computer network protocol)—Congresses

I Dixit, Sudhir II Prasad, Ramjee III International Symposium on Wireless Personal Multimedia Communications (4th : 2001 : Aalborg, Denmark)

621.3'845

ISBN 1-58053-354-X

Cover design by Igor Valdman Text design by Darrell Judd

© 2003 Sudhir Dixit and Ramjee Prasad

All rights reserved Printed and bound in the United States of America No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the publisher.

All terms mentioned in this book that are known to be trademarks or service marks have been appropriately capitalized Artech House cannot attest to the accuracy of this informa- tion Use of a term in this book should not be regarded as affecting the validity of any trade- mark or service mark.

International Standard Book Number: 1-58053-354-X

Library of Congress Catalog Card Number: 2002027778

10 9 8 7 6 5 4 3 2 1

To my wife Asha, daughter Sapna, and son Amar, who have been endless sources of strength, encouragement, and purpose in my life

—Sudhir Dixit

To my wife Jyoti, to our daughter Neeli, to our sons Anand and Rajeev, and to our granddaughters Sneha and Ruchika

—Ramjee Prasad

.

1.5 Seamless Mobility and IP 101.6 Ubiquity and Dynamic Ad Hoc Networks 111.7 Security Considerations 12

2.2 Motivation for High Data Rates and IP 182.3 Radio Interface Technologies 202.4 Cost Advantages of 3G Wireless IP 222.5 Technology Trade-Offs for 3G Voice and Data 23

2.7 Open Application Platforms for Wireless Devices 25

Chapter 6 Beyond 3G: 4G IP-Based Mobile Networks 87

6.2 Drivers for the 4G Architecture 88

6.2.3 Support for Many Different Wireless Technologies 90

6.3 4G Architecture and Research Issues 92

7.3.5 Protocols That Make Use of the Known Physical Location 117

7.6.6 Example 125

Chapter 8 Differentiated and Integrated Services for IP Applications over UMTS 133

8.3 UMTS Service Classes and Parameters 144

8.3.1 Mapping the End-to-End Service to Local Bearer Services 144

8.5 Proposed Integrated Services Parameters and Mapping 150

8.5.2 Proposed Additional Parameters for Real-Time Traffic 153 8.5.3 Proposed Additional Parameters for Non-Real-Time Traffic 154

9.3.2 Service-Based Local Policy and RSVP Sender/Receiver Proxy 170

10.3.2 VoIP over Wireless WANs (Cellular Networks) 190

10.4 QoS for Wireless VoIP Using Selective Packet

10.4.1 Analysis of the G.729 Frame Loss Concealment 191

11.2 Mobile Ad Hoc Networks 205

11.2.3 Addressing Issues and Mobility Management 207

11.3.5 A Flexible QoS Model for Mobile Ad Hoc Networks 215

11.5.1 Core-Extraction Distributed Ad Hoc Routing 219

12.4.1 Limits of UL Radio Resources for Non-Real-Time

12.4.2 Dynamic Feedback Information Multiple Access 251

13.2 A Multidimensional Concept of Radio Channels in

13.2 Future Wireless IP Networks 25713.3 The Impact of MCA Systems on Wireless IP Networks 26013.4 System Model and Performance Measures 26213.5 Frequency Diversity for an MCA System 264

13.6 Power Control for an MCA System 267

13.6.2 Scheme 2: MCA with Frequency Diversity and Power Control 270

13.7 Carrier-Grouping for an MCA System 272

Chapter 14 TCP/IP over Next-Generation Broadband Wireless Access Networks 279

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

xii CONTENTS

14.2 TCP Background 281

14.3 Design Features for 2G-BWA Networks 284

15.2 Background and Related Studies 305

15.2.1 TCP Congestion Control and the TCP Formula 305

15.2.3 Window-Based Multicast Congestion Control 307

15.3 Reliable Multicast Congestion Control Schemes 307

15.4.3 Configuration 2—A Wired/Wireless Network with 15.4.3 Wireless Links of Different Loss Rates 319 15.4.4 Configuration 3—A Large Network with Two Base Stations 321 15.4.5 Configuration 4—A Wired/Wireless/Mobile Network

