Analysis Methods Of Topology Discovery And Implementation For Administration In Networks

As the advent of the information age, the network technology is developing rapidly. The new network products and network model emerge unceasingly; People's learning, working and social lives are contacted with network. It is extremely important to keep network smooth and reliable. A stable, secure, reliable, and supervisory network management system is required urgently. And network management is primarily achieved by monitoring network topology.The main purpose of network topology discovery is to access and maintenance information about network nodes and the existent links between their relationships and finally to map out the entire network topology map. Network manager can quickly location and repair network nodes under the topology.This paper analyses the problems about network management such as management of a wide range of heterogeneous network, lacking of abilities to diagnose network layer 2, information processing shortages and little methods to monitor status of network real-timely. It points out that a perfect solution for discovering the network topology is very important.Base on this perspective, the paper introduces SNMP network management model, and the basic knowledge structure of topology discovery, analysis the problems of current methods of topology discovery which are commonly used, such as discover slowly and not completely, especially it can not assure its discovery accuracy because of restrictions in switch table. A network topology discovery method based on the physical address table is discussed in details. The basic idea is that we gain integrated indirect connections between network switches thru some logic judge, filter out information of redundant links, reduce consumption of resources and build topology tree of the network at last.Several problems such as topology discovery in Multi-VLAN-Network and HUB-discovery are given the corresponding solution in the paper.Base on improved topology discovery strategy, we design a distributed network management system frame and implement a automatic topology discovery model. Many test cases have run and proved that the system can work well in a heterogeneous network. And the results of the algorithms for topology discovery are correct. Compared to similar software, the module can quickly and accurately builds network topology, achieves intelligent network management.