Autonomic Networks

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

• Introduction xvChapter 1. Artificial Intelligence and Monitoring of Telecommunications Networks1Hassine MOUNGLA1.1. Introduction 11.2. Network management goals 21.3. Monitoring needs of telecommunications networks 31.4. The telecommunications management network (TMN) 61.4.1. TMN management functions 61.4.2. TMN architecture 71.5. Control in telecommunications networks 71.6. Some AI techniques for monitoring telecommunications networks 91.6.1. Chronos: an expert system generator for monitoring telecommunications networks 91.6.2. Monitoring with model-based techniques 111.6.3. Agent technology 121.6.4. Example of agent-based telecommunications network monitoring architecture 141.6.5. Telecommunications network management with mobile agents 151.7. Conclusion 181.8. Bibliography 18Chapter 2. Adaptive and Programmable Management of IP Quality of Service 23Miguel CASTRO, Dominique GAÏTI, Abdallah M’HAMED and Djamal ZEGHLACHE2.1. Introduction 232.2. Open and programmable network technology 242.3. Active and programmable QoS management over IP 252.3.1. Programmable modules 282.4. Architecture for adaptive and programmable management 312.4.1. Legacy mechanisms 332.4.2. MMB 332.4.3. MAPI 342.4.4. Management kernel 342.4.5. Core control 342.4.6. Hardware 352.5. CLAM: a new language for adaptive and programmable management 352.6. Related studies 362.6.1. Behavioral networks 362.6.2. Smart packets 362.6.3. SENCOMM 372.6.4. General evaluation 382.7. Case studies 392.7.1. Case study 1: web service optimization 392.7.2. Case study 2: maximization of a given objective function 452.7.3. Case study 3: adaptive control of equity 492.8. Conclusion and perspectives 572.9. Bibliography 58Chapter 3. Software Agents for IP Management 61Anneli LENICA3.1. Introduction 613.2. IP networks and their management 623.2.1. IP networks 623.2.2. IP network evolution and associated problems 633.2.3. IP network management 653.3. The multi-agent paradigm 663.3.1. What is an agent? 663.3.2. When should MAS be used? 683.4. MAS for IP network management 713.4.1. MAS for specific network problems 713.4.2. Existing applications 723.5. Perspectives and conclusion 783.6. Bibliography 79Chapter 4. The Use of Agents in Policy-based Management 83Francine KRIEF4.1. Introduction 834.2. Policy-based management 854.2.1. The policies 854.2.2. Information model 864.2.3. Architecture 874.2.4. COPS protocol 884.2.5. Advantages and challenges 894.2.6. The agents and their advantage in network management 904.3. Provisioning and service control 914.3.1. Dynamic QoS provisioning in wired networks 924.3.2. Dynamic QoS provisioning in wireless networks 954.3.3. Prediction layer 954.3.4. Adaptation layer 964.3.5. Monitoring layer 964.3.6. Mobile agents for policy-based QoS provisioning 974.3.7. Dynamic service provisioning for mobile users 984.3.8. Intelligent agents for dynamic security control 994.4. Agents and service contract negotiation 1004.4.1. Service contract 1004.4.2. An intelligent negotiation interface 1014.4.3. Client-provider dynamic negotiation 1044.4.4. Dynamic negotiation between providers 1054.4.5. Dynamic services negotiation for mobile users 1074.5. Management of emerging services 1074.5.1. Emerging services 1084.5.2. Dynamic management of emerging services 1094.5.3. Dynamic management of group multimedia services 1104.6. Conclusion 1114.7. Bibliography 112Chapter 5. Multi-agent Platforms 117Zeina EL FERKH JRAD5.1. Introduction 1175.2. Towards a standardization of multi-agent technology 1185.2.1. FIPA model 1185.2.2. KAOS model 1215.2.3. General Magic model 1225.3. Characteristics of a multi-agent platform 1225.3.1. Methodological requirements for a multi-agent simulation platform 1235.3.2. Other forms of requirements for an agent platform 1245.4. Multi-agent platform evaluation 1255.5. Examples of MAS platforms 1275.5.1. Platforms for simulation 1275.5.2. Implementation platforms 1315.5.3. Mobility platforms 1385.6. Conclusion 1395.7. Bibliography 140Chapter 6. Behavioral Modeling and Multi-agent Simulation 143Leila MERGHEM-BOULAHIA6.1. Introduction 1436.2. Traditional network modeling and simulation approaches 1446.2.1. Queuing theory 1456.2.2. Modeling by Petri nets 1456.2.3. Modeling by process algebra 1456.2.4. Limits 1466.3. Multi-agent modeling and simulation 1476.3.1. Multi-agent simulation steps 1476.3.2. Contributions 1486.4. Behavioral modeling 1496.4.1. Principle 1496.4.2. Contributions 1506.5. Two-level behavioral model of a network node 1516.5.1. Introduction 1516.5.2. Role of the two behavioral levels 1536.5.3. Agents 1546.5.4. Model of two behavioral levels. 