非协作网络的链路性能状态推断技术研究综述  

Review of Link Performance State Inference Technologies for Non-Collaborative Networks

作　　者：张志勇 吴佑珂 刘方 徐锐 ZHANG Zhi-yong;WU You-ke;LIU Fang;XU Rui(Cyberspace Security Key laboratory of Sichuan Province,Chengdu Sichuan 610041,China;China Cyber Security,Chengdu Sichuan 610041,China)

机构地区：[1]网络空间安全四川省重点实验室,四川成都610041 [2]中国电子科技网络信息安全有限公司,四川成都610041

出　　处：《通信技术》2019年第2期259-268,共10页Communications Technology

基　　金：四川省应用基础研究项目(No.2018JY0379);国家科技部重点研发计划(No.2016YFB0801302)~~

摘　　要：网络层析成像技术通过端到端测量获得路径性能参数,采用统计推断方法推测目标网络的内部链路性能状态,可以在无需对目标网络内部元素访问控制的前提下实现对目标网络运行态势的准确感知,适用于非协作网络环境下识别目标网络的脆弱区域。对网络链路性能状态参数的层析成像技术进行分析,从链路时延、丢包率和拥塞状态等方面介绍现有研究,从性能参数模型、参数推断方法等角度分析相关方法,并对比其优缺点,以加强对相关领域的认识。In network tomography,path performance parameters are obtained by end-to-end measurements.At the same time,the statistical inference method is used to infer the internal link performance state of the target network.In this way,accurate sensing of the target network operating situation can be realized without requiring access control to the internal elements of the target network.This method is suitable for identifying vulnerable areas of the target network in a non-cooperative network environment.The tomographic techniques of network link performance status parameters are analyzed;the existing explorations introduced from the aspects of link delay,packet loss rate and congestion status;the related methods analyzed from theangles of performance parameter model and parameter inference method,and their advantages and disadvantages are compared,thus to enhance the awareness of relevant fields.

关 键 词：网络测量 网络层析成像 链路性能参数 统计推断 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]