Advanced Wireless Networks

Transcription

1 Advanced Wireless Networks 4G Technologies Savo G. Glisic University of Oulu, Finland

2

3 Advanced Wireless Networks

4

5 Advanced Wireless Networks 4G Technologies Savo G. Glisic University of Oulu, Finland

6 Copyright C 2006 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England Telephone (+44) (for orders and customer service enquiries): cs-books@wiley.co.uk Visit our Home Page on 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 ed to permreq@wiley.co.uk, or faxed to (+44) Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Other Wiley Editorial Offices John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA , USA Wiley-VCH Verlag GmbH, Boschstr. 12, D 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 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 (HB) ISBN (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.

7 To my family

8

9 Contents Preface xix 1 Fundamentals G Networks and Composite Radio Environment Protocol Boosters One-element error detection booster for UDP One-element ACK compression booster for TCP One-element congestion control booster for TCP One-element ARQ booster for TCP A forward erasure correction booster for IP or TCP Two-element jitter control booster for IP Two-element selective ARQ booster for IP or TCP Hybrid 4G Wireless Network Protocols Control messages and state transition diagrams Direct transmission The protocol for one-hop direct transmission Protocols for two-hop direct-transmission mode Green Wireless Networks 20 References 22 2 Physical Layer and Multiple Access Advanced Time Division Multiple Access-ATDMA Code Division Multiple Access Orthogonal Frequency Division Multiplexing Multicarrier CDMA Ultrawide Band Signal MIMO Channels and Space Time Coding 41 References 42

10 viii CONTENTS 3 Channel Modeling for 4G Macrocellular Environments (1.8 GHz) Urban Spatal Radio Channels in Macro/MicroCell Environment (2.154 GHz) Description of environment Results MIMO Channels in Micro- and PicoCell Environment (1.71/2.05 GHz) Measurement set-ups The eigenanalysis method Definition of the power allocation schemes Outdoor Mobile Channel (5.3 GHz) Path loss models Path number distribution Rotation measurements in an urban environment Microcell Channel (8.45 GHz) Azimuth profile Delay profile for the forward arrival waves Short-term azimuth spread for forward arrival waves Wireless MIMO LAN Environments (5.2 GHz) Data evaluation Capacity computation Measurement environments Indoor WLAN Channel (17 GHz) Indoor WLAN Channel (60 GHz) Definition of the statistical parameters UWB Channel Model The large-scale statistics The small-scale statistics The statistical model Simulation steps Clustering models for the indoor multipath propagation channel Path loss modeling 90 References 93 4 Adaptive and Reconfigurable Link Layer Link Layer Capacity of Adaptive Air Interfaces The MAC channel model The Markovian model Goodput and link adaptation Switching hysteresis Link service rate with exact mode selection Imperfections in the adaptation chain Estimation process and estimate error Channel process and estimation delay Feedback process and mode command reception Link service rate with imperfections Sensitivity of state probabilities to hysteresis region width Estimation process and estimate error Feedback process and acquisition errors 118

11 CONTENTS ix 4.2 Adaptive Transmission in Ad Hoc Networks Adaptive Hybrid ARQ Schemes for Wireless Links RS codes PHY and MAC frame structures Error-control schemes Performance of adaptive FEC Simulation results Stochastic Learning Link Layer Protocol Stochastic learning control Adaptive link layer protocol Infrared Link Access Protocol The IrLAP layer IrLAP functional model description 142 References Adaptive Medium Access Control WLAN Enhanced Distributed Coordination Function Adaptive MAC for WLAN with Adaptive Antennas Description of the protocols MAC for Wireless Sensor Networks S-MAC protocol design Periodic listen and sleep Collision avoidance Coordinated sleeping Choosing and maintaining schedules Maintaining synchronization Adaptive listening Overhearing avoidance and message passing Overhearing avoidance Message passing MAC for Ad Hoc Networks Carrier sense wireless networks Interaction with upper layers 174 References Teletraffic Modeling and Analysis Channel Holding Time in PCS Networks 179 References Adaptive Network Layer Graphs and Routing Protocols Elementary concepts Directed graph Undirected graph Degree of a vertex Weighted graph Walks and paths 193

