一种基于流量行为分析的P2P流媒体识别方法  

A Kind of P2P Streaming Media Identification Method Based on Traffic Behavior Analysis

作　　者：宫婧[1] 孙知信[2] 陈二运 

机构地区：[1]南京邮电大学理学院,江苏南京210003 [2]南京邮电大学计算机技术研究所,江苏南京210003 [3]南京中兴软创公司,江苏南京210012

出　　处：《计算机技术与发展》2009年第9期128-131,214,共5页Computer Technology and Development

基　　金：国家自然科学基金(60572131);科技型中小企业创新基金(08C26213200495);江苏省科技攻关项目(BE2007058);江苏省高校自然科学基础研究项目(08KJB520005);南京邮电大学基金项目(NY206050)

摘　　要：P2P流媒体的识别问题已经成为P2P研究领域中的热点问题。以目前主流的P2P流媒体软件:PPLive、PPStream、QQLive为蓝本,进行流量行为特征分析,总结出相应的流量特征和特征字符串,提出一种基于流量行为分析的P2P流媒体识别方法。该方法的核心思想是:首先通过流量的统计算法将捕获到的流媒体数据包进行分类统计,再运用特征字匹配算法进行二次识别匹配,从而得出结果。随后,开发出基于流量行为分析的检测系统,试验结果证明该方法能够有效地识别P2P流媒体。How to identify P2P streaming media has become a hot issue in the field of P2P. Taking several most popular P2P streaming applications like PPLive,PPStream,and QQLive as the main research objects, analyzes their traffic behavior and generalized related traffic profiles and payload signatures. Then a novel P2P streaming media identification method based on traffic behavior analysis is proposed. The core of this method is two aspects. First, using traffic statistics algorithm,classify and statistics captured streaming media packets. Then, using features of characteristic matching algorithm,can identify and matching these packets,so r^ult can come out. Subsequently,a detection system adopting this traffic behavior analysis based method has been developed to evaluate it, and the experiment results show that the method proposed in this paper can identify P2P streaming media traffic effectively.

关 键 词：P2P 流媒体 行为分析 流量检测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]