IPTV平台实时检测系统的设计与实现  

Design and Implementation of IPTV Platform Real-time Detection System

作　　者：何杰[1] 刘朝斌[2] 杨岳湘[1] 唐川[3] 

机构地区：[1]国防科学技术大学计算机学院,长沙410073 [2]第二军医大学信息中心,上海200433 [3]国防科学技术大学信息中心,长沙410073

出　　处：《计算机工程》2011年第6期256-259,共4页Computer Engineering

基　　金：国家"863"计划基金资助重点项目(2008AA01Z407);湖南省自然科学基金资助项目(07JJ5084);广东省科技计划基金资助项目(2009B080701031)

摘　　要：基于P2P的网络电视直播(IPTV)业务快速发展并带来诸多安全隐患,精确检测IPTV流使用的平台是对其实施有效监管的前提。为此,采用支持向量机方法,依据网络流量的流量特征和行为特征,选取PPlive、PPstream、QQlive、Sopcast和UUsee作为IPTV平台的代表,设计并实现一种IPTV平台实时检测系统。实验结果证明,该系统能够在5 min内检测出被管理节点是否在使用IPTV平台及平台类型,平均检测精度为90.62%。The rapid development of P2P based Internet Protocol Television（IPTV） has brought many potential security risks.The accurate detection of IPTV platform is the prerequisite for its effective supervision.In this paper,based on the flow features and behavior features of network traffic,by using support vector machine method,selecting PPlive,PPstream,QQlive,Sopcast and UUsee as representatives for IPTV platforms,it designs and implements an IPTV platform real-time detection system.Experiment results show the system can detect whether the managed node is using an IPTV platform,and it is with 90.62% correct rate within 5 minutes.

关 键 词：实时检测 P2P网络 IPTV业务 支持向量机 

分 类 号：N945[自然科学总论—系统科学]