15.5.3 Future Directions 330

17.1 Introduction and Background 359

17.2.1 A Generic HBR Solution for IP Micro-Mobility 366

17.5 Conclusions and Future Directions 388

18.3 Experimental Results in Test Bed 403

18.4 Optimizing Protocol Configuration 40818.5 Building Wide Experiment 410

19.2.5 Mapping: Resolving Interface Locator from Node Identifier 420

19.3 LIN6: An Application of LINA to IPv6 422

19.3.6 Compatibility with Traditional IPv6 Nodes 427

19.4 Communication Example of LIN6 427

19.6 Protocol Evaluation of Mobility Handling 432

19.7 Comparison of LIN6 and Mobile IPv6 437

19.7.1 Single Point of Failure: Mapping Agent and HA 437

21.2 The New Cost-Efficient Scheme 460

21.3 Performance Evaluation 464

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

xvi CONTENTS

Chapter 22 Enabling WAP Handoffs Between GSM and IEEE 802.11b Bearers

23.6 Interconnection Between UMTS and WLAN

23.4 Through Mobility Gateway 49423.7 Interconnection Between UMTS and WLAN Based on

24.3.3 Architecture Components for SIP-Based Push 515

25.6.3 Capacity of the Networks as a Communication System 538 25.6.4 Geographical Coverage of Service and BAN Capacity 538

26.2 Content Distribution in Wireline Web 547

26.3 Content Distribution in Wireless IP Networks 553

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

xviii CONTENTS

26.3.2 Content Adaptation and Proxy Services for Wireless Terminals 555

28.4 Standardization Issues 57628.5 Video Transcoding System Architecture and Control

28.6 Features of the Video Transcoding System 582

28.6.2 Characteristics of the Experimental System 583

Chapter 29 On Security in Wireless and IP Networks 587

29.6 DDoS, Firewalls, and Private Addressing 611

Preface

indriyanam hi charatamyan mano ‘nuvidhiyatetad asya harati prajnamvayur navam ivambhasi

As a boat on the water is swept away by a strong wind, even one of theroaming senses on which the mind focuses can carry away a man’sintelligence

—The Bhagvad Gita (2.67)

The Technical Program Committee of the Fourth International posium on Wireless Personal Multimedia Communications (WPMC’01),held September 9–12, 2001, in Aalborg, Denmark, decided to organize spe-cial invited sessions to highlight the trend of the fusion of the packet IPwireless networks This task was assigned to us We realized the importance

Sym-of the area, as well as the shortage Sym-of technical material in a single place inthe field of wireless IP and closely related technologies that form the criticalsuccess factors Therefore, we decided to invite the experts who are trulyactive in the field: the equipment manufacturers, mobile operators, andthose working in research laboratories and universities

Wireless IP and Building the Mobile Internet is the first book to take a

com-prehensive look at the convergence of wireless and Internet technologiesgiving rise to the mobile wireless Internet as we know it In short, the bookendeavors to provide an overview of all the elements required to understandand develop the future IP-based wireless multimedia communications andservices

The primary audience of this book is practicing engineers and designers,

as well as engineering managers The book is organized, however, in a mat that makes it easily adaptable for a graduate-level textbook We believethat this book provides sufficient exposure and knowledge in multiple

for-technologies with a good mix of theory and practice to understand theinternal working of the wireless Internet This book attempts to bridge thegap between research in wireless and IP communications by includingchapters from experts who have hitherto confirmed their work in theirrespective specialties with their peers.

