Automatic Fault Detection And Elimination Of Network Management

Network fault management is very important to ensure the stable operation of the network resources. With the expansion of the scale of computer network, the requirement of network fault management is highly. How to discover network fault automatically and handle all kinds of faults in network resources equipment effectively have become the focus that the researchers need to pay attention to. Nowadays, automatic detection and recovery in exited network fault management model shortcomings still defects, especially in the network fault management model based on rules, the defects include two aspects, which are rules need to be manually defined and it lack the effective method of fault automatic recovery. The capacity of automation will be lower due to fault recovery needs a substantial amount of manual management. In order to overcome the problems of the rule automatic generation and fault recovery in the model, this paper proposes a novel network fault management model based on rules, the actual experiment verified the effectiveness of the new model.Processing capacity of rules is the key to evaluate the performance of model. In the existed model, rules definition always depends on manual definition. However, the rules definition by manual defined often contains a number of redundancies of fault diagnosis information. The redundant information will inevitably influence the efficiency of the rule matching. In the method of rules automatic generation which is proposed in this paper, we used neighborhood rough set to execute the attribute reduction for network fault diagnosis attributes, which can reduce redundancy of the information of network fault diagnosis. Experiments results show that the attribute reduction algorithm based on neighborhood rough set has good ability of network fault attribute reduction.Rules automatic generation needs a method which can convert processing information of network fault into the rules with processing ability. In the method of rules automatic generation, this paper proposes a structured semantic grammar to describe the classification and processing information of network fault. Experiments results show that the method has good capability of rules automatic generation and the generated rules can reduce rules matching processing time effectively meanwhile without reducing the fault diagnostic rate.The existed network management model lack the method of automatic recover long-distance fault, this paper proposes the mechanism of network fault long-distance automation recovery. In the rules automation generation, the generated recovery rules have defined a series of processing approaches of fault recovery. We realized an automaton based on rules, which use the method of long-distance delivery based on WMI to dispatch the automaton and rules to fault device, and it realizes the long-distance fault recovery. The proposed method has a good capacity of fault automation recovery.In conclusion, we applied the proposed network fault management model based on rules into the network environment, and then we verify the model is effective and analyze the experiment results. The experiment results show that the proposed model can improve efficiency and scalability of network fault management.