Study And Implement On The Design Of Distributed Network Fault Management System DFMS

With the rapid expansion of the network application, the operation and maintenance of network are getting more and more complex. As a result, the role of network management becomes more important and its contents expand greatly. As a basal component of the network management, the fault management helps to ensure the the safety and stabilization of the network.This paper presents my work and conclusion on the project of system management and network management. My research focuses on the framework and function of the sub-system of network fault management. In this paper, a distributed network fault management system DFMS is presented which is based on policy enforcement. The architecture design and several efficient ways to get fault imformation are concretely introduced. In the process of events getting deployment, I raise a self-adjusting MIB algorithm on MIB limit setting other than the usual ways. This algorithm can adjust the limit dynamicly and automatically. After that, the primary Event Correlation and network fault diagnosis are analysed briefly. Finally, the concrete realization of the architecture and each functional modules are discussed, and then, the peculiarity and performance of the system is tested by examples of application.DFMS system adopts the framework with four levels of MOL(Managed Object Layer), EML(Element Management Layer), NML(Network Management Layer) and SML(System Management Layer). The collection of fault information includes four ways of ICMP packets parsing, Trap notification reports, Ping polling and MIB limit overstepping warning. The communication mechanisms can be either RMI(Remote Method Invocation) or SOAP(Simple Object Access Protocol) according to different network environments. As for the further process of fault diagnosis, the means of Bayesian network is an advisable way.Utilizing several new techniques, the result of this study can provide theoretical and practical supports for network fault management system under large-scale network environment, and the current work will contribute to the further study of network fault analysis, fault diagnosis and fault rectification.