计算机应用与软件2017，Vol.34Issue(4)：110-117,156,9.DOI:10.3969/j.issn.1000-386x.2017.04.020
P2P应用流量的高效分类方法研究
RESEARCH ON EFFICIENT CLASSIFICATION METHOD OF P2P APPLICATION TRAFFIC
摘要
Abstract
Network application has been showing a lot of categories because of the widespread use of Internet applications, especially P2P applications traffic growth.Traditional traffic classification and application identification methods can't considerable reach stable application classification precision.In order to improve classification accuracy and stability of P2P application traffic and manage network scientifically, this paper proposes WMFA (sliding window multi-flow association) algorithm.Using P2P application traffic statistics feature by reducing the flow statistics feature dimension and reducing the number of packets in each network traffic flow, C4.5 algorithm is used to classify P2P applications.We use WMFA algorithm to mine misrecognized flows, and carry on the multi-flow association on the second recognition to improve the P2P application traffic classification accuracy.Experimental results show that with a decrease in P2P traffic flow characteristics dimension and reducing the number of each flow data packets, WMFA algorithm average classification precision for the domestic mainstream P2P application is more than 96%, the average rate of accuracy than the existing method of 3%.关键词
P2P流量分类/应用识别/WMFA算法/多流关联Key words
P2P traffic classification/Application recognition/WMFA algorithm/Multi-flow association分类
信息技术与安全科学引用本文复制引用
陈金富,赵慧,常鹏,张永铮..P2P应用流量的高效分类方法研究[J].计算机应用与软件,2017,34(4):110-117,156,9.基金项目
国家自然科学基金项目(61572496) （61572496）
国家高技术研究发展计划基金项目(2013AA014703). （2013AA014703）