12 x CONTENTS Connected graphs Trees Spanning tree MST computation Shortest path spanning tree Graph Theory Routing with Topology Aggregation Network and Aggregation Models Line segment representation QoS-aware topology aggregation Mesh formation Star formation Line-segment routing algorithm Performance measure Performance example 224 References Effective Capacity Effective Traffic Source Parameters Effective traffic source Shaping probability Shaping delay Performance example Effective Link Layer Capacity Link-layer channel model Effective capacity model of wireless channels Physical layer vs link-layer channel model Performance examples 253 References Adaptive TCP Layer Introduction A large bandwidth-delay product Buffer size Round-trip time Unfairness problem at the TCP layer Noncongestion losses End-to-end solutions Bandwidth asymmetry TCP Operation and Performance The TCP transmitter Retransmission timeout Window adaptation Packet loss recovery TCP-OldTahoe (timeout recovery) TCP-Tahoe (fast retransmit) TCP-Reno fast retransmit, fast (but conservative) recovery 269

13 CONTENTS xi TCP-NewReno (fast retransmit, fast recovery) Spurious retransmissions Modeling of TCP operation TCP for Mobile Cellular Networks Improving TCP in mobile environments Mobile TCP design The SH-TCP client The M-TCP protocol Performance examples Random Early Detection Gateways for Congestion Avoidance The RED algorithm Performance example TCP for Mobile Ad Hoc Networks Effect of route recomputations Effect of network partitions Effect of multipath routing ATCP sublayer ATCP protocol design Performance examples 289 References Crosslayer Optimization Introduction A Cross-Layer Architecture for Video Delivery 296 References Mobility Management Introduction Mobility management in cellular networks Location registration and call delivery in 4G Cellular Systems with Prioritized Handoff Channel assignment priority schemes Channel reservation CR handoffs Channel reservation with queueing CRQ handoffs Performance examples Cell Residing Time Distribution Mobility Prediction in Pico- and MicroCellular Networks PST-QoS guarantees framework Most likely cluster model 347 Appendix: Distance Calculation in an Intermediate Cell 355 References Adaptive Resource Management Channel Assignment Schemes Different channel allocation schemes Fixed channel allocation Channel borrowing schemes 371

14 xii CONTENTS Hybrid channel borrowing schemes Dynamic channel allocation Centralized DCA schemes Cell-based distributed DCA schemes Signal strength measurement-based distributed DCA schemes One-dimensional cellular systems Fixed reuse partitioning Adaptive channel allocation reuse partitioning (ACA RUP) Resource Management in 4G Mobile Agent-based Resource Management Advanced resource management system CDMA Cellular Multimedia Wireless Networks Principles of SCAC QoS differentiation paradigms Traffic model Performance evaluation Related results Modeling-based static complete-sharing MdCAC system Measurement-based complete-sharing MsCAC system Complete-sharing dynamic SCAC system Dynamic SCAC system with QoS differentiation Example of a single-class system NRT packet access control Assumptions Estimation of average upper-limit (UL) data throughput DFIMA, dynamic feedback information-based access control Performance examples Implementation issues Joint Data Rate and Power Management Centralized minimum total transmitted power (CMTTP) algorithm Maximum throughput power control (MTPC) Statistically distributed multirate power control (SDMPC) Lagrangian multiplier power control (LRPC) Selective power control (SPC) RRM in multiobjective (MO) framework Multiobjective distributed power and rate control (MODPRC) Multiobjective totally distributed power and rate control (MOTDPRC) Throughput maximization/power minimization (MTMPC) Dynamic Spectra Sharing in Wireless Networks Channel capacity Channel models Diversity reception Performance evaluation Multiple access techniques and user capacity Multiuser detection 442

