一种复合的自治域级拓扑发现方法

摘要

With the development of network technology,traditional network seems powerless when facing new challenges. Future network cames out at this time led by software defined network(SDN)with many advantages,but topology measure-ment is as important as it was in traditional environment.In an AS-level topology,we can simply take every single AS as a node,and edges between two nodes means adjacency of ASes.Large numbers of different topology discovery algorithms have been shown recently in various researches.In this paper,a simple but efficient method utilizing high speed collector is deployed in the network to collect border gateway protocol(BGP)routing tables on BGP routers and BGP update messages among them to deduce network topological structure is proposed,and it could also differentiate transit-node and stub-node in the network.Experiments are conducted to prove this method could achieve expected effect which is precisely and com-pletely inferring the topological structure at AS-level.%随着网络技术的发展，面对新的挑战，传统网络逐渐力不从心，软件定义网络（software-defined network， SDN）领衔的未来网络应运而生，随之而来的是各类网络测量技术纷纷针对未来网络发生演变，但拓扑结构测量在传统网络环境下的作用仍然不可忽视。在自治域级的网络拓扑中，每个自治域都可以简化为一个点，而用两点之间的连线表示自治域间的邻接关系。近年来有许多相关的研究展示了不同的拓扑发现算法。提出了一种简单高效的方法来推断自治域级的拓扑，利用在网络中部署高速采集器采集边界网关协议（border gateway protocol，BGP）路由器上的路由表以及BGP协议的更新信息来推断网络拓扑结构，并判定自治域的相关属性。实验证明了该方法能够达到预期效果，全面、准确地推断网络在自治域级的拓扑结构。