基于机器学习的P2P网络流问题的研究  

Research on P2P network flow problem based on machine learning

作　　者：邓小盾[1] 

机构地区：[1]西安外事学院,陕西西安710077

出　　处：《电子设计工程》2017年第8期109-114,共6页Electronic Design Engineering

基　　金：2015-2016年度高水平民办大学建设研究项目(15GJ044);2016年度西安市社会科学规划基金项目(16IN13)

摘　　要：P2P应用的出现和蓬勃发展使互联网流量组成发生显著变化,P2P流量己跃居成为互联网第一大流量,这给网络管理带来诸多问题,对流量监控提出更高要求。同时为逃避检测,P2P应用正朝着端口动态化、负载加密化的方向迅速发展。传统的流量识别技术己经难以有效识别出P2P流量,而基于机器学习的P2P流量识别技术不依赖端口和负载信息,因此,成为近年来的研究热点。针对基于机器学习的P2P流量识别的问题,采用查阅文献的方法,通过对国内外关于P2P网络流量识别的研究成果的研究,总结出流量识别的方法,并且分析P2P网络未来的发展趋势,有非常重要的理论意义。The emergence of P2P applications and the vigorous development of Internet traffic has undergone significant changes in the composition of, P2P traffic has leapt to become the first Internet traffic, which network management to bring many problems, put forward higher requirement for traffic monitoring. At the same time, in order to avoid detection, P2P applications are moving towards the port dynamic, the direction of the load encryption rapid development. Traditional traffic identification technology has been difficult to effectively identify the P2P traffic, and the P2P traffic identification technology based on machine learning does not rely on the port and load information, so it has become a hot research topic in recent years.This paper is supported by the National 863 project ＂high credibility network business control system, based on machine learning to identify P2P traffic problems, using the method of literature review, through the study of domestic and foreign research results on P2P network traffic identification, summed up the traffic identification method, and analyze the future development trend of the P2P network has very important theoretical significance.

关 键 词：P2P网络 流量识别 识别方法 未来趋势 

分 类 号：TN711[电子电信—电路与系统]