15 CONTENTS xiii Interference and coexistence Channel estimation/imperfections Signal and interference model Receiver structure Interference rejection circuit model Performance analysis Performance examples 451 References Ad Hoc Networks Routing Protocols Routing protocols Reactive protocols Hybrid Routing Protocol Loop-back termination Early termination Selective broadcasting (SBC) Scalable Routing Strategies Hierarchical routing protocols Performance examples FSR (fisheye routing) protocol Multipath Routing Clustering Protocols Introduction Clustering algorithm Clustering with prediction Cashing Schemes for Routing Cache management Distributed QoS Routing Wireless links reliability Routing Routing information Token-based routing Delay-constrained routing Tokens Forwarding the received tokens Bandwidth-constrained routing Forwarding the received tickets Performance example 527 References Sensor Networks Introduction Sensor Networks Parameters Pre-deployment and deployment phase Post-deployment phase Re-deployment of additional nodes phase 539

16 xiv CONTENTS 14.3 Sensor Networks Architecture Physical layer Data link layer Network layer Transport layer Application layer Mobile Sensor Networks Deployment Directed Diffusion Data propagation Reinforcement Aggregation in Wireless Sensor Networks Boundary Estimation Number of RDPs in P Kraft inequality Upper bounds on achievable accuracy System optimization Optimal Transmission Radius in Sensor Networks Back-off phenomenon Data Funneling Equivalent Transport Control Protocol in Sensor Networks 575 References Security Authentication Attacks on simple cryptographic authentication Canonical authentication protocol Security Architecture Key Management Encipherment Modification detection codes Replay detection codes Proof of knowledge of a key Point-to-point key distribution Security Management in GSM Networks Security Management in UMTS Security Architecture for UMTS/WLAN Interworking Security in Ad Hoc Networks Self-organized key management Security in Sensor Networks 622 References Active Networks Introduction Programable Networks Reference Models IETF ForCES Active networks reference architecture Evolution to 4G Wireless Networks 635

17 CONTENTS xv 16.4 Programmable 4G Mobile Network Architecture Cognitive Packet Networks Adaptation by cognitive packets The random neural networks-based algorithms Game Theory Models in Cognitive Radio Networks Cognitive radio networks as a game Biologically Inspired Networks Bio-analogies Bionet architecture 656 References Network Deployment Cellular Systems with Overlapping Coverage Imbedded Microcell in CDMA Macrocell Network Macrocell and microcell link budget Performance example Multitier Wireless Cellular Networks The network model Performance example Local Multipoint Distribution Service Interference estimations Alternating polarization Self-organization in 4G Networks Motivation Networks self-organizing technologies 691 References Network Management The Simple Network Management Protocol Distributed Network Management Mobile Agent-based Network Management Mobile agent platform Mobile agents in multioperator networks Integration of routing algorithm and mobile agents Ad Hoc Network Management Heterogeneous environments Time varying topology Energy constraints Network partitioning Variation of signal quality Eavesdropping Ad hoc network management protocol functions ANMP architecture 717 References Network Information Theory Effective Capacity of Advanced Cellular Networks G cellular network system model The received signal 730

18 xvi CONTENTS Multipath channel: near far effect and power control Multipath channel: pointer tracking error, rake receiver and interference canceling Interference canceler modeling: nonlinear multiuser detectors Approximations Outage probability Capacity of Ad Hoc Networks Arbitrary networks Random networks Arbitrary networks: an upper bound on transport capacity Arbitrary networks: lower bound on transport capacity Random networks: lower bound on throughput capacity Information Theory and Network Architectures Network architecture Definition of feasible rate vectors The transport capacity Upper bounds under high attenuation Multihop and feasible lower bounds under high attenuation The low-attenuation regime The Gaussian multiple-relay channel Cooperative Transmission in Wireless Multihop Ad Hoc Networks Transmission strategy and error propagation OLA flooding algorithm Simulation environment Network Coding Max-flow min-cut theorem (mfmct) Achieving the max-flow bound through a generic LCM The transmission scheme associated with an LCM Memoryless communication network Network with memory Construction of a generic LCM on an acyclic network Time-invariant LCM and heuristic construction Capacity of Wireless Networks Using MIMO Technology Capacity metrics Capacity of Sensor Networks with Many-to-One Transmissions Network architecture Capacity results 793 References Energy-efficient Wireless Networks Energy Cost Function Minimum Energy Routing Maximizing Network Lifetime Energy-efficient MAC in Sensor Networks Staggered wakeup schedule 810 References 812

