John wiley and sons - Advanced wireless networks 4g technologies jun 2006

John wiley and sons - Advanced wireless networks 4g technologies jun 2006

Advanced Wireless Networks

4G Technologies

Savo G Glisic

University of Oulu, Finland

iii

ii

Advanced Wireless Networks

i

ii

Advanced Wireless Networks

4G Technologies

Savo G Glisic

University of Oulu, Finland

iii

Copyright  C 2006 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester,

West Sussex PO19 8SQ, England Telephone (+44) 1243 779777 Email (for orders and customer service enquiries): cs-books@wiley.co.uk Visit our Home Page on www.wiley.com

All Rights Reserved No part of this publication may be reproduced, stored in a retrieval system

or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988

or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing of the Publisher.

Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed

Other Wiley Editorial Offices

John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 22 Worcester Road, Etobicoke, Ontario, Canada M9W 1L1 Wiley also publishes its books in a variety of electronic formats Some content that appears

in print may not be available in electronic books.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library ISBN-13 978-0-470-01593-3 (HB)

ISBN-10 0-470-01593-4 (HB) Typeset in 10/12pt Times by TechBooks, New Delhi, India.

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire.

This book is printed on acid-free paper responsibly manufactured from sustainable forestry

in which at least two trees are planted for each one used for paper production.

iv

To my family

v

vi

1.2.5 A forward erasure correction booster for IP or TCP 10

1.2.7 Two-element selective ARQ booster for IP or TCP 10

vii

3.5.3 Short-term azimuth spread for forward arrival waves 65

3.9.5 Clustering models for the indoor multipath propagation channel 87

4.1.11 Sensitivity of state probabilities to hysteresis region width 114

4.3 Adaptive Hybrid ARQ Schemes for Wireless Links 126

9.2.7 TCP-Reno fast retransmit, fast (but conservative) recovery 269

9.2.9 Spurious retransmissions 270

9.4 Random Early Detection Gateways for Congestion Avoidance 279

11.1.2 Location registration and call delivery in 4G 310

11.2.3 Channel reservation with queueing – CRQ handoffs 333

11.4 Mobility Prediction in Pico- and MicroCellular Networks 344

Appendix: Distance Calculation in an Intermediate Cell 355

xii CONTENTS

12.1.8 Signal strength measurement-based distributed DCA schemes 380

12.1.11 Adaptive channel allocation reuse partitioning (ACA RUP) 385

12.4.9 Dynamic SCAC system with QoS differentiation 412

12.5.1 Centralized minimum total transmitted

12.5.3 Statistically distributed multirate power control (SDMPC) 430

12.5.7 Multiobjective distributed power and rate control (MODPRC) 43312.5.8 Multiobjective totally distributed power and rate

12.5.9 Throughput maximization/power minimization (MTMPC) 436

15.1.1 Attacks on simple cryptographic authentication 592

16.5 Cognitive Packet Networks 640

16.5.2 The random neural networks-based algorithms 644

18.3.3 Integration of routing algorithm and mobile agents 709

19.3.5 Multihop and feasible lower bounds under high attenuation 760

19.4 Cooperative Transmission in Wireless Multihop Ad Hoc Networks 76419.4.1 Transmission strategy and error propagation 767

19.5.2 Achieving the max-flow bound through a generic LCM 77419.5.3 The transmission scheme associated with an LCM 777

19.5.6 Construction of a generic LCM on an acyclic network 77919.5.7 Time-invariant LCM and heuristic construction 78019.6 Capacity of Wireless Networks Using MIMO Technology 783

21.1 Blind QoS Assessment System 817

21.4 QoS in OFDMA-based Broadband Wireless Access Systems 834

xviii

The major expectation from the fourth generation (4G) of wireless communication networks

is to be able to handle much higher data rates, which will be in the range of 1Gb inthe WLAN environment and 100 Mb in cellular networks A user, with a large range ofmobility, will access the network and will be able to seamlessly reconnect to differentnetworks, even within the same session The spectra allocation is expected to be moreflexible, and even flexible spectra sharing among the different subnetworks is anticipated