The major objective of this book is to focus not only on the latest opments in mobility, wireless, and Internet technologies, but also to inte-grate these to provide workable end-to-end solutions We have encouragedthe authors to be concerned about the adjoining layers and technologieseven though their primary interest is to focus deeply on only one aspect ofthe technology spectrum To meet this objective of seamless interworking,

devel-we felt that it was good to have some overlap in the subject materials sincethe context is often different from one focus area to another and the percep-tion of a certain issue or problem can vary widely

This book covers a broad range of topics and has been organized intoseveral sections: wireless IP evolution, quality of service (QoS) and resourcemanagement, TCP/IP in wireless IP networks, handoff, mobility, signaling,and services and applications We illustrate the coverage of this book withthe help of Figure P.1

We have tried our best to make each chapter complete in itself Thisbook is a single authoritative source of information both for the industryprofessional and the academic in the combined field of wireless and Internetprotocols Any feedback that would improve the book or correct any errors

is greatly appreciated

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

xxii PREFACE

PREFACE xxiii

Figure P.1 Illustration of the topics covered in the book and their organization into sections.

.

Acknowledgments

The material in this book originates from the Proceedings of the Fourth

International Symposium on Wireless Personal Multimedia Communications

(WPMC’01), held September 9–12, 2001, in Aalborg, Denmark Wewould like to thank all our colleagues involved in the Organizing Commit-tee and the Technical Program Committee for their support and coopera-tion, which made this project not only possible but a huge success as well.Also, the authors and reviewers of each chapter are deeply acknowledgedfor their diligent and expert work

WPMC’01 was organized by the Center of Personkomunikation(CPK), the Yokosuka Research Park (YRP) R&D Committee, and theCommunication Research Laboratory (CRL) Their support is greatlyappreciated

Junko Prasad helped to prepare the manuscript, freeing us from theenormous editorial burden, for which we are immensely grateful We thankthe wonderful editorial staff of Artech House for their continuous supportand advice during the course of writing this book We especially thank

Dr Julie Lancashire and Ms Ruth Harris for keeping us on track, and

Ms Rebecca Allendorf for producing this book in record time

Sudhir Dixit Burlington, Massachusetts

Ramjee Prasad Aalborg, Denmark November 2002

.

Chapter 1

Wireless IP

Sudhir Dixit and Ramjee Prasad

The Internet era started in 1969 A family tree of the Internet is shown in

Figure 1.1 The key driver for the Internet protocol (IP)-based network is a

common application and service environment across multiple types of works In fact, the IP has created an open platform for innovative, flexible,and fast service creation, has enabled existing services to be supported, andhas provided IP-based mobility for all types of wired and wireless transport

net-in both the access network and the core network

The target setting is to (1) create a world class all-IP system with rapidtime-to-market and future-proof design, (2) enable flexibility for providingnew, revolutionary services while ensuring smooth network evolution andservice continuity, (3) provide access independent design for globally seam-less services, and (4) enable growth of revenue-generating systems now byleveraging the newly emerged wireless data market

The basic concept of wireless IP is shown in Figure 1.2 It is basically a erful confluence of the network interworking layer and the tetherless con-nectivity with or without mobility in a heterogeneous networkingenvironment with the promise of seamless connectivity across network sub-domains Combining the best of both wireless and IP technologies hasbrought us into the era of wireless IP Wireless IP will enable cost-effective,high-quality IP-based wireless multimedia services, including voice over IP,

pow-in large volumes [1–8]

In most developed markets the volume of data traffic has already passed that of voice traffic, and this trend will only continue to accelerate

sur-Consequently, operators and vendors have realized that there are major nomic advantages to multiplexing all types of traffic over packet switchednetworks rather than over circuit switched networks No doubt, the sametrend is being observed in the wireless mobile world This evolution is beingsupported in the core network and the radio access network by retrofitting

eco-the various second-generation (2G) standards and networks and by building new infrastructure based on third-generation (3G) standards.