19 CONTENTS xvii 21 Quality-of-Service Management Blind QoS Assessment System System modeling QoS Provisioning in WLAN Contention-based multipolling Polling efficiency Dynamic Scheduling on RLC/MAC Layer DSMC functional blocks Calculating the high service rate Heading-block delay Interference model Normal delay of a newly arrived block High service rate of a session QoS in OFDMA-based Broadband Wireless Access Systems Iterative solution Resource allocation to maximize capacity Predictive Flow Control and QoS Predictive flow control model 843 References 847 Index 853

20

21 Preface 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 in the WLAN environment and 100 Mb in cellular networks. A user, with a large range of mobility, will access the network and will be able to seamlessly reconnect to different networks, even within the same session. The spectra allocation is expected to be more flexible, 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 and reconfigurable solutions on all layers in the network. For this reason the first part of the book deals with adaptive link, MAC, network and TCP layers including a chapter on crosslayer optimization. This is followed by chapters on mobility management and adaptive radio resource 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 like programmable networks, reference models, evolution to 4G wireless networks, 4G mobile network architecture, cognitive packet networks, the random neural networks based algorithms, game theory models in cognitive radio networks, cognitive radio networks as a game and 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, mobile agents 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 the chapter 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

22 xx PREFACE MIMO technology and capacity of sensor networks with many-to-one transmissions. Two additional chapters, energy efficient wireless networks and QoS management, are also included in the book. As an extra resource a significant amount of material is available on the book s companion website at in the form of three comprehensive appendices: Appendix A provides a review of the protocol stacks for the most important existing wireless networks, Appendix B presents a comprehensive review of results for the MAC layer and Appendix C provides an introduction to queueing theory. The material included in this book is a result of the collective effort of researchers across the 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 Science Laboratory, University of Illinois at Urbana-Champaign, had a significant impact, especially on the network information theory material presented in the book. Professor Imrich Chlamtac, of University of Texas at Dallas helped a great deal with the material regarding 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 WLAN technology. Dr. Moe Win of Massachusetts Institute of Technology provided a set of original diagrams on Ultra Wide Band Channel measurements. The author would also like to thank Professor P. Leppanen, J.P. Mäkelä, P. Nissinaho and Z. Nikolic, for their help with the graphics. Savo G. Glisic Oulu

23 1 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 reconfigurability concepts. The CRE assumes that different radio networks can be cooperating components in a heterogeneous wireless access infrastructure, through which network providers can more efficiently achieve the required capacity and quality of service (QoS) levels. Reconfigurability enables terminals and network elements to dynamically select and adapt to the most appropriate radio access technologies for handling conditions encountered in specific service area regions and time zones of the day. Both concepts pose new requirements on the management of wireless systems. Nowadays, a multiplicity of radio access technology (RAT) standards are used in wireless communications. As shown in Figure 1.1, these technologies can be roughly categorized into four sets: Cellular networks that include second-generation (2G) mobile systems, such as Global System for Mobile Communications (GSM) [1], and their evolutions, often called 2.5G systems, 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 Telecommunications Systems (UMTS; WCDMA and cdma2000) [3] are based on CDMA technology that provides 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 picocellular networks and home (HAN) and personal area networks (PAN). 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 on orthogonal frequency division multiple access (OFDMA) and space time coding. Advanced Wireless Networks: 4G Technologies C 2006 John Wiley & Sons, Ltd. Savo G. Glisic