In such a ‘composite radio environment’ (CRE), there will be a need for more adaptive andreconfigurable solutions on all layers in the network For this reason the first part of the bookdeals with adaptive link, MAC, network and TCP layers including a chapter on crosslayeroptimization This is followed by chapters on mobility management and adaptive radioresource management The composite radio environment will include presence of WLAN,

cellular mobile networks, digital video broadcasting, satellite, mobile ad hoc and sensor

networks

Two additional chapters on ad hoc and sensor networks should help the reader understand

the main problems and available solutions in these fields The above chapters are followed

by a chapter on security, which is a very important segment of wireless networks

Within the more advanced solutions, the chapter on active networks covers topics likeprogrammable networks, reference models, evolution to 4G wireless networks, 4G mobilenetwork architecture, cognitive packet networks, the random neural networks based algo-rithms, game theory models in cognitive radio networks, cognitive radio networks as a gameand biologically inspired networks, including bionet architecture

Among other topics, the chapter on networks management includes self-organization

in 4G networks, mobile agent-based network management, mobile agent platform, mobileagents in multioperator networks, integration of routing algorithm and mobile agents and

ad hoc network management.

Network information theory has become an important segment of the research, and thechapter covering this topic includes effective capacity of advanced cellular network, capacity

of ad hoc networks, information theory and network architectures, cooperative transmission

in wireless multihop ad hoc networks, network coding, capacity of wireless networks using

xix

xx PREFACE

MIMO technology and capacity of sensor networks with many-to-one transmissions Twoadditional chapters, energy efficient wireless networks and QoS management, are alsoincluded in the book

As an extra resource a significant amount of material is available on the book’s

com-panion website at www.wiley.com/go/glisic in the form of three comprehensive appendices:

Appendix A provides a review of the protocol stacks for the most important existing less networks, Appendix B presents a comprehensive review of results for the MAC layerand Appendix C provides an introduction to queueing theory

wire-The material included in this book is a result of the collective effort of researchers acrossthe globe Whenever appropriate, the reference to the original work, measurement results

or diagrams is made The lists of references includes approximately 2000 titles

Discussions and cooperation with Professor P R Kumar, of the Coordinated ScienceLaboratory, University of Illinois at Urbana-Champaign, had a significant impact, espe-cially on the network information theory material presented in the book Professor ImrichChlamtac, of University of Texas at Dallas helped a great deal with the material regard-ing bioinspired nets Professor Carlos Pomalaza-Raes, of Indiana-Purdue University, USA,

inspired the presentation on ad hoc and sensor networks Professor Kaveh Pahlavan of

Worchester Polytechnic Institute, Massachusetts, inspired the presentations of the WLANtechnology Dr Moe Win of Massachusetts Institute of Technology provided a set of originaldiagrams on Ultra Wide Band Channel measurements

The author would also like to thank Professor P Leppanen, J.P M¨akel¨a, P Nissinahoand Z Nikolic, for their help with the graphics

Savo G Glisic

Oulu

Fundamentals

1.1 4G NETWORKS AND COMPOSITE RADIO ENVIRONMENT

In the wireless communications community we are witnessing more and more the existence

of the composite radio environment (CRE) and as a consequence the need for bility concepts The CRE assumes that different radio networks can be cooperating compo-

reconfigura-nents in a heterogeneous wireless access infrastructure, through which network providerscan more efficiently achieve the required capacity and quality of service (QoS) levels Re-configurability enables terminals and network elements to dynamically select and adapt

to the most appropriate radio access technologies for handling conditions encountered inspecific service area regions and time zones of the day Both concepts pose new require-ments on the management of wireless systems Nowadays, a multiplicity of radio accesstechnology (RAT) standards are used in wireless communications As shown in Figure 1.1,these technologies can be roughly categorized into four sets:

r Cellular networks that include second-generation (2G) mobile systems, such as Global

System for Mobile Communications (GSM) [1] , and their evolutions, often called 2.5Gsystems, such as enhanced digital GSM evolution (EDGE), General Packet Radio Service(GPRS) [2] and IS 136 in the USA These systems are based on TDMA technology.Third-generation (3G) mobile networks, known as Universal Mobile Telecommunica-tions Systems (UMTS; WCDMA and cdma2000) [3] are based on CDMA technology thatprovides up to 2 Mbit/s In these networks 4G solutions are expected to provide up to

100 Mbit/s The solutions will be based on a combination of multicarrier and space–time signal formats The network architectures include macro- micro- and picocellularnetworks and home (HAN) and personal area networks (PAN)

r Broadband radio access networks (BRANs) [4], or wireless local area networks (WLANs)

[5], which are expected to provide up to 1 Gb/s in 4G These technologies are based onorthogonal frequency division multiple access (OFDMA) and space–time coding