The vision of enabling end-to-end connectivity has propelled IP to beadopted as a unifying layer to support a multitude of link layer standards andtechnologies This end-to-end “all-IP” vision has caused everyone to begin

looking beyond 3G systems, commonly referred to as fourth-generation

sys-tems Though the 3G systems are primarily limited to cellular/GSM wireless

access limited to terminal mobility, the next generation all-IP systems of thefuture will enable both terminal and user mobility across a range of wireless

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

2 WIRELESS IP

Figure 1.1

Family tree of the

Internet Branches and

leaves of the tree are

not shown in

chrono-logical order.

Figure 1.2

What is wireless IP?

access networks [e.g., wireless local area network (LAN) in hot spots, fixed

access, ad hoc networks]

Figure 1.3 shows an all-IP network architecture where access is through

a variety of wireless technologies—with the intelligence residing in theaccess and the backbone primarily providing the packet transport The vari-ous functions, features, and capabilities of the network and services will besupported by specialized servers potentially attached anywhere in the IPnetwork The new architecture will need to be based on the paradigm thatevery user is potentially mobile, and the network is to be able to carry all

types of traffic, some requiring strict quality of service (QoS) and others

requiring only best effort service Some key areas that are being studiedaggressively are (1) transporting heterogeneous traffic in a heterogeneousaccess network using Internet protocols, (2) seamless mobility, (3) seamlessQoS and resource management, (4) ubiquitous networking for dynamic adhoc networks, (5) security, and (6) network services and end user applica-tions

The IP has emerged as a unifying network layer protocol that ently works over heterogeneous link layer and physical layer protocols.Efforts are underway to ensure that QoS, signaling, routing, resource man-agement, mobility, and security functions and features are provided at the IPlayer and above and are mapped suitably to the lower layers so as to be con-sistent and meaningful end-to-end In this book we have focused on the dif-ferent aspects of wireless IP Keeping this objective in mind, the book isdivided into five different sections: (1) wireless IP evolution, (2) QoS andresource management, (3) TCP/IP in wireless IP networks, (4) handoff,mobility, and signaling, and (5) services and applications

transpar-1.2 Wireless IP 3

Figure 1.3

All-IP network

architecture.

This chapter provides a high-level overview of the challenges that less IP poses for the mobile Internet to become a reality for the masses.

Figure 1.4 illustrates an example of a hierarchical, heterogeneous network.The present-day terrestrial mobile systems operate in the licensed radiospectrum It is anticipated, however, that networks whose spectrum alloca-tion is not regulated in the ISM band will be abundant, requiring terminals

to support multiple air interfaces and corresponding medium access control

(MAC) layer and radio standards The problem of incompatibility with ferent frequency bands and standards in different parts of the world has led tothe development of self-configuring multimode phones and terminals capa-ble of adapting to the spectrum and standard available in a particular operat-ing region Some key technologies operating in the ISM band areBluetooth, IEEE 802.11a and b, ETSI HIPERLAN I and II, and broadbandwireless [1]—all of which are quite suitable for “hot-spot” locations wherethere is a large concentration of users carrying mobile devices (PDAs,phones, and laptops) and who could benefit from locally available low-costhigh bandwidth Examples of such hot spots are major transportation cen-ters, conference/exhibition halls, museums, and shopping malls

dif-Most wireless access networks today are point-to-point, and the coreand the backbone networks are mostly mesh based Point-to-point accesstopology suffers from unreliable connectivity and unbalanced traffic

distribution resulting in poor utilization of network resources It is clear thatlonger-term mesh topology will be the preferred topology, and the artificial

boundaries of radio access network (RAN) and core network (CN) will

disap-pear The flat distributed IP network, as it exists today in the Internet bone, will extend all the way to the end user connected via wireless access.Increasingly the access networks will be heterogeneous, and the IP layer,which will be the integrating common layer across the networks, will need

back-to deal with different access back-topologies, from full mesh back-to point-back-to-point,from dedicated bandwidth to shared bandwidth, and from best effort service

to guaranteed QoS, across different link layer technologies Since differentaccess networks will offer varying bandwidth capabilities, the resulting traf-fic profiles will be varying as well Adaptation of (and to) the appropriateradio interface, power control, radio resource, and mobility-enabled accessand handover are some of the key requirements stipulated by the technol-ogy developers and the operators alike

