Research And Application Of Network Fault Intelligent Diagnosis Technology

With the rapid development of information technology, network fault diagnosis has become the core of the network fault management. The expansion of network scale and the increase of network heterogeneous degree are increasingly high demand for network security. How to achieve a rapid network fault diagnosis scheme with the minimum cost and the shortest time has become a research hotspot by network expert in today’s large-scale network environment.The thesis after that deep summary the present network fault intelligent diagnosis technology. Firstly, it introduces the dependency graph-base fault localization algorithm. Secondly, it considers the dependency graph-base fault localization algorithm will be invalid when the occurred fault in network has not dependency connection. According to this problem, the paper designed the dependency graph-base fault localization algorithm in what independent fault situation. Induct a concept of information cost to locate fault. Due to the dependency graph-base fault localization algorithm need to know failure probability of every object and the extent of interdependency between objects in advance. Therefore, it has big limitation in practice.Next, the paper based on further study the association graph and case base-base fault localization technology. It presents the taking time factor into association graph and case base-base fault localization algorithm to deal with associate graph can’t locate fault when it has occurred ring. The algorithm according to representative occur time of fault event of each vertex of association graph to locate fault. The earliest fault is likely to be the source of fault.The paper applied the taking time factor into association graph and case base-base fault localization algorithm to fault localization function of sub-network fault diagnosis system after the comprehensive comparison of the two algorithms. Although its principle is simple, it has high practical in reality.