The Research On Network Topology Discovery Of Decreasing Redundancy And Implementation Of Network Management System Topology Module

With the increasingly development of network, the network structure is becoming more complicated and network management becomes the critical element of maintaining the normal state of network. And network topology discovery is not only the basis of network management, but also the precondition of analyzing network performance and traffic. More and more people are focusing on discovering network topology completely, accurately and efficiently.In this thesis, some network topology discovery methods are discussed, and a Doubltree algorithm used for decreasing probe redundancy is analyzed and improved. Meanwhile, a topology discovery module based on SNMP of a network element monitoring subsystem is designed and implemented. The following are main tasks:1.A network topology method based on improved DoubleTree is proposed in this paper. DoubleTree, an important method for network topology discovery of IP level, can decrease the probe redundancy effectively, however, it also brings about numerous communication traffic resulting from exchanging global stop set between probes, increasing the network burden. To solve this problem, every destination address is assigned a BloomFilter and a update flag, and only updated BloomFilter need to be exchanged between probes, therefore reducing the communication traffic. The experimental result shows that the improved methodcan reduce 50% probe redundancy and decrease the network traffic among probe sources, while covering 95% network nodes and links.2.The topology discovery module based on SNMP of a network element monitoring subsystem is designed and implemented. The module mainly includes Communication based on SNMP, DataBase Store, topology discovery and visualization of topology layout. The main task of this module is to construct network topology map via routing information of routes and spanning tree information of switches, and layout the network map in order to visualize it. At present, the system is used in the real network.