Although in the foreseeable future a vast number of applications will betransaction-oriented of short duration, in the long-term the multimediaflows will be of longer duration requiring stringent QoS The content of the

World Wide Web (WWW) is already sourced from different locations during

the same session; the same will happen in the wireless Internet much soonerthan anyone can imagine The model based on setting up connections prior

to data transfer is clearly not workable because of the long latenciesinvolved, especially for short flows Therefore, an always-connected con-nectionless model will most likely be the dominant approach Traffic man-agement and optimum use of radio resources are certainly going to be majorchallenges since the same infrastructure will need to deal with a variety offlows of different lengths, durations, QoS requirements, and subscriptionagreements Wireless spectrum, a finite resource, is closely regulated andlicensed to the operators For example, we illustrate in Figure 1.5 the fre-quency spectrum allocated to second- and third-generation and beyondnetworks We strongly advise the reader to peruse the chapters in Part I ofthis book for an overview of the evolution of the wireless IP technologies

Future wireless Internet architecture will be based on IP and will utilizeIPv6 mobility Although the legacy IPv4 is by far the dominant andaccepted protocol and will probably remain as such for a long time, it hasmany weaknesses, such as limited address space, lack of mobility support,and poor or unproven support for guaranteed QoS over both wireless andwired links [2] Many of these limitations will be overcome when IPv6 isuniversally deployed From a QoS perspective, the various services and

1.4 QoS and Resource Management 5

applications can be categorized in real-time and nonreal-time classes withdifferent QoS requirements The real-time applications (and consequentlythe resulting traffic) can be symmetrical or asymmetrical, putting stringentrequirements on delay, delay jitter, loss, and maintenance of QoS in inter-domain/intertechnology handovers The IP was originally designed as aconnectionless best-effort network layer transport protocol without anyQoS guarantees, with the vision of keeping the protocol simple, resilient,distributed, self-configuring, and plug-and-play This has resulted in nonde-terministic performance guarantees, and any packet- or bit-level reliabletransfer is assured at the transport (TCP) layer Because of IP’s proliferationand the embedded base of deployed IP-enabled network and end user ter-minals, it is only natural to add QoS support at the IP layer to meet the vary-ing requirements of the users and applications This would enable theoperators and the service providers to start charging for those value-added,guaranteed-quality services depending upon the users’ willingness and abil-ity to pay Technology developers, however, are faced with formidabledifficulties in meeting this challenge in a heterogeneous network environ-ment When and if the IP QoS standards are in place and implemented, themapping of the QoS mechanisms and parameters to the link layers and thephysical layers will be important These will need to be coordinated withthe radio resource and connection admission mechanisms Uniform

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

6 WIRELESS IP

Figure 1.5 Frequency ranges of second- and third-generation networks and beyond.

mapping of IP QoS over the layer 2 QoS classes (at radio layer and core work) will enable the terminals to roam globally and operate across hetero-geneous networks Parts II and III of this book address many of the aboveissues in QoS and resource management, and TCP/IP in wireless IP net-works, respectively.

net-Currently proposed QoS techniques, such as integrated service (IntServ),

differentiated service (DiffServ), and multiprotocol label switching (MPLS), are

either nonscalable, too immature, or both, rendering them unable toenforce and manage end-to-end QoS throughout an IP-based heterogene-ous network Commercial-grade IP telephony requires linkage between callsetup, end-to-end QoS setup, interdomain authorization, and accounting.This section describes the IP network model for interdomain QoS for accessand the backbone network necessary to support QoS-aware applicationservices in both public and business contexts In particular, this section pro-vides a survey of exciting protocols and addresses how to effectively com-bine them across IP-based networks