Advanced Wireless Networks: 4G Technologies Savo G Glisic

C

 2006 John Wiley & Sons, Ltd.

1

2 FUNDAMENTALS

Cellular network Access

BRAN/

WLAN Access

TDMA IS 136 EDGE, GPRS UMTS WCDMA

up to 2MBit/s cdma2000

MC CDMA Space-Time diversity

4G (100Mb)

IEEE 802.11 2.4GHz (ISM) FHSS & DSSS 5GHz

Reconfigurable Mobile Terminals

Network Reconfiguration

&

Dynamic SpectraAllocation

DVB

Sensor networks

Ad hoc networks

IP Network

Private Network PSTN

satellite

PLMN

Cellular network macro/micro/

Space–time–frequency coding,

WATM

UWB/impulse radio

IEEE 802.15.3 and 4

4G (1 Gbit)

Figure 1.1 Composite radio environment in 4G networks

r Digital video broadcasting (DVB) [6] and satellite communications.

r Ad hoc and sensor networks with emerging applications.

Although 4G is open for new multiple access schemes, the CRE concept remains tive for increasing the service provision efficiency and the exploitation possibilities of theavailable RATs The main assumption is that the different radio networks , GPRS, UMTS,BRAN/WLAN, DVB, and so on, can be components of a heterogeneous wireless accessinfrastructure A network provider (NP) can own several components of the CR infras-tructure (in other words, can own licenses for deploying and operating different RATs),and can also cooperate with affiliated NPs In any case, an NP can rely on several alterna-tive radio networks and technologies to achieve the required capacity and QoS levels, in acost-efficient manner Users are directed to the most appropriate radio networks and tech-nologies, at different service area regions and time zones of the day, based on profilerequirements and network performance criteria The various RATs are thus used in a

attrac-handset can make a handoff between different RATs The deployment of CRE systems can

be facilitated by the reconfigurability concept, which is an evolution of software-defined

radio [7, 8] CRE requires terminals that are able to work with different RATs and theexistence of multiple radio networks, offering alternative wireless access capabilities toservice area regions Reconfigurability supports the CRE concept by providing essentialtechnologies that enable terminals and network elements to dynamically (transparently andsecurely) select and adapt to the set of RATs, that is most appropriate for the conditionsencountered in specific service area regions and time zones of the day According to thereconfigurability concept, RAT selection is not restricted to those technologies pre-installed

in the network element In fact, the required software components can be dynamically loaded, installed and validated This makes it different from the static paradigm regardingthe capabilities of terminals and network elements

down-The networks provide wireless access to IP-based applications, and service continuity inlight of intrasystem mobility Integration of the network segments in the CR infrastructure isachieved through the management system for CRE (MSCRE) component attached to eachnetwork The management system in each network manages a specific radio technology;however, the platforms can cooperate The fixed (core and backbone) network will consist ofpublic and private segments based on IPv4 and IPv6-based infrastructure Mobile IP (MIP)will enable the maintenance of IP-level connectivity regardless of the likely changes in theunderlying radio technologies used that will be imposed by the CRE concept Figures 1.2and 1.3 depict the architecture of a terminal that is capable of operating in a CRE context.The terminals include software and hardware components (layer 1 and 2 functionalities)for operating with different systems The higher protocol layers, in accordance with theirpeer entities in the network, support continuous access to IP-based applications Differentprotocol busters can further enhance the efficiency of the protocol stack There is a need

to provide the best possible IP performance over wireless links, including legacy systems

bandwidth reasignment

Terminal management system

• Network discovery support

• Network selection

• Mobility management intersystem (vertical) handover

• QoS monitoring

• Profile management user preferences, terminal characteristics

Application Enhanced for TMS interactions and Information flow synchronization

Transport layer TCP/UDP

Network layer

IP Mobile IP

GPRS support protocol Layers 2/1

UMTS support protocol Layers 2/1

WLAN/BRAN Support protocol Layers 2/1

DVB-T Support protocol Layers 2/1

protocol boosters &

conversion

Figure 1.2 Architecture of a terminal that operates in a composite radio environment

4 FUNDAMENTALS

Bandwidth reassignment

Termin

(a)

al management system

• Network discovery support

• Network selection

• Mobility management intersystem (vertical) handovers

• QoS monitoring

• Profile management

• Functionality for software download, installation, validation

• Security, fault/error recovery

Application enhanced for TMS interactions and information flow synchronization

Transport layer TCP/UDP

Network layer

IP, Mobile IP

Reconfigurable modem

Interface Activeconfigurations Repository

Protocol busters and conversion

• Reconfiguration commands

• Monitoring information

Software components for communication through the selected RATs

RAT-specific and generic software components and parameters

Figure 1.3 (a) Architecture of terminal that operates in the reconfigurability context

