摘要
P2P流量已经占据了整个网络流量的60%—70%,HTTP、EMAIL等传统的流量已经受到P2P流量的影响,同时伴随大量非授权内容的传播和安全问题。网络运营商、企业网和校园网为保证传统应用的性能,需要对P2P流量进行有效管理。要完成这一目标,首要工作是对P2P流的检测,把它和传统的流量区分出来。该文对P2P流检测主要方法进行了总结,将当前的检测方法分为基于报文层面、流层面和节点层面3大类,针对各类中具体的方法,分析了各自的优点和缺点,并进行了比较。提出将多种方法结合使用有效对P2P流进行检测。最后分析了P2P流检测技术进一步的研究方向。
P2P flows occupy 60% -70% of the entire network traffic with transfers of non-licensed content and security problems affecting HTTP, EMAIL, and other traditional applications. ISP and network operators need to manage P2P traffic to ensure the performance of traditional applications. To accomplish this goal, the system must first identify the P2P traffic. Current identification technology can be generally divided into three categories as packet level methods, flow level methods, and node level methods. This paper surveys the advantages and disadvantages of the current P2P flow identification methods. A variety of methods are combined to more effectively detect P2P flow. The survey also indicates new directions in P2P detection.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第4期616-620,共5页
Journal of Tsinghua University(Science and Technology)
基金
国家"八六三"高技术项目(2006AA01Z209)
关键词
对等网络技术(P2P)
网络流量分类
净荷
统计特征
流检测
peer-to-peer （P2P）
network traffic classification
payload
statistical characteristic
flow identification