There is widespread consumer expectation that commercial IP-based3G devices and fixed IP devices will need to provide QoS equal to that of

cellular digital circuit switched telephony, such as the Global System for

Mobile Communications (GSM) In order to achieve this, a mechanism is

needed to incorporate end-to-end QoS in IP networks [3] The support ofQoS in any network requires the use of network resources, and the primaryobjective is to allocate and manage all dedicated bandwidth, control jitter-ing, and bound latency (required for real-time and interactive traffic), aswell as to meet data rate and reliability commitments QoS support enablespremium services to prioritize the delivery of certain IP packets at theexpense of packets carrying best-effort traffic Thus, best-effort packets willsuffer from degraded performance (e.g., delay) when traffic is heavy Conse-quently, service providers must either over-dimension the network toensure adequate capacity, or limit the admission of best-effort users, in order

to ensure an acceptable QoS for applications that are not real-time sensitive,such as ftp, e-mail, and Web browsing QoS delivery in access networks

such as Universal Mobile Telecommunications System (UMTS) and wireless

LANs (WLANs) is determined by local network usage policy

Implement-ing Session Initiation Protocol (SIP) servers in such networks can help to

enforce policy for all SIP calls SIP call parameters, such as endpointaddresses, call ID, time, authorization requests, and tokens, must be ex-changed with policy servers, trusted authorization and accounting servers,

in order to install and tear down QoS policy in Resource Reservation Protocol

(RSVP)-enabled routers These parameters may apply to either mobile orfixed end access points Mobile IP has been standardized by the InternetEngineering Task Force (IETF) to support mobile users and Internetdevices

1.4 QoS and Resource Management 7

1.4.1 QoS Network Model

Figure 1.6 illustrates a simple QoS reference model for various real-time IPcommunications; this covers call setup, QoS reservation, policy, authoriza-tion, and payments The approach used here applies standard end-to-endRSVP in access networks and DiffServ with MPLS in the backbonenetworks Key elements in the hierarchical model include the access net-work, backbone network, and clearinghouse (centralized QoS managementunit)

Strict priority for real-time traffic (such as voice) facilitates simple androbust QoS implementation in a private IP network where policy controland individual specific accounting are not required Extension, however, ofthe strict priority to real-time traffic between the different domains acrossthe Internet would prevent service providers from exercising policy controland accounting, which would make the services very costly and reduce theincentive for service providers to deploy QoS mechanisms in their net-works Furthermore, the model of strict priority service without policy andaccounting may be under pressure when applied to high-bandwidth appli-cations, such as video on demand Economic trade-offs between the sub-scription charge and the guaranteed QoS level will be required

•Access network: An IP network to which users directly connect their

hosts/clients for IP connectivity The access network is part of a single

administrative domain, such as those operated by Internet service

provid-W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

8 WIRELESS IP

Figure 1.6

A reference model for

QoS support for IP

telephony in an

IP-based heterogeneous

network.

ers (ISPs), corporate networks, the government, and educational

organizations

•Backbone network: One or more backbone networks may be between

two or more access networks The backbone network in our modelhas no knowledge of individual microflows, such as phone calls be-tween parties connected to access networks

•Clearinghouse: Given the large number of access networks belonging to

different administrative domains, it is not possible to have service-level

specifications (SLSs) between all domains on the Internet

Clearing-houses can facilitate the authorization and logging or accounting tween domains for premium services, such as QoS Current SLSs arestatic in nature, although there is interest in signaling for dynamic de-livery of QoS between service providers, such as in the case ofbandwidth-broker-mediated services

be-1.4.2 Resource Management Problems

Spectrum resource management is an important topic and will continue to

be in the near and distant future Resource management takes on newdimensions and can no longer be restricted to being a matter of spectrumutilization only Other important components are mobile equipment man-agement and infrastructure deployment and cost structure

Future systems are expected to require much higher data rates than rent systems, since most of the current resource management systems are notdirectly tied to any specific development of new methods Data rates in per-sonal communication systems, however, will in many cases be limited bypropagation conditions such as distance loss and multipath The primaryconstraining factor is the link budget Since the required transmitted powerincreases linearly with the bandwidth, high-speed wireless access will havebut a limited range This will increase the complexity of the resource man-agement schemes

cur-If the bandwidth as such is not important to the design and performance

of radio resource management (RRM) algorithms, the traffic characteristics are.