(b) Fixed spectrum allocation compared to contiguous and fragmented DSA.(c) DSA operation configurations: (1) static (current spectrum allocations);(2) continuous DSA operations; (3) discrete DSA operations

Figure 1.3 (Continued )

Within the performance implications of link characteristics (PILC) IETF group, the cept of a performance-enhancing proxy (PEP) [9–12] has been chosen to refer to a set ofmethods used to improve the performance of Internet protocols on network paths wherenative TCP/IP performance is degraded due to the characteristics of a link Different types

con-of PEPs, depending on their basic functioning, are also distinguished Some con-of them try tocompensate for the poor performance by modifying the protocols themselves In contrast,

a symmetric/asymmetric boosting approach, transparent to the upper layers, is often bothmore efficient and flexible A common framework to house a number of different protocolboosters provides high flexibility, as it may adapt to both the characteristics of the trafficbeing delivered and the particular conditions of the links In this sense, a control plane foreasing the required information sharing (cross-layer communication and configurability)

is needed Furthermore, another requirement comes from the appearance of multihop munications as PEPs have traditionally been used over the last hop, so they should beadapted to the multihop scenario Most communications networks are subject to time andregional variations in traffic demands, which lead to variations in the degree to which thespectrum is utilized Therefore, a service’s radio spectrum can be underused at certaintimes or geographical areas, while another service may experience a shortage at the sametime/place Given the high economic value placed on the radio spectrum and the importance

com-of spectrum efficiency, it is clear that wastage com-of radio spectrum must be avoided Theseissues provide the motivation for a scheme called dynamic spectrum allocation (DSA),which aims to manage the spectrum utilized by a converged radio system and share it

6 FUNDAMENTALS

between participating radio networks over space and time to increase overall spectrumefficiency, as shown in Figure 1.3(b, c)

Composite radio systems and reconfigurability, discussed above, are potential enablers

of DSA systems Composite radio systems allow seamless delivery of services through themost appropriate access network, and close network cooperation can facilitate the sharingnot only of services, but also of spectrum Reconfigurability is also a very important issue,since with a DSA system a radio access network could potentially be allocated any frequency

at any time in any location It should be noted that the application layer is enhanced with themeans to synchronize various information streams of the same application, which could betransported simultaneously over different RATs The terminal management system (TMS)

is essential for providing functionality that exploits the CRE On the user/terminal side,the main focus is on the determination of the networks that provide, in a cost-efficientmanner, the best QoS levels for the set of active applications A first requirement is thatthe MS-CRE should exploit the capabilities of the CR infrastructure This can be done in areactive or proactive manner Reactively, the MS-CRE reacts to new service area conditions,such as the unexpected emergence of hot spots Proactively, the management system cananticipate changes in the demand pattern Such situations can be alleviated by using alternatecomponents of the CR infrastructure to achieve the required capacity and QoS levels Thesecond requirement is that the MS-CRE should provide resource brokerage functionality

to enable the cooperation of the networks of the CR infrastructure Finally, parts of theMS-CRE should be capable of directing users to the most appropriate networks of the

CR infrastructure, where they will obtain services efficiently in terms of cost and QoS Toachieved the above requirements an MS architecture such as that shown in Figure 1.4 isrequired

Mobile terminal

Managed network (component of CR infrastructure) — legacy element and network management systems

Session manager

Resource brokerage

Profile and service-level information

Status monitoring

Service configuration traffic distribution Netwotk configuration

User and control plane interface

Management Plane interface

Management plane interface

Management plane interface

Short-term operation

Mid-term operation

MS-CRE

Figure 1.4 Architecture of the MS-CRE

MS-CRE MS-CRE MS-CRE

1 Identification of new condition in service area

2 Extraction of status of Network and of SLAs

3b Offer request 3a Offer request

3c Offer request 4a Optimization request

4b Determination of new service provision pattern (QoS levels, traffic distribution to networks) Computation of Tentative reconfigurations 4c Reply

5 Solution acceptance phase Reconfiguration of managed Network and managed components

Figure 1.5 MS-CRE operation scenario

The architecture consists of three main logical entities:

r monitoring, service-level information and resource brokerage (MSRB);

r resource management strategies (RMS);

r session managers (SMs).

