Research And Implementation Of Network Topology Discovery

With the increasingly wide range of Internet applications, the scale of Internet is growing rapidly, and structure of Internet is becoming more complicated.Accurate description of Internet Topology is very important for management and planning of Internet and the solution of Internet problems.At the same time,the technical problems in Internet topology discovery are different from the technical problems in traditional network topology discovery.The ways that traditional network topology discovery uses cannot be used in Internet topology discovery because of the Internet's heterogeneous, dynamic nature and the large size of Internet.This article describes the Network Topology Discovery Technology Research, and classifies the research depending on whether it has the permissions of managing the network.For the topology discovery within the management of domain, many ways are introduced, such as ways based on SNMP protocol, ICMP protocol and so on and the advantages and disadvantages of them are analyzed.For the topology discovery out of the management of domain, many faced problems are analyzed, such as the alias resolution problems, detecting redundancy issues, and the DoubleTree algorithm is focused on.At the same time, the Autonomous system level topology discovery techniques is described, and analyze the characteristics of Autonomous system level topology discovery and the research status.Based on the existing solutions, an improved distributed topology discovery programs TopoFinder is proposed.TopoFinder makes many strategies for the problems that a distributed topology discovery program encounts, such as the detecting source selection strategy, target address selection strategy, improved combination of multi-packet detection, alias resolve method, alias filtering methods and so on.Then according to the proposed distributed topology discovery program TopoFinder, a distributed topology discovery system is designed and implemented, and the system design and architecture are described.In addition, based on the existing solutions, a large-scale network critical node discovery technique is proposed. According to the autonomous system level topology discovery, the key nodes of large-scale network are analyzed, such as the border routers, in order to provide the basis for network planning.