1546.5.5. Ensuring adaptability 1566.6. Perspectives and conclusion 1586.7. Bibliography 159Chapter 7. Behavioral Modeling and Simulation: An Example in Telecommunications Networks 163Leila MERGHEM-BOULAHIA7.1. Introduction 1637.2. Basic behaviors adapted to networks 1647.2.1. Queue management basic behaviors 1647.2.2. Scheduling basic behaviors 1677.2.3. Routing basic behaviors 1687.3. Metabehaviors 1697.3.1. Queue management metabehavior 1697.3.2. Scheduling metabehavior 1707.3.3. Routing metabehavior 1717.4. Simulation components and parameters 1717.4.1. Objects 1717.4.2. Agents 1727.4.3. Parameters 1737.5. A few results 1747.5.1. Impact of queue management basic behaviors 1747.5.2. Impact of scheduling basic behaviors 1767.5.3. Impact of queue management metabehavior rules 1787.5.4. Impact of scheduling metabehavior rules 1797.6. Discussion 1797.7. Conclusion and perspectives 1817.8. Bibliography 182Chapter 8. Multi-agent System in a DiffServ Network: Behavioral Models and Platform 185Nada MESKAOUI8.1. Introduction 1858.2. Quality of service – existing solutions and their problems 1868.2.1. RTP/RTCP 1868.2.2. IntServ/RSVP 1878.2.3. DiffServ 1878.3. Agents, multi-agent systems and architectures 1888.3.1. Agents 1888.3.2. MAS 1908.4. Towards intelligent and cooperative telecommunications networks 1918.4.1. Node structure 1928.4.2. Agent components 1938.4.3. Agent behavioral model 1948.5. Simulation – platform, topology and results 2008.5.1. Platform 2008.5.2. Topology and configuration 2018.5.3. Simulation results 2038.6. Conclusion 2098.7. Bibliography 209Chapter 9. Intelligent Agent Control Simulation in a Telecommunications Network 213Hugues LECARPENTIER9.1. Introduction 2139.2. Network management and control by intelligent software agents 2159.2.1. Agent-based admission control 2159.2.2. Project Tele-MACS 2159.2.3. Project Hybrid 2159.2.4. Route selection by mobile agents 2169.2.5. Cooperative mobile agents for network mapping 2169.2.6. Project MAGNA 2169.3. Simulating the behavior of intelligent agents in a communication network 2179.3.1. Simulation of behavioral quality of service network control 2179.3.2. Intelligent control simulation of a DiffServ network 2179.3.3. Comparison and choice of a platform 2189.4. Detailed simulator presentation 2189.4.1. Structure of an INET node 2199.5. Software agent architecture 2249.5.1. Events monitor 2269.5.2. Cleaner 2279.5.3. Message interface 2279.5.4. Task interface 2299.5.5. Manager 2299.6. Illustration 2299.6.1. Quality of service control for voice over IP 2299.6.2. Presentation of agents and routers used 2309.7. Conclusion 2319.8. Bibliography 231Chapter 10. Agents and 3rd and 4th Mobile Generations 233Badr BENMAMMAR10.1. Introduction 23310.2. Agent technology 23410.2.1. Definition of an agent 23410.3. Introduction to UMTS 23810.3.1. VHE 23910.3.2. Application of agents in UMTS 24110.4. Introduction to WLAN 25310.4.1. Application of agents in wireless networks 25410.4.2. Problems related to the application of MAS in wireless environments 25610.5. 4th generation mobile network 25610.5.1. Definition of 4th generation 25610.5.2. User expectations for mobile 4G networks 25710.5.3. Technical conditions to achieve 4th mobile generation 25810.5.4. Application of agents in 4G mobile networks 25810.6. Conclusion 26310.7. Bibliography 264Chapter 11. Learning Techniques in a Mobile Network 267Sidi-Mohammed SENOUCI11.1. Introduction 26711.2. Learning 26911.2.1. Unsupervised learning 26911.2.2. Supervised learning 26911.2.3. Reinforcement learning 27011.3. Call admission control 27511.3.1. Problem formulation 27511.3.2. Implementation of algorithm 27611.3.3. Experimental results 27811.4. Dynamic resource allocation 28011.4.1. Problem formulation 28111.4.2. Algorithm implementation 28211.4.3. Experimental results 28311.5. Conclusion 28411.6. Bibliography 286Chapter 12. An Experimental Example of Active Networks: The Amarrage Project 289Nadjib ACHIR, Yacine GHAMRI-DOUDANE and Mauro FONSECA12.1. Introduction 28912.2. Description of the Amarrage project 29112.2.1. Objectives 29112.2.2. Contributions 29212.3. Active networks: active architecture example for the control and management of DiffServ networks 29612.3.1. DiffServ 29812.3.2. Policy-based control 30012.3.3. Description of architecture components 30212.3.4. Capsule filtering at the level of data plan 30512.3.5. Active router resource monitoring 30512.3.6. Definition of QoS policies 30612.3.7. Definition and deployment of TCB 30712.3.8. Sensor deployment 30912.3.9. Implementation of DACA architecture 31012.3.10. Evaluation of DACA architecture behavior 31212.4. Conclusion 31512.5. Bibliography 315List of Authors 317Index 319