The MSRB entity identifies the triggers (events) that should be handled by the MS-CREand provides corresponding auxiliary (supporting) functionality The RMS entity providesthe necessary optimization functionality The SM entity is in charge of interacting withthe active subscribed users/terminals The operation steps and cooperation of the RMScomponents are shown in Figures 1.5 and 1.6, respectively In order to get an insight intothe scope and range of possible reconfigurations, we review in Appendix A (please go towww.wiley.com/go/glisic) the network and protocol stack architectures [1–58] of the basicCRE components as indicated in Figure 1.1

1.2 PROTOCOL BOOSTERS

As pointed out in Figure 1.2, an element of the reconfiguration in 4G networks is protocolbooster A protocol booster is a software or hardware module that transparently improvesprotocol performance The booster can reside anywhere in the network or end systems, andmay operate independently (one-element booster), or in cooperation with other protocol

3a Request for checking the feasibility of solution

3c Reply on feasibility

of solution

4 Selection of best feasible solution

5 Reply

6 Solution acceptance phase

7 Network configuration

Figure 1.6 Cooperation of the RMS components

Protocol messages

Booster messagesFigure 1.7 Two-element booster

boosters (multielement booster) Protocol boosters provide an architectural alternative toexisting protocol adaptation techniques, such as protocol conversion A protocol booster is asupporting agent that by itself is not a protocol It may add, delete or delay protocol messages,but never originates, terminates or converts that protocol A multielement protocol boostermay define new protocol messages to exchange among themselves, but these protocols areoriginated and terminated by protocol booster elements, and are not visible or meaningfulexternal to the booster Figure 1.7 shows the information flow in a generic two-elementbooster A protocol booster is transparent to the protocol being boosted Thus, the elimination

of a protocol booster will not prevent end-to-end communication, as would, for example,the removal of one end of a conversion [e.g transport control protocol/Internet protocol(TCP/IP) header compression unit [13]] In what follows we will present examples ofprotocol busters

UDP has an optional 16-bit checksum field in the header If it contains the value zero, itmeans that the checksum was not computed by the source Computing this checksum may

be wasteful on a reliable LAN On the other hand, if errors are possible, the checksumgreatly improves data integrity A transmitter sending data does not compute a checksumfor either local or remote destinations For reliable local communication, this saves thechecksum computation (at the source and destination) For wide-area communication, thesingle-element error detection booster computes the checksum and puts it into the UDPheader The booster could be located either in the source host (below the level of UDP) or

in a gateway machine

1.2.2 One-element ACK compression booster for TCP

On a system with asymmetric channel speeds, such as broadcast satellite, the forward (data)channel may be considerably faster than the return (acknowledgment, ACK) channel Onsuch a system, many TCP ACKs may build up in a queue, increasing round-trip time, andthus reducing the transmission rate for a given TCP window size The nature of TCP’scumulative ACKs means that any ACK acknowledges at least as many bytes of data as anyearlier ACK Consequently, if several ACKs are in a queue, it is necessary to keep only theACK that has arrived most recently A simple ACK compression booster could insure thatonly a single ACK exists in the queue for each TCP connection (A more sophisticated ACKcompression booster allows some duplicate ACKs to pass, allowing the TCP transmitter toget a better picture of network congestion.) The booster increases the protocol performancebecause it reduces the ACK latency, and allows faster transmission for a given window size

1.2.3 One-element congestion control booster for TCP

Congestion control reduces buffer overflow loss by reducing the transmission rate at thesource when the network is congested A TCP transmitter deduces information about net-work congestion by examining ACKs sent by the TCP receiver If the transmitter seesseveral ACKs with the same sequence number, then it assumes that network congestionhas caused a loss of data messages If congestion is noted in a subnet, then a congestioncontrol booster could artificially produce duplicate ACKs The TCP receiver would thinkthat data messages had been lost because of congestion, and would reduce its window size,thus reducing the amount of data it injected into the network

1.2.4 One-element ARQ booster for TCP

TCP uses ARQ to retransmit data unacknowledged by the receiver when a packet loss issuspected, such as after a retransmission time-out expires If we assume the network of

Figure 1.7 (except that booster B does not exist), then an ARQ booster for TCP: (1) will cache packets from host Y ; (2) if it sees a duplicate acknowledgment arrive from host X

and it has the next packet in the cache, then deletes the acknowledgment and retransmits

the next packet (because a packet must have been lost between the booster and host X ); and (3) will delete packets retransmitted from host Y that have been acknowledged by host X

The ARQ booster improves performance by shortening the retransmission path A typical

