采用负熵目标函数和自相关成分的攻击流检测  

Detection of Attack Flows Based on Autocorrelation Components Analysis and Negative Entropy Objective Function

作　　者：彭天炜[1] 

机构地区：[1]成都职业技术学院软件学院,成都610041

出　　处：《科技通报》2014年第8期83-85,共3页Bulletin of Science and Technology

基　　金：四川省自然科学基金(2013CDZ088)

摘　　要：降质服务RoQ作为一种新型攻击方式,比传统的拒绝服务DoS攻击隐蔽性更强,传统检测方法难以实现高效定位检测攻击源。为快速准确地将攻击流分离检测,提出一种采用负熵作为目标函数,基于自相关成分分析的攻击流特征检测算法。建立自相关成分分析和RoQ攻击流特征盲源分离算法盲源分离数学模型,采用负熵目标函数,通过自适应调整分离系数,最终达到快速分离源信号的目的,实现了对RoQ攻击流特征分离和准确定,最后进行系统流程设计。实验对比表明,算法能很准确地实现攻击信号与合法信号的分离检测,攻击特征检测概率远高于传统算法的检测概率,检测性能具有明显提高。在计算机网络安全防护领域特别是对RoQ类攻击方面有很好的应用价值。Reduction of Quality （RoQ） attacking method a new means of network attack, it was more concealment than the traditional Denial of Service （DoS） attack, it was more difficult to achieve efficient detection result. Especially for the tradi-tