多源网络断层扫描技术在网络拓扑推断中的研究

Research of Multiple Source Network Tomography in Topology Inference

网络断层扫描技术不需要网络内部节点的配合，通过网络外部端到端的测量来推断内部特性。多源断层扫描技术可获得更多的网络内部链路信息，能够更精准的得到网络的拓扑结构。该文重点论述了多源网络断层扫描技术，将多源网络划为几种2-by-2结构，并通过一种根据hamIning距离的链路利用率算法得出了网络拓扑结构。通过仿真实验证明了该方法的精准性。

Network tomography can induce network internal characteristics based end-to-end measurement without the cooperation of nodes. Multiple source network tomography can obtain plenty information of the internal link performance, can get network topology accuracy. The paper introduces multiple source network tomography, and convert into 2-by-2 structures. The algorithm infer network topology using hamming distance. The simulation improve the inference accuracy.