10 FUNDAMENTALS

application would be if host X were on a wireless network and the booster were on the

interface between the wireless and wireline networks

1.2.5 A forward erasure correction booster for IP or TCP

For many real-time and multicast applications, forward error correction coding is desirable.The two-element forward error correcting (FEC) booster uses a packet forward error cor-rection code and erasure decoding The FEC booster at the transmitter side of the networkadds parity packets The FEC booster at the receiver side removes the parity packets andregenerates missing data packets The FEC booster can be applied between any two points

in a network (including the end systems) If applied to IP, then a sequence number boosteradds sequence number information to the data packets before the first FEC booster If ap-plied to TCP (or any protocol with sequence number information), then the FEC boostercan be more efficient because: (1) it does not need to add sequence numbers; and (2) itcould add new parity information on TCP retransmissions (rather than repeating the sameparities) At the receiver side, the FEC booster could combine information from multipleTCP retransmissions for FEC decoding

1.2.6 Two-element jitter control booster for IP

For real-time communication, we may be interested in bounding the amount of jitter thatoccurs in the network A jitter control booster can be used to reduce jitter at the expense

of increased latency At the first booster element, timestamps are generated for each datamessage that passes These timestamps are transmitted to the second booster element, whichdelays messages and attempts to reproduce the intermessage interval that was measured bythe first booster element

1.2.7 Two-element selective ARQ booster for IP or TCP

For links with significant error rates, using a selective automatic repeat request (ARQ)protocol (with selective acknowledgment and selective retransmission) can significantlyimprove the efficiency compared with using TCP’s ARQ (with cumulative acknowledg-

ment and possibly go-back-N retransmission) The two-element ARQ booster uses a

se-lective ARQ booster to supplement TCP by: (1) caching packets in the upstream booster;(2) sending negative acknowledgments when gaps are detected in the downstream booster;and (3) selectively retransmitting the packets requested in the negative acknowledgments(if they are in the cache)

1.3 HYBRID 4G WIRELESS NETWORK PROTOCOLS

As indicated in Appendix A (please go to www.wiley.com/go/glisic), there are two basictypes of structure for WLAN:

(1) Infrastructure WLAN – BS-oriented network Single-hop (or cellular) networks that

require fixed base stations (BS) interconnected by a wired backbone

has BSs providing coverage for mobile hosts (MHs), ad hoc networks do not have

any centralized administration or standard support services regularly available onthe network to which the hosts may normally be connected MHs depend on eachother for communication

The BS-oriented network is more reliable and has better performance However, the ad hoc network topology is more desirable because of its low cost, plug-and-play property,

flexibility, minimal human interaction requirements, and especially battery power efficiency

It is suitable for communication in a closed area – for example, on a campus or in abuilding

To combine their strength, possible 4G concepts might prefer to add BSs to an ad hoc

network To save access bandwidth and battery power and have fast connection, the MHs

could use an ad hoc wireless network when communicating with each other in a small area.

When the MHs move out of the transmitting range, the BS could participate at this time andserve as an intermediate node The proposed method also solves some problems, such as

a BS failure or weak connection under ad hoc networks The MHs can communicate with

one another in a flexible way and freely move anywhere with seamless handoff

There have been many techniques or concepts proposed for supporting a WLAN with

and without infrastructure, such as IEEE802.11 [14], HIPERLAN [15], and ad hoc WATM

LAN [16] The standardization activities in IEEE802.11 and HIPERLAN have recognized

the usefulness of the ad hoc networking mode IEEE 802.11 enhances the ad hoc function to

the MH HIPERLAN combines the functions of two infrastructures into the MH Contrary

to IEEE802.11 and HIPERLAN, the ad hoc WATM LAN concept is based on the same

centralized wireless control framework as the BS-oriented system, but insures that MH

designed for the BS-oriented system can also participate in ad hoc networking Both the

BS oriented and ad hoc networks have some drawbacks In the BS-oriented networks, BS

manages all the MHs within the cell area and controls handoff procedures It plays a veryimportant role for WLAN If it does not work, the communication of MHs in this area will

be disrupted Under this situation, some MHs could still transmit messages to each otherwithout BS Therefore, to increase the reliability and efficiency of the BS-oriented network,MH-to-MH direct transmission capability can be added However, this is restricted to atmost two hops such that this new enhancement will not increase the protocol complexitytoo much

In the ad hoc networks, it is not easy to rebuild or maintain a connection When the

connection is built, it will be disrupted any time one MH moves out of the connectionrange So, as a compromise, the MHs could communicate with each other over the wirelessmedia, without any support from the infrastructure network components within the signaltransmission range Yet when the transmission range is less than the distance between thetwo MHs, the MHs could change back to the BS-oriented systems MH would be able to

operate in both ad hoc and BS-oriented WLAN environments.

