Mobile Agent based Intelligent System for Autonomic Network Management Thesis Submitted to the Kumaun University, Nainital By Naveen Kumar Gondhi For the Award of Degree of Doctor of Philosophy in Computer Science Under the Supervision of Dr. Durgesh Pant Professor & Director School of Computer Science & IT Uttarakhand Open University, Haldwani Campus Dehradun, Uttarakhand
DECLARATION I declare that the present thesis submitted to the Kumaun University, Nainital for the degree of Doctor of Philosophy in Computer Science is original in its contents and has not been submitted before, either in parts or in full to any university for any research degree or diploma. (Naveen Kumar Gondhi)
CERTIFICATE This is to certify that the thesis entitled MOBILE AGENT BASED INTELLIGENT SYSTEM FOR AUTONOMIC NETWORK MANAGEMENT submitted to Kumaun University, Nainital for the degree of DOCTOR OF PHILOSOPHY IN COMPUTER SCIENCE is a record of the bonafide work carried out by Mr. Naveen Kumar Gondhi, under my guidance and supervision. It is further certified that the candidate has fulfilled the norms required for the submission of the thesis. (Dr. Durgesh Pant) Date: 31.12.2010 Professor & Director School of Computer Science & IT Uttarakhand Open University, Haldwani Campus Dehradun, Uttarakhand
I Acknowledgements Completing a task is never a one man effort It is often the result of valuable contribution of a number of Individuals in a direct or indirect manner That helps in shaping and achieving A GOAL I, take this opportunity to express my utmost indebtedness to this Great Institution- Kumaun University- Nainital which provided me as opportunity to fulfill the most cherished dream of reaching my covetous goal. It gives me an immense pleasure to express my sincere and wholehearted sense of gratitude to my esteemed Guide Dr. Durgesh Pant, Professor & Director, School of Computer Science & IT, Uttarakhand Open University, Haldwani, Campus Dehradun, Uttarakhand for his invaluable guidance and untiring supervision throughout my project. My most profound gratitude goes to my family which always supports me during the course of the study and I would like to express my sincere thanks and full appreciation to Shri Mata Vaishno Devi University and all those people who extended their cooperation, moral support assistance whenever and wherever need aroused. and rendering their Naveen Kumar Gondhi
II List of Tables Table 3.1 Comparison of MAs vs. RMI-based approaches in terms of response time as a function of the transferred data Table 3.2 Comparison of collected bytes from each host having varied payloads Table 3.3. Comparison of MAs vs. RMI-based approaches in terms of network overhead Table 3.4. Synchronous SNMP time measurements Table 3.5. Asynchronous SNMP time measurements Table 5.1 ITU-T FCAPS Model for Accounting & Performance Management Table 5.2: Calculation of Turn around Cost of Jobs Table 5.3: Load distributions on five servers using SI, ESI, SyI, RI Approach
III Fig. 1.1 Network Management Paradigms Fig. 2.1 Composition of Mobile Agent List of figures Fig. 2.2 Network Load Balancing using Mobile Agent Fig. 2.3 Asynchronous Operations of Mobile Agent Fig. 2.4 Intelligent System overview Fig. 2.5 Components of Knowledge based system Fig. 2.6 Neuro-fuzzy network Fig. 2.7 Generic Network Management System Fig. 2.8 Internet Management Model Fig. 2.9 Basic MIB Structure Fig. 3.1 Schematic Architecture of MAIS Fig. 3.2 Aglet Proxy / Context Layer Architecture Fig. 3.3 ATP layers Communication Methods Fig. 3.4 WebNMS SNMP API Architecture Fig. 3.5 MAs vs. RMI-based approaches comparison in terms of response time Fig. 3.6 Response time for multi-hop MAs as a function of amount of encapsulated information Fig. 3.7 Response time for multi-hop MAs as a function of network size Fig. 3.8 Graphical representation of Mobile Agents vs. RMI comparison in terms of network overhead Fig. 3.9 Response time as a function of the network size and the collected MIB objects per host for synchronous SNMP polling operations. Fig. 3.10 Response time as a function of the network size and the collected MIB objects per host for asynchronous SNMP polling operations Fig. 3.11 Multi-hop MAs response time measurements as a function of the network size and collected MIB objects per host with TCP protocol Fig. 4.1 Fault Management Process Fig. 4.2 Classification of fault isolation techniques. Fig. 4.3 Schema of intelligent mobile agent. Fig. 4.4 Schema Architecture of MAIS for automated fault management Fig. 5.1 Generic Three tier Accounting Infrastructure Fig. 5.2 Accounting Management Architecture
III Fig. 5.3 Performance Management Architecture Fig. 5.4 System architecture of EMAS in Cluster Computing Fig. 5.5 Information collection policy Fig. 5.6 Genetic Algorithm for Server Selection Fig. 5.7 EMAS System Throughput Fig. 5.8 Normalized average response time under varying load Fig. 5.9 Normalized average waiting time under varying load Fig. 5.10 Job Migration under varying load Fig. 5.11 System throughput of ESI approach and the case without load balancing. Fig. 6.1 RMF Schematic Architecture Fig. 6.2 Policy Execution Module Fig. 6.3 Average End to end delay for RMF, AODV, DSR Fig. 6.4 Normalized Routing Load for RMF, AODV, DSR Fig. 6.5 Packet Delivery Fraction for RMF, AODV, DSR
IV Preface An explosive growth in the number of computers and the need to share information has triggered the development of computer networks. Moreover, today s network applications demand a wide variety of efficient services which requires that the network must be operating efficiently at all times. The Network failures cause various losses such as efficiency and productivity. Thus, Network Management solutions have to be developed to minimize the network failure time and monetary losses to both the clients and network access provider. Network Management is a critical issue in today s rapidly changing network environment. Existing centralized client server based network management suffers from problems such as insufficient scalability, interoperability, reliability and flexibility, as networks become more geographically distributed. Considering the above issues, research has been done for using mobile agents efficiently for network management. Mobile Agents decentralize management tasks and distribute load over the network but they too suffers from the problem of limited Intelligence, objectives and non utilization on existing commercial networks. The present thesis proposes Mobile Agents based Intelligent System for Autonomic Network Management leveraging the benefits of client server paradigm and mobile agent by incorporating the expert system technology resulting into Intelligent Network Management solution, capable to initiate corrective actions for diagnosing of network faults on its own thereby, achieving a rapid response time, efficient resource utilization and consequently higher throughput. The proposed system explores the areas of Network management viz. Fault Management, Accounting Management, Performance Management and Resource Management. The main goal of our effort is to apply artificial intelligence approach to reduce the severity of the network management tasks by analyzing and capturing the statistics of the network along with its services, properly weigh them according to expert system intelligence and finally implements the solutions with the help of mobile agents leading to a leap towards high performance network management solution.
