基于加性指标的网络断层扫描的研究  

Study of Network Tomography Based on Additive Index

作　　者：吴辰文 谢云燕[1] 

机构地区：[1]兰州交通大学

出　　处：《哈尔滨师范大学自然科学学报》2015年第1期67-72,共6页Natural Science Journal of Harbin Normal University

基　　金：甘肃省自然科学基金资助项目(1308RJZA111);兰州市科技计划基金项目(2009-1-5)

摘　　要：为了提高网络拓扑推断的准确率,在加性指标的网络断层扫描基础上,改进了原有推断二叉树的邻接算法,同时有进一步改进后可以推断一般树的拓扑结构.通过端到端的方法进行测量,运用加性指标进行求得终端节点的距离,最后将计算得到的距离作为改进算法的输入,推断出网络拓扑结构.并充分进行仿真实验证明算法的正确性以及比现有算法推断的网络拓扑准确率有所提高.To improve the efficiency of the network tomography inference based on the network tomography technology of additive metrics,the NJ algorithm previously inferred the binary tree topology is improved,and the further improved NJ algorithm can infer a general tree. Measured by the end-to-end method,the use of the terminal node distance measure is additive,finally the calculated distance acts the improved algorithm input to infer the network topology. The simulation experiments show that the algorithm is correct and the network topology inference is more accuracy than existing algorithms significantly.

关 键 词：性能的估计 邻接法 网络断层扫描 拓扑结构 逻辑 

分 类 号：TP393.02[自动化与计算机技术—计算机应用技术]