The key resource management problems in multimedia type systems arerelated to the data rates, and delay constraints of traffic in small cell environ-ments will exhibit very large peak-to-average capacity demands Videousers with absolute delay requirements may require considerable portions ofthe spectrum that they share with e-mail message traffic with no such abso-lute constraints Dynamic channel allocation (i.e., statistical multiplexing)will provide even larger capacity gains in these situations than in today’smobile phone scenarios

1.4 QoS and Resource Management 9

1.5 Seamless Mobility and IP

Fast-forward to a few years in the future, and we will be living in an erawhere seamless mobility across heterogeneous wireless networks will betaken for granted [2, 4] In today’s networks, though, mobility is supported

at layer 2 in WLANs and 2G/3G networks, which prohibits roaming acrossheterogeneous access networks and routing domains In contrast, mobilitysupport at the network layer (IP) will allow Internet-wide (global) mobilitywhere the layer 2 and physical layers will be completely transparent, albeit atthe cost of increased complexity and longer propagation delays Any issues

of complexity, cost, and performance will eventually become irrelevantwith the everyday advances in the technologies involved and the volumesthat are expected There are three key reference models for mobility under

study in the IETF: Mobile IP (MIP), Handoff Aware Wireless Access Internet

Infrastructure (HAWAII), and Cellular IP (CIP) Each reference model has its

own pros and cons Supporting them will require the mobile terminal to bemobility-aware, and the legacy IP protocol stacks that have already beenimplemented in them will have to be replaced with the new referenceimplementations The reader will learn more about seamless mobility inmany of the chapters devoted to this topic in Part IV of this book

Most work until now has focused on terminal mobility, which hasalready been successfully proven in commercial networks Terminal mobil-ity allows the network to route calls or packets to a mobile device regardless

of the type of network to which it is attached [2] If we add to this the bility of user or personal mobility (i.e., the user is not tied to a personal ter-minal), an additional set of requirements and complexities have to be dealtwith Personal mobility enables the users to access their services regardless oftheir point of attachment or the type of terminal they are using [5], wherethe device they use may not belong to them It is akin to the very successfuluse of the e-mail alias in the Internet today (e.g., Yahoo!, MSN, Hotmail,and AOL) where one can access his or her e-mail from any terminal, anytime, and from anywhere in the world as long as he or she is connected tothe Internet Personal mobility allows a user’s calls and environment to beforwarded from one terminal to another The user should be contactablewith only one identity that should be mappable to an address where hispackets can be routed, and this mapping should not be tied to a single termi-nal or a single operator since a user may roam from one operator’s network

capa-to another and his packets may follow paths that may cross many differentoperators’ networks of one or more types The control of the identity map-ping must remain with the user who should ultimately decide who canreach him or her on which network during what times and with what level

of security/confidentiality

W i r e l e s s I P a n d B u i l d i n g t h e M o b i l e I n t e r n e t

10 WIRELESS IP

A vision of a seamless network of complementary access systems isdepicted in Figure 1.7.

Ubiquitous computing or networking has received a great deal of attentionrecently [9] Ubiquitous networking refers to the dynamic ad hoc formation

of collaborating entities (people and devices), which adapt to network ditions and network types, and are basically access-network agnostic Theyself-configure autonomously when nodes and services appear, negotiate,migrate, and disappear [2, 3, 9, 10] Although there are many differentdirections being pursued toward a ubiquitous networking infrastructure,there are some key fundamental characteristics that are common to them all.Architecturally, these fundamental characteristics may be classified into the

con-collaboration-level infrastructure and communication-level infrastructure [10].

