摘要
随着P2P技术的发展,很多P2P软件采用了新的技术改善局域网用户的传输质量,导致P2P数据传输中出现了新的特征。针对STUN协议,在多元分类的基础上,提出了一种基于会话的P2P流量识别算法。首先介绍了STUN协议建立会话的消息格式,并描述了几种不同类型NAT建立地址映射的过程。算法采用多元组描述会话,建立二叉树结构对收发的数据包依次进行分析,识别P2P会话,最后以哈希表存储会话状态。在Linux环境下,以迅雷为例进行实验,结果表明采用本算法可以很好地识别P2P数据流。
With the development of P2P technology, a lot of P2P applications introduced new technology to improve the transmission quality of the local area network,leading P2P data transmission in the emergence of new features. Presents a P2P identification algorithm based on session behavior characterization which targets STUN by multiple classification analysis. Firstly, STUN packet format for establishing a session is included to illustrate how all kinds of NATs acquire mapped address. Algorithm uses multi - group description of the session, the establishment of a binary tree structure on the analysis of data packets sent and received to identify P2P session, and uses the hash table to store session state. Finally, in Linux circumstance, many experiments on thunder are processed. The experiment results indicate that the algorithm can identify P2P traffic perfectly.
出处
《计算机技术与发展》
2010年第3期5-8,共4页
Computer Technology and Development
基金
国家自然科学基金(60773041)
江苏省自然科学基金(BK2008451)
国家高科技863项目(2007AA01Z404
2007AA01Z478)
2006江苏省软件专项项目
南京市高科技项目(2007软资127)
现代通信国家重点实验室基金(9140C1105040805)
江苏高校科技创新计划项目(CX08B-085Z
CX08B-086Z)
江苏省六大高峰人才项目
