Research On Immunity Agents For Network Fault Diagnosis

With the continuous development of computer network technology, expanding of network scale and increasing of network applications, the existing network fault management system with single function, complex operations, and Inefficiency was becoming more and more difficult to adapt for users. How to ensure network running rapidly, safely and steadily has become the focus of attention of researchers. However, there are so many reasons in network failures, such as complex structure and a huge amount of information, which make the existing network troubleshooting tools and related technologies difficult to resolve all network failures. Thus, applicating intelligent technology to network fault diagnosis has become an inevitable trend in the field.This thesis proposes a Network Fault Diagnosis Model using combination of immune algorithm and agent technology, based on analysis of biological immunity mechanism in-depth and study of agent technology with highly adaptive, self-control, self-learning for the above characteristics of network and its failure presently. In the model, there are different agents that each agent has its own division, that is, data collection agent, diagnostic agent, communication agent, decision-making agent, knowledge-managing agent and system-managing agent.The main features of this model as following:firstly, Using the artificial immune system, information processing mechanism for self-learning, self-organizing, adaptive and so it can adapt to a complex network environment; secondly, in a distributed environment, in accordance with the different characteristics of the network and different environmental, different agents with different functions and categories diagnose the faults, while the agent to collaborate among each other, exchange information, and collect data,in order to diagnose the type of fault, Greatly improving the efficiency of real time network fault diagnosis.In this thesis, the agent communication mechanism was researched, and the immune agent system architecture and immune agent model were designed. On the mean while, the concept of vaccine introduced to immune algorithm is applied to the center of antibody base of the model, in order to diagnosis the faults of interface group of network. Through simulation experiments, the model can diagnosis the network interface group fault type effectively. At the same time through studying the unknown fault type, the effectiveness of the algorithm was further proved.