The former enables devices to automatically discover each other, form laborative regions, and interact with each other at the application or servicelevel; the latter enables the exchange of collaboration-level messages and ofapplication data itself by means of the networking infrastructure

col-Both the collaboration-level and the communication-level tures will have major implications on all the layers of the protocol stack(including the wireless IP) and vice versa, depending on how soon and towhat extent the protocols can be enhanced In the meantime, the industry ismaking every effort to use whatever standards are available today to providethe necessary enhancements to deliver dynamic ad hoc networks using the

infrastruc-1.6 Ubiquity and Dynamic Ad Hoc Networks 11

Figure 1.7

Seamless network of

complementary access

systems.

concepts of ubiquitous networking For example, work is already in

prog-ress to interconnect multiple wireless devices on one’s person by a personal (or body) area network (PAN or BAN) PAN is a person-centered network

concept that enhances our personal experience by connecting all knownand future personal devices and equipment within a limited range (of, say,2m) using wireless techniques (e.g., Bluetooth) PAN will cover the per-sonal space surrounding the person within the distance to which the voicereaches It will have a capacity in the range of 10 bps to 10 Mbps Figure 1.8illustrates the position of PAN with respect to other systems (B-PAN standsfor broadband-PAN)

The following are the challenges and open issues for PAN:

•Low-power, low-cost radio integration;

•Definition of possible physical layers and access techniques;

•Ad hoc networking;

•Middleware architecture;

•Security (different security techniques, gatekeeping functionalities);

•Overall system concept;

•Human aspects

Authentication, data security, and privacy are of paramount interest in thewireless networks [2] Security features in 2G and 3G systems focus onmainly two aspects: one to authenticate the user with a billing system andthe other to encrypt the data so that it cannot be eavesdropped If the call isprepaid, then there is no association of the billing system with the identity ofthe caller In future IP networks, this dependence on the operator (and the

associated billing system) would need to be minimized as much as possible,and any negotiation with respect to authentication and identity disclosureshould rest with the end users The security mechanisms present in today’swireless networks are very different from those present in the fixed net-works There are typically four levels of security applicable in both fixed andwireless networks: (1) authentication at login, (2) end-to-end security at theapplication level, (3) network level security, and (4) link level security.

In a heterogeneous environment some of the major issues are: (1) how

to get two access routers that may have never known each other to trusteach other, (2) the level and type of security support may be different at dif-ferent access routers connected to different types of access networks, (3) theprovision of security and trust features for access routers at the client leveland ramifications of this at the IP layer, (4) the enabling of different levels ofidentity authentication and data security (either on demand or depending

on the context) without reference to operator, and (5) the optimization ofend-to-end security while removing security at certain layers to minimizeduplication while the data path crosses multiple technologies and operatordomains The implications on the IP or vice versa may be significant Chap-ter 29 examines the various aspects of security in wireless and IP networks

An open IP-enabled wireless network will provide vast opportunities for amyriad of new services and applications to be developed, upon which onecannot begin to speculate The younger generation is already comfortableusing mobile devices and PCs while continuously challenging established

business models and ethical boundaries On the data front, short message

serv-ices (SMS) have already been a huge success, and similar efforts are underway

now to develop and offer mobile multimedia messaging with a big push for

multimedia messaging services (MMS) [11] The MMS can use up to 57.6 Kbps

for sending and retrieving data MMS-compatible handsets are alreadybeginning to ship It will not be long before location-aware services begin toappear as well Part V of this book examines the issues of wireless IP fromthe standpoint of some of the key services and applications that will need to

be supported in the near future The real challenge for the future can beexplained by the following equation to achieve IP-based wireless multime-dia communications: