一种基于多播推测丢包率的算法

网络层析是近年新兴的一个网络研究领域,它利用端到端的性能测试结果推导网络内部性能特征或拓扑结构,克服了传统网络测量技术的一些缺陷.丢包率层析的主要方法是利用最大似然估计(MLE),但是计算复杂度高且计算时间较长;基于伪似然估计(PMLE)方法可以较快估计各链路丢包率,但是在非叶节点链路的误差较大.为了克服以上缺点,本文基于多播网络的端对端测量,结合MLE和PMLE提出一种推算网络内部各链路的丢包率算法.通过仿真证实该算法估测的结果能真实地反应网络内部丢包趋势,在推测精度较好的情况下,计算量减少,计算复杂度降低.

An Algorithm of Network Loss Ratio Based on Multicast

Network tomography is an emerging network research area in recent years.They use end-to-end performance measurement result to derive network internal performance distribution or topology,thus it overcomes traditional network measurement's shortcomings.The arithmetic described in this paper is based on multicast network tomography end-to-end measurement,computing the loss rate.Through comparison and simulation,it is obtained that the loss inferences correctly show the network loss trend.