Research And Implementation Of Network Health Assessment And Fault Prediction

As the network involved in all aspects of our life constantly and problems of network security emerge in endlessly. It is an important problem that how to evaluate the safety conditions of network and predict the fault and make preventive strategies to make the network operate efficiently and stably. For this problem, this paper put forward a model which can do the network health assessment and fault prediction based on the state of present or past. It is a new technology to predict the fault and diagnose the system. Since this method analysis the network health assessment according to the original information, it is helpful to understand the security situation of network. What is more, fault prediction can forecast the fault to be coming and does something advance. For example, some works about the maintenance can improve the stability of the network and prevent network security to be affected by the hidden danger. It is sure that the method has important significance.This paper is based on the State Science and Technology Support Projects and Shaanxi Province International Cooperation Key Projects, and is combined with the latest research achievements and development trends in the field of network security. This paper will study on building the network health evaluation and fault prediction model and other aspects about it.This paper can be conducted by five aspects.Firstly, the paper expounds the developing situation in the field of network security in detail, and introduces the related theory of technology including monitoring products, health assessment, fault forecast technology and the significance of the research and analysis. Then, it embarks from the network monitoring system for the various aspects of information network to establish index system of evaluation and prediction such as selecting network real-time performance information, early warning information, log information, vulnerability information loopholes, and host running service level indicators like network health assessment and prediction index.After this, the paper will probe the network health assessment and fault prediction model building, and the model of system structure and the process in detail. Since the existing network health assessment method has the limitation on the overall assessment cannot be accurately understood by the network of local shortage, the new one will put forward a kind of network health assessment model based on directed graph, the two aspects of fusion node assessment and evaluation of network, network health assessment can be divided into healthy nodes, and the node is an importance weight calculation method and comprehensive health nodes and node importance network health model function is used to determine the health of the network.Then, the method will use real-time performance to monitor information by using fuzzy neural network algorithm, and achieve the classification of network fault prediction.Lastly, the health assessment and prediction model of simulation experiments have been carried out to verify the network health by using independent research and development achievements service security monitoring systems, network monitoring, network information combined with the project team. At the same time, the method will test and verify the possibility of fault simulation of network in the future using the fault prediction algorithm.