基于支持向量机的P2P网络DoS攻击检测  

Detection Mechanism of DoS Attacks in Peer-to-Peer Networks Based on Support Vector Machine

作　　者：吴敏[1] 王汝传[1] 王治平[2] 

机构地区：[1]南京邮电大学计算机学院,江苏南京210003 [2]中兴通讯股份有限公司,江苏南京210012

出　　处：《计算机技术与发展》2009年第11期151-154,共4页Computer Technology and Development

基　　金：国家自然科学基金(60573141;60773041);江苏省自然科学基金(BK2008451);国家高科技863项目(2007AA01Z404;2007AA01Z478);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);中兴通讯高校合作基金

摘　　要：对等网络技术近年来发展迅速,但其安全性问题一直是P2P网络进一步发展和应用急待解决的重要问题之一。由于对等网络的松散性,基于洪泛式请求的拒绝服务(DoS)攻击已经成为主要威胁之一。文章首先介绍和分析了DoS攻击在对等网络下的特点,然后提出了一种基于支持向量机的P2P环境下DoS攻击的检测模型,该模型能够通过以离线的方式对发生DoS攻击时流的统计特性进行特征提取,并能实时识别攻击的发生。实验证明,这种模型具有较高的检测率和较低的误检率。Peer- to- peer network technology has improved quickly in recent years, however the security problem is always the most important issue that impede the further improvement of P2P application which should be solved immediately. Among all these security issues the denial -of- service（DoS） attacks are becoming one of the major threats in peer - to - peer networks which are loosely connected. First describe the characters of DoS attack in P2P environment, then propose a DoS attack detection model based on support vector machine mechanism. The model can extract flow statistics feature when DoS attack occurs offline and can identify the event online. The result of the experiment shows that this model has high detection rate and low false reject rate in detecting DoS attack in P2P environment.

关 键 词：对等网络 拒绝服务 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]