基于社区挖掘的网络业务监控方法  

Research on network traffic monitoring based on community minging

作　　者：王永程 褚衍杰 

机构地区：[1]盲信号处理重点实验室,四川成都610041

出　　处：《重庆邮电大学学报（自然科学版）》2016年第3期426-430,共5页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金(61372076)~~

摘　　要：网络业务监控通常应用于大型企业级网络监控,通过分析企业网中承载的业务数据,实现对网络中业务类型及不同业务对应的终端分布情况的监测。提出了一种基于社区挖掘的网络业务监控方法,该方法的输入为企业网中IP通联数据,通过构建IP通联图,并基于IP通联图进行社区挖掘,输出IP通联社区,每个社区代表一种业务类型,社区内节点代表相应的业务终端。通过对某大型跨国企业网络数据的实证分析,发现与传统业务监控方法相比,该方法不仅能够有效发现各业务网络,实时监控业务网络状态,且能对网络中出现的新业务进行预警。Network traffic monitoring is usually applied to large enterprise network in monitoring traffic types and correponding terminals through the analysis of traffic data in network. We put forward a network traffic monitoring method based on community mining, with the IP communication data as input. This method constructs IP communication graph to mining IP community output IP communities. Each community represents a traffic type where nodes represent the corresponding service terminals. Through empirical analysis of the network data of a large multinational enterprise, the method can not only effectively discover the network traffic types, monitoring the network state in real time, but also give early warnings for new emerging traffic type.

关 键 词：网络监控 社区挖掘 业务网络 深度包检测(DPI)监测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]