Two different methods – one-hop and two-hop direct transmission within the BS-orientedconcept – will be considered The first method is simple and controlled by the signal strength.The second method should include the data forwarding and implementation of routingtables

12 FUNDAMENTALS

1.3.1 Control messages and state transition diagrams

To integrate the BS-oriented method and the direct transmission method, we define somecontrol messages:

(1) ACK/ACCEPT/REJECT – used to indicate the acknowledgment, acceptance, ordenial of connection or handoff request;

(2) CHANGE – used by MH to inform the sender to initiate the handoff procedure;(3) DIRECT – used by MH to inform BS that the transmission is in direct transmissionmode;

(4) SEARCH – used to find the destination; each MH receiving this message must checkthe destination address for a match;

(5) SETUP – used to establish a new connection;

(6) TEARDOWN – used for switching from BS-oriented handoff to direct transmission;

it will let BS release the channel and buffer;

(7) AGENT – used by the MH whose BS fails to accept another MH acting as a surrogatefor transmission;

(8) BELONG – used by a surrogate MH to accept another MH’s WHOSE-BS-ALIVErequest;

(9) WHOSE-BS-ALIVE – used by the MH whose BS failed to find a surrogate MH.Since a mobile host may be in BS-oriented, one-hop direct-transmission mode or two-hop direct-transmission mode, it is important that we understand the timing for modetransition

Figure 1.8 shows the state transition diagram The meaning and timing of each transitionare explained below:

(1) The receiver can receive the sender’s signal directly

(2) The receiver is a neighbor of a neighbor of the sender

(3) Neither case 1 nor 2

(4) The receiver can no longer hear the sender’s signal; however, a neighbor of the sendercan communicate with the receiver directly

(5) The receiver discovers that it can hear the sender’s signal directly

(6) The receiver can no longer hear the sender’s signal, and none of the sender’s bors can communicate directly with the receiver

neigh-(7) The receiver discovers that it can hear the sender’s signal directly

(8) No neighbors of the sender can communicate with the receiver directly

(9) The sender’s original relay neighbor fails However, the sender can find anotherneighbor that can communicate with the receiver directly

1-hop transmission mode

2-hop transmission mode

direct-BS-oriented mode Start

1-hop transmission mode

2-hop transmission mode

direct-BS-oriented mode Start

(1)

(2)

(3) (4)

(5)

(6) (7)

(8) (9)

(10)

Figure 1.8 Transition diagram for transmission mode

(10) The handoff from one BS to another occurs In Figure 1.8, we note the followingtwo points: (a) when a mobile host starts communication, it could be in any modedepending on the position of the receiving mobile host; and (b) the transition fromthe BS-oriented to two-hop direct-transmission mode is not possible because thecommunicating party cannot know that a third mobile host exists and is withinrange

1.3.2 Direct transmission

Direct transmission defines the situation where two mobile hosts communicate directly oruse a third mobile host as a relay without the help of base stations This section considersthe location management and handoff procedures when the MH moves around These func-tions are almost the same as the traditional ones However, the system must decide whetherone-hop direct transmission, two-hop direct transmission or BS-oriented transmissionmethod should be used When the sender broadcasts the connection request message, boththe BS and the MH within the sender’s signal covering area receive this message Each

MH receiving the message checks the destination ID If the destination ID matches itself,the transmission uses the one-hop direct-transmission method (If we allow two-hop directtransmission, each receiving mobile host must check its neighbor database to see if thedestination is currently a neighbor of itself.) Otherwise, the BS is used for connection.When the destination moves out of the covering area, the BS has to take over On the otherhand, when the MH moves into the covering range, the receiver has the option to stop goingthrough the BS and changes to one-hop direct transmission

Figure 1.9 Using one-hop direct-transmission mode.