Mobile Agents based Intelligent System for Autonomic Network Management Contents Acknowledgments. I List of Figures.II List of Tables.III Preface. IV Chapter 1 Introduction 1.1. Network Management Paradigms 2 1.2. Background & Motivation...5 1.3. Research Objectives 6 1.4. Research Methodology 7 1.5. Organization of Dissertation 8 1.6. List of Publications..9 Chapter 2 Literature Survey 2.1 Computing Paradigms..12 2.1.1. Mobile Agent Definition, Composition & Characteristics 13 2.1.2. Mobile Agent based Applications.18 2.1.3. Mobile Agent Systems...20 2.2 Intelligent System.24 2.2.1. Knowledge Based Systems 25 2.2.1.1. Expert Systems 27 2.2.2. Computation Intelligence / Soft computing...28 2.2.2.1. Genetic Algorithm..29 2.2.2.2. Neural Network...30 2.2.2.3. Fuzzy Logic..31 2.2.2.4. Neuro fuzzy System.31 2.3 Network Management...32 2.3.1. ISO Network Management Model-FCAPS.33 2.3.2. Network Management Architecture 34 2.3.2.1. Managed Devices, Agents & Managers 35 2.3.3. Simple Network Management Protocol 36 2.3.3.1. MIB Modules & Object Identifiers 38 9
Mobile Agents based Intelligent System for Autonomic Network Management 2.3.3.2. SNMP versions 39 2.3.4. Current Trends in Network Management 41 2.3.5. References...43 Chapter 3 Mobile Agent Based Intelligent System for Autonomic Network Management 3.1 Introduction..48 3.2 Main Goals of MAIS Framework..50 3.3 System Architecture MAIS Framework..52 3.3.1. Platform for Mobile Agent Module 53 3.3.1.1. Mobile Agent Pool 54 3.3.1.2. Management Policies 56 3.3.1.3. Communications & Coordination Layers 57 3.3.2. Intelligent System Module..58 3.3.3. Network Manager Module...59 3.3.4. Implementation Tools.60. 3.3.4.1. IBM Aglets Layer Architecture..61 3.3.4.2. WebNMS SNMP API...63 3.3.4.3. Visual Prolog...66 3.4 Implementation & Performance Evaluation 68 3.4.1. Response Time 68 3.4.1.1. Comparison of MAs vs. RMI-based approaches..69 3.4.1.2. Response time for multi-hop MAs..70 3.4.2. Network Overhead...72 3.4.2.1. Comparison of MAs vs. RMI-based approaches..73 3.4.2.2. Response time for multi-hop MAs..74 3.4.3. SNMP time measurements...75 3.4.3.1. synchronous SNMP polling...76 3.4.3.2. Asynchronous SNMP polling..77 3.4.4. Multi-hop MAs time measurements.78 3.5 Conclusions.80. 3.6 References 82 10
Mobile Agents based Intelligent System for Autonomic Network Management Chapter 4 Fault Management 4.1 Introduction.85 4.2 Fault Management Process 86 4.2.1. Fault Detection..86 4.2.2. Fault Isolation 87 4.2.3. Fault correction..90 4.3 Automated Fault Management..91 4.4 Mobile Agent based System Architecture.93 4.4.1. Mobile Agent Pool. 94 4.4.2. Intelligent System Module.97 4.5 Implementation & Performance Evaluation..99 4.6 Conclusions...100 4.7 References..101 Chapter 5 Accounting & Performance Management 5.1 Introduction..103 5.2 Accounting Management 104 a. Accounting Management Architecture 106 b. Purposes of Accounting Management.107 5.3 Performance Management...108 a. Performance Management Architecture.109 b. Purposes of Performance Management...109 5.4 Relation Between Accounting & Performance Management.110 5.5 Accounting & Performance Management in Cluster Computing 112 5.6 Mobile Agent based Management in Cluster Computing 114 5.7 Evolutionary Mobile Agent System Architecture...118 a. Mobile Agent Pool...119 b. Management Policies.120 c. Server Selection Policies...122 5.8 Implementation & Performance Evaluation.124 5.9 References.128 11
Mobile Agents based Intelligent System for Autonomic Network Management Chapter 6 Resource Management. 6.1 Introduction 131 6.2 Resource Management..132 6.3 Resource Management Architecture..133 a. Agent Management System.135 b. Policy Execution Module 137 6.4 Implementation & Performance Evaluation in MANET.138 6.5 Conclusions.140 6.6 References 140 Chapter 7 Conclusions...143 Annexure A1 Bibliography..147 Annexure A2 Publications..163 12