24 2 FUNDAMENTALS Sensor networks PLMN PSTN Ad hoc networks IP Network Private Network Cellular network macro/micro/ Pico/PAN TDMA IS 136 EDGE, GPRS UMTS WCDMA up to 2MBit/s cdma2000 MC CDMA Space-Time diversity 4G (100Mb) satellite Cellular network Access DVB Network Reconfigur ation & Dynamic Spectra Allocation Reconfigurable Mobile Terminals BRAN/ WLAN Access IEEE GHz (ISM) FHSS & DSSS 5GHz WLAN, WPAN OFDM > 10 Mbit/s Hiperlan and IEEE 802.x 54 Mb (indoor) Hiperaccess (wider area) Hiperlink 155 Mb Space time frequency coding, WATM UWB/impulse radio IEEE and 4 4G (1 Gbit) Figure 1.1 Composite radio environment in 4G networks. Digital video broadcasting (DVB) [6] and satellite communications. Ad hoc and sensor networks with emerging applications. Although 4G is open for new multiple access schemes, the CRE concept remains attractive for increasing the service provision efficiency and the exploitation possibilities of the available 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 access infrastructure. A network provider (NP) can own several components of the CR infrastructure (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 alternative radio networks and technologies to achieve the required capacity and QoS levels, in a cost-efficient manner. Users are directed to the most appropriate radio networks and technologies, at different service area regions and time zones of the day, based on profile requirements and network performance criteria. The various RATs are thus used in a

25 4G NETWORKS AND COMPOSITE RADIO ENVIRONMENT 3 complementary manner rather than competing with each other. Even nowadays a mobile 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 the existence of multiple radio networks, offering alternative wireless access capabilities to service area regions. Reconfigurability supports the CRE concept by providing essential technologies that enable terminals and network elements to dynamically (transparently and securely) select and adapt to the set of RATs, that is most appropriate for the conditions encountered in specific service area regions and time zones of the day. According to the reconfigurability concept, RAT selection is not restricted to those technologies pre-installed in the network element. In fact, the required software components can be dynamically downloaded, installed and validated. This makes it different from the static paradigm regarding the capabilities of terminals and network elements. The networks provide wireless access to IP-based applications, and service continuity in light of intrasystem mobility. Integration of the network segments in the CR infrastructure is achieved through the management system for CRE (MSCRE) component attached to each network. The management system in each network manages a specific radio technology; however, the platforms can cooperate. The fixed (core and backbone) network will consist of public 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 the underlying radio technologies used that will be imposed by the CRE concept. Figures 1.2 and 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 their peer entities in the network, support continuous access to IP-based applications. Different protocol 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. 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 protocol boosters & conversion bandwidth reasignment 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 Figure 1.2 Architecture of a terminal that operates in a composite radio environment.

26 4 FUNDAMENTALS (a) Application enhanced for TMS interactions and information flow synchronization Terminal 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 Transport layer TCP/UDP Network layer IP, Mobile IP Protocol busters and conversion Active Interface configurations Reconfigurable modem Repository Bandwidth reassignment Reconfiguration commands Monitoring information Software components for communication through the selected RATs RAT-specific and generic software components and parameters (b) Fixed Contiguous Fragmented Frequency RAN1 RAN1 RAN1 RAN1 RAN1 RAN1 RAN2 RAN2 RAN2 RAN2 RAN2 RAN2 Frequency RAN1 RAN2 RAN1 RAN2 RAN1 RAN2 RAN1 RAN2 RAN1 RAN2 RAN1 RAN2 Frequency RAN1 RAN2 RAN3 RAN1 RAN2 RAN1 RAN3 RAN1 RAN2 RAN3 RAN1 RAN2 RAN3 RAN1 RAN2 RAN3 RAN1 RAN3 RAN2 RAN3 Time or region Time or region Time or region 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.