1.3.3 The protocol for one-hop direct transmission

For one-hop direct-transmission mode, each case of protocol operations is described inmore detail as follows:

1.3.3.1 One-hop direct-transmission mode

(1) The sender broadcasts a SEARCH message Every node in the signal covering range(including the BS) receives the message, as shown in Figure 1.9

(2) If the receiver is within the range, it receives the message and finds out the destination

is itself It responds with the message ACK back to the sender

(3) At the same time, the BS also receives the SEARCH message It locates the MH andsends the SETUP message to the destination For direct transmission, the destinationreceives the SETUP message and sends the DIRECT message to BS Otherwise, itsends an ACCEPT message to the BS, and the communication is in BS-orientedmode

(4) The sender continues transmitting directly until the MH moves out of the coveringarea

1.3.3.2 BS-oriented Mode

(1) The sender broadcasts a SEARCH message If the receiver is out of the coveringrange, it will not receive the message, as in Figure 1.10 (It is possible that two-hopdirect-transmission mode can be used This will be explained later.)

(2) However, the BS of the sender always receives the SEARCH message It queries thereceiver’s position and sends the SETUP message to the destination

(3) When the destination receives the SETUP message, it sends an ACCEPT message

to the BS

Figure 1.10 Using the BS-oriented mode.

(4) The communication continues by BS-oriented mode until the distance between thetwo MHs is close enough and the receiver wants to change to direct transmissionmode

1.3.3.3 Handoff – out of direct transmission range

In direct transmission mode, when the destination detects that the strength of the signal isless than an acceptable value, handoff should be executed The procedures are described asfollows:

(1) The destination sends the CHANGE message to its sender

(2) As the sender receives the CHANGE request, it will send out the SEARCH messageagain Then a BS-oriented mode will be used or a two-hop direct-transmission mode,

as explained later

1.3.3.4 Handoff-BS to one-hop direct transmission

If the receiver detects that it is within the covering region of the sender’s signal and thesignal is strong enough, it has the option of switching from the BS-oriented mode to one-hopdirect-transmission mode Each step is described below and presented in Figure 1.11.(1) The sender sends the SEARCH message out Then the one-hop direct-transmissionwill be established

(2) After the sender receives the ACCEPT message from the receiver, it sends aTEARDOWN message to the BS and breaks the connection along the path

1.3.4 Protocols for two-hop direct-transmission mode

Two-hop direct-transmission mode will cover a wider area than one-hop direct-transmissionmode It allows two mobile hosts to communicate through a third mobile host acting as arelay Therefore, each mobile host must implement a neighbor database to record its current

Figure 1.11 BS-oriented handoff to one-hop direct transmission.

neighbors (A neighbor of a mobile host is another mobile host that can be connected directlywith radio waves and without the help of base stations.) Furthermore, we must handle thecase when the two-hop direct-transmission connection is disrupted because of mobility, forexample, the relay MH moves out of range or the destination moves out of range Withone-hop or two-hop direct transmission, the system reliability is increased since somemobile hosts can still communicate with others even if their base stations fail However, westill limit the number of hops in direct transmission to two for the following reasons:(1) The routing will become complicated in three or more hops direct transmission Inmultihop direct communications, handling many routing paths wastes the bandwidth

in exchanging routing information, time stamps, avoiding routing update loop, and

1.3.4.1 The neighbor database

In two-hop direct-transmission mode, each MH maintains a simple database (see Table 1.1)

to store the information of neighboring MHs within its radio covering area Each mobilehost must broadcast periodically to inform the neighboring MHs of its related information.For example, the neighbor database of MH1 in Figure 1.12 is shown in Table 1.1 In thetable, the BS-down field indicates whether or not the neighboring mobile host can detect

a nearby base station A value True means that the mobile host cannot connect to a base

station

Figure 1.12 Two-hop direct-transmission zone.

1.3.4.2 Two-hop direct-transmission mode

This situation applies when the sender and destination are both within an intermediate’scoverage area The sender transmits the data to the destination through an intermediate MH.The connection setup procedures are as follows (see Figure 1.13):

(1) If an MH is within the transmission range, it receives the sender’s message There aretwo cases: (a) when the receiver finds the destination is itself, it sends the messageACCEPT back to the sender; the connection will be the one-hop direct transmission;and (b) if the destination is not itself, the mobile host checks the neighbor database

If the destination is in the database, it forwards the SETUP message to make tion to the destination After the destination accepts the connection setup, it sendsACCEPT back to the sender

connec-(2) If the destination receives many copies of SETUP, it only accepts the first one Theredundant messages are discarded The other candidate intermediate nodes will give

up after timeout

(3) At the same time, the BS receives the SEARCH message It queries the MH andsends the SETUP message to the destination For direct transmission, the destinationreceives the SETUP message and sends the DIRECT message to its BS Otherwise,the destination accepts the connection from the base station

(4) The communication continues transmitting directly until the transmission path isbroken