Research On Distributed Network Self-Management Model And Some Key Problems

With the fast development of network to large-scale and heterogeneity, the traditional network management is not working well for the new requirement. The distributed network management becomes an important research topic and development trend. Based on the consideration of distributed network characteristics and analyse of current distributed network management techniques, the research object of this thesis is to propose a distributed self-management solution for large-scale and heterogeneous network. A model is designed to apply the solution for providing the function of automation, autonomy and self-adaptation to network management. Based on the key aspects, architecture and mechanism, of network management, some original research as follow:(1) Distributed Network Self-Management ModelBecause network nodes in distributed network management are large, dispersed and diverse, traditional operator-intensive management mode requires much intervention from labors, who have to not only master the exact and deep knowledge of network resources and management rules, but also adjust management task based on the real situation of network. This operator-intensive network management mode results in low efficiency and high proneness to error. A network self-management model is presented to solve it. This model improves the automatic, autonomous and self-adapting management in two aspects: Automatic Initialization and Dynamic Self-management. In the process of Automatic Initialization, Management Locating Service (MLS) is presented to locate the proper managers efficiently in a multi-manager environment, as well as activate the initial management tasks automatically. In the process of Dynamic Self-management, the event-driven management task configuration mechanism is presented to extend management task on-the-fly to the managed node. In addition, the architecture based on hierarchical management station and distributed management task is proposed to well adapt the large-scale network.