27 4G NETWORKS AND COMPOSITE RADIO ENVIRONMENT 5 (c) DAB DAB DAB Analog TV and DVB-T Analog TV and DVB-T Analog TV and DVB-T WLAN UMTS GSM GSM UMTS Contiguous DSA GSM Fragmented DSA (2) GSM GSM GSM Contiguous DSA (1) Contiguous DSA WLAN UMTS Analog TV and DVB-T UMTS GSM WLAN WLAN WLAN Fragmented DSA (3) Figure 1.3 (Continued ) Within the performance implications of link characteristics (PILC) IETF group, the concept of a performance-enhancing proxy (PEP) [9 12] has been chosen to refer to a set of methods used to improve the performance of Internet protocols on network paths where native TCP/IP performance is degraded due to the characteristics of a link. Different types of PEPs, depending on their basic functioning, are also distinguished. Some of them try to compensate for the poor performance by modifying the protocols themselves. In contrast, a symmetric/asymmetric boosting approach, transparent to the upper layers, is often both more efficient and flexible. A common framework to house a number of different protocol boosters provides high flexibility, as it may adapt to both the characteristics of the traffic being delivered and the particular conditions of the links. In this sense, a control plane for easing the required information sharing (cross-layer communication and configurability) is needed. Furthermore, another requirement comes from the appearance of multihop communications as PEPs have traditionally been used over the last hop, so they should be adapted to the multihop scenario. Most communications networks are subject to time and regional variations in traffic demands, which lead to variations in the degree to which the spectrum is utilized. Therefore, a service s radio spectrum can be underused at certain times or geographical areas, while another service may experience a shortage at the same time/place. Given the high economic value placed on the radio spectrum and the importance of spectrum efficiency, it is clear that wastage of radio spectrum must be avoided. These issues 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

28 6 FUNDAMENTALS between participating radio networks over space and time to increase overall spectrum efficiency, 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 the most appropriate access network, and close network cooperation can facilitate the sharing not 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 the means to synchronize various information streams of the same application, which could be transported 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-efficient manner, the best QoS levels for the set of active applications. A first requirement is that the MS-CRE should exploit the capabilities of the CR infrastructure. This can be done in a reactive 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 can anticipate changes in the demand pattern. Such situations can be alleviated by using alternate components of the CR infrastructure to achieve the required capacity and QoS levels. The second 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 the MS-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. To achieved the above requirements an MS architecture such as that shown in Figure 1.4 is required. Short-term operation Mid-term operation MS-CRE Session manager Management Plane interface Resource brokerage Status monitoring Profile and service-level information Management plane interface MSRB Service configuration traffic distribution Netwotk configuration RMS Management plane interface Mobile terminal Managed network (component of CR infrastructure) legacy element and network management systems User and control plane interface Figure 1.4 Architecture of the MS-CRE.

29 PROTOCOL BOOSTERS 7 Session manager MSRB RMS MS-CRE MS-CRE MS-CRE 1. Identification of new condition in service area 2. Extraction of status of Network and of SLAs 3a. Offer request 3c. Offer request 4a. Optimization request 3b. Offer 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: monitoring, service-level information and resource brokerage (MSRB); resource management strategies (RMS); session managers (SMs). The MSRB entity identifies the triggers (events) that should be handled by the MS-CRE and provides corresponding auxiliary (supporting) functionality. The RMS entity provides the necessary optimization functionality. The SM entity is in charge of interacting with the active subscribed users/terminals. The operation steps and cooperation of the RMS components are shown in Figures 1.5 and 1.6, respectively. In order to get an insight into the scope and range of possible reconfigurations, we review in Appendix A (please go to the network and protocol stack architectures [1 58] of the basic CRE components as indicated in Figure PROTOCOL BOOSTERS As pointed out in Figure 1.2, an element of the reconfiguration in 4G networks is protocol booster. A protocol booster is a software or hardware module that transparently improves protocol performance. The booster can reside anywhere in the network or end systems, and may operate independently (one-element booster), or in cooperation with other protocol

30 8 FUNDAMENTALS MSRB Service configuration traffic distribution Network configuration 1. Optimization request 2. Service configuration and traffic distribution: allocation to QoS and networks 3a. Request for checking the feasibility of solution 3c. Reply on feasibility of solution 3b. Computation of tentative network reconfiguration 5. Reply 4. Selection of best feasible solution 6. Solution acceptance phase 7. Network configuration Figure 1.6 Cooperation of the RMS components. Protocol messages Host X Booster A Booster B Host Y Booster messages Figure 1.7 Two-element booster. boosters (multielement booster). Protocol boosters provide an architectural alternative to existing protocol adaptation techniques, such as protocol conversion. A protocol booster is a supporting 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 booster may define new protocol messages to exchange among themselves, but these protocols are originated and terminated by protocol booster elements, and are not visible or meaningful external to the booster. Figure 1.7 shows the information flow in a generic two-element booster. 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 of protocol busters.