基于统计分析建立流量动态临界线的蠕虫检测机制研究  

Worm detection technology research of net-flow dynamic critical line established based on statistical analytic method

作　　者：王勇超[1] 谢永凯[1] 朱之平[1] 林怀忠[2] 

机构地区：[1]浙江大学网络信息中心,杭州310027 [2]浙江大学人工智能所,杭州310027

出　　处：《计算机应用研究》2010年第3期1032-1034,共3页Application Research of Computers

基　　金：国家"863"计划资助项目(2007AA01Z197;2008AA01Z416)

摘　　要：提出了一种基于正态分布进行异常流量检测,从而判断当前内网中是否存在蠕虫感染的方法。该方法根据历史流量的正态分布统计特性,计算出网络内数据流量的一般行为的可信区间,如果监控的流量超出该可信区间,则判断为异常流量并作出蠕虫威胁的报警。结合该方法,进一步分析了如何以双因素模型分析网络中蠕虫的数量。This paper raised a method detect the abnormal net-flow based on normal distribution, then estimated the existence of Internet worm in internal network. According to the normal distribution character of the history flow, this method computed the normal behavior trusted zone of data flow in network, judged the inspected flow abnormal flow if it went beyond the trusted zone, and alarmed the threat of Internet worm. Combined with this method, further analyzed how to use two-factor model analysis of the number of Interuet worms in network.

关 键 词：正态分布 异常流量 可信区间 蠕虫 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]