Autonomic Networks

As the Internet becomes larger and larger, and consequently more difficult to control and to manage, telecommunication operators, manufacturers and companies require tools to perform management and control tasks. A large number of tools coming from different areas have been proposed, but these are not sufficient to handle an evolving and dynamic environment. This book presents and explains all the techniques which integrate a certain level of intelligence (through intelligent software agents for example) in order to represent knowledge, take appropriate decisions, communicate with other entities and achieve a self-managing network.

Table of Contents:
Introduction xv Chapter 1. Artificial Intelligence and Monitoring of Telecommunications Networks1 Hassine MOUNGLA 1.1. Introduction 1 1.2. Network management goals 2 1.3. Monitoring needs of telecommunications networks 3 1.4. The telecommunications management network (TMN) 6 1.4.1. TMN management functions 6 1.4.2. TMN architecture 7 1.5. Control in telecommunications networks 7 1.6. Some AI techniques for monitoring telecommunications networks 9 1.6.1. Chronos: an expert system generator for monitoring telecommunications networks 9 1.6.2. Monitoring with model-based techniques 11 1.6.3. Agent technology 12 1.6.4. Example of agent-based telecommunications network monitoring architecture 14 1.6.5. Telecommunications network management with mobile agents 15 1.7. Conclusion 18 1.8. Bibliography 18 Chapter 2. Adaptive and Programmable Management of IP Quality of Service 23 Miguel CASTRO, Dominique GAÏTI, Abdallah M’HAMED and Djamal ZEGHLACHE 2.1. Introduction 23 2.2. Open and programmable network technology 24 2.3. Active and programmable QoS management over IP 25 2.3.1. Programmable modules 28 2.4. Architecture for adaptive and programmable management 31 2.4.1. Legacy mechanisms 33 2.4.2. MMB 33 2.4.3. MAPI 34 2.4.4. Management kernel 34 2.4.5. Core control 34 2.4.6. Hardware 35 2.5. CLAM: a new language for adaptive and programmable management 35 2.6. Related studies 36 2.6.1. Behavioral networks 36 2.6.2. Smart packets 36 2.6.3. SENCOMM 37 2.6.4. General evaluation 38 2.7. Case studies 39 2.7.1. Case study 1: web service optimization 39 2.7.2. Case study 2: maximization of a given objective function 45 2.7.3. Case study 3: adaptive control of equity 49 2.8. Conclusion and perspectives 57 2.9. Bibliography 58 Chapter 3. Software Agents for IP Management 61 Anneli LENICA 3.1. Introduction 61 3.2. IP networks and their management 62 3.2.1. IP networks 62 3.2.2. IP network evolution and associated problems 63 3.2.3. IP network management 65 3.3. The multi-agent paradigm 66 3.3.1. What is an agent? 66 3.3.2. When should MAS be used? 68 3.4. MAS for IP network management 71 3.4.1. MAS for specific network problems 71 3.4.2. Existing applications 72 3.5. Perspectives and conclusion 78 3.6. Bibliography 79 Chapter 4. The Use of Agents in Policy-based Management 83 Francine KRIEF 4.1. Introduction 83 4.2. Policy-based management 85 4.2.1. The policies 85 4.2.2. Information model 86 4.2.3. Architecture 87 4.2.4. COPS protocol 88 4.2.5. Advantages and challenges 89 4.2.6. The agents and their advantage in network management 90 4.3. Provisioning and service control 91 4.3.1. Dynamic QoS provisioning in wired networks 92 4.3.2. Dynamic QoS provisioning in wireless networks 95 4.3.3. Prediction layer 95 4.3.4. Adaptation layer 96 4.3.5. Monitoring layer 96 4.3.6. Mobile agents for policy-based QoS provisioning 97 4.3.7. Dynamic service provisioning for mobile users 98 4.3.8. Intelligent agents for dynamic security control 99 4.4. Agents and service contract negotiation 100 4.4.1. Service contract 100 4.4.2. An intelligent negotiation interface 101 4.4.3. Client-provider dynamic negotiation 104 4.4.4. Dynamic negotiation between providers 105 4.4.5. Dynamic services negotiation for mobile users 107 4.5. Management of emerging services 107 4.5.1. Emerging services 108 4.5.2. Dynamic management of emerging services 109 4.5.3. Dynamic management of group multimedia services 110 4.6. Conclusion 111 4.7. Bibliography 112 Chapter 5. Multi-agent Platforms 117 Zeina EL FERKH JRAD 5.1. Introduction 117 5.2. Towards a standardization of multi-agent technology 118 5.2.1. FIPA model 118 5.2.2. KAOS model 121 5.2.3. General Magic model 122 5.3. Characteristics of a multi-agent platform 122 5.3.1. Methodological requirements for a multi-agent simulation platform 123 5.3.2. Other forms of requirements for an agent platform 124 5.4. Multi-agent platform evaluation 125 5.5. Examples of MAS platforms 127 5.5.1. Platforms for simulation 127 5.5.2. Implementation platforms 131 5.5.3. Mobility platforms 138 5.6. Conclusion 139 5.7. Bibliography 140 Chapter 6. Behavioral Modeling and Multi-agent Simulation 143 Leila MERGHEM-BOULAHIA 6.1. Introduction 143 6.2. Traditional network modeling and simulation approaches 144 6.2.1. Queuing theory 145 6.2.2. Modeling by Petri nets 145 6.2.3. Modeling by process algebra 145 6.2.4. Limits 146 6.3. Multi-agent modeling and simulation 147 6.3.1. Multi-agent simulation steps 147 6.3.2. Contributions 148 6.4. Behavioral modeling 149 6.4.1. Principle 149 6.4.2. Contributions 150 6.5. Two-level behavioral model of a network node 151 6.5.1. Introduction 151 6.5.2. Role of the two behavioral levels 153 6.5.3. Agents 154 6.5.4. Model of two behavioral levels. 154 6.5.5. Ensuring adaptability 156 6.6. Perspectives and conclusion 158 6.7. Bibliography 159 Chapter 7. Behavioral Modeling and Simulation: An Example in Telecommunications Networks 163 Leila MERGHEM-BOULAHIA 7.1. Introduction 163 7.2. Basic behaviors adapted to networks 164 7.2.1. Queue management basic behaviors 164 7.2.2. Scheduling basic behaviors 167 7.2.3. Routing basic behaviors 168 7.3. Metabehaviors 169 7.3.1. Queue management metabehavior 169 7.3.2. Scheduling metabehavior 170 7.3.3. Routing metabehavior 171 7.4. Simulation components and parameters 171 7.4.1. Objects 171 7.4.2. Agents 172 7.4.3. Parameters 173 7.5. A few results 174 7.5.1. Impact of queue management basic behaviors 174 7.5.2. Impact of scheduling basic behaviors 176 7.5.3. Impact of queue management metabehavior rules 178 7.5.4. Impact of scheduling metabehavior rules 179 7.6. Discussion 179 7.7. Conclusion and perspectives 181 7.8. Bibliography 182 Chapter 8. Multi-agent System in a DiffServ Network: Behavioral Models and Platform 185 Nada MESKAOUI 8.1. Introduction 185 8.2. Quality of service – existing solutions and their problems 186 8.2.1. RTP/RTCP 186 8.2.2. IntServ/RSVP 187 8.2.3. DiffServ 187 8.3. Agents, multi-agent systems and architectures 188 8.3.1. Agents 188 8.3.2. MAS 190 8.4. Towards intelligent and cooperative telecommunications networks 191 8.4.1. Node structure 192 8.4.2. Agent components 193 8.4.3. Agent behavioral model 194 8.5. Simulation – platform, topology and results 200 8.5.1. Platform 200 8.5.2. Topology and configuration 201 8.5.3. Simulation results 203 8.6. Conclusion 209 8.7. Bibliography 209 Chapter 9. Intelligent Agent Control Simulation in a Telecommunications Network 213 Hugues LECARPENTIER 9.1. Introduction 213 9.2. Network management and control by intelligent software agents 215 9.2.1. Agent-based admission control 215 9.2.2. Project Tele-MACS 215 9.2.3. Project Hybrid 215 9.2.4. Route selection by mobile agents 216 9.2.5. Cooperative mobile agents for network mapping 216 9.2.6. Project MAGNA 216 9.3. Simulating the behavior of intelligent agents in a communication network 217 9.3.1. Simulation of behavioral quality of service network control 217 9.3.2. Intelligent control simulation of a DiffServ network 217 9.3.3. Comparison and choice of a platform 218 9.4. Detailed simulator presentation 218 9.4.1. Structure of an INET node 219 9.5. Software agent architecture 224 9.5.1. Events monitor 226 9.5.2. Cleaner 227 9.5.3. Message interface 227 9.5.4. Task interface 229 9.5.5. Manager 229 9.6. Illustration 229 9.6.1. Quality of service control for voice over IP 229 9.6.2. Presentation of agents and routers used 230 9.7. Conclusion 231 9.8. Bibliography 231 Chapter 10. Agents and 3rd and 4th Mobile Generations 233 Badr BENMAMMAR 10.1. Introduction 233 10.2. Agent technology 234 10.2.1. Definition of an agent 234 10.3. Introduction to UMTS 238 10.3.1. VHE 239 10.3.2. Application of agents in UMTS 241 10.4. Introduction to WLAN 253 10.4.1. Application of agents in wireless networks 254 10.4.2. Problems related to the application of MAS in wireless environments 256 10.5. 4th generation mobile network 256 10.5.1. Definition of 4th generation 256 10.5.2. User expectations for mobile 4G networks 257 10.5.3. Technical conditions to achieve 4th mobile generation 258 10.5.4. Application of agents in 4G mobile networks 258 10.6. Conclusion 263 10.7. Bibliography 264 Chapter 11. Learning Techniques in a Mobile Network 267 Sidi-Mohammed SENOUCI 11.1. Introduction 267 11.2. Learning 269 11.2.1. Unsupervised learning 269 11.2.2. Supervised learning 269 11.2.3. Reinforcement learning 270 11.3. Call admission control 275 11.3.1. Problem formulation 275 11.3.2. Implementation of algorithm 276 11.3.3. Experimental results 278 11.4. Dynamic resource allocation 280 11.4.1. Problem formulation 281 11.4.2. Algorithm implementation 282 11.4.3. Experimental results 283 11.5. Conclusion 284 11.6. Bibliography 286 Chapter 12. An Experimental Example of Active Networks: The Amarrage Project 289 Nadjib ACHIR, Yacine GHAMRI-DOUDANE and Mauro FONSECA 12.1. Introduction 289 12.2. Description of the Amarrage project 291 12.2.1. Objectives 291 12.2.2. Contributions 292 12.3. Active networks: active architecture example for the control and management of DiffServ networks 296 12.3.1. DiffServ 298 12.3.2. Policy-based control 300 12.3.3. Description of architecture components 302 12.3.4. Capsule filtering at the level of data plan 305 12.3.5. Active router resource monitoring 305 12.3.6. Definition of QoS policies 306 12.3.7. Definition and deployment of TCB 307 12.3.8. Sensor deployment 309 12.3.9. Implementation of DACA architecture 310 12.3.10. Evaluation of DACA architecture behavior 312 12.4. Conclusion 315 12.5. Bibliography 315 List of Authors 317 Index 319

About the Author :
Dominique Gaïti is a Professor at the University of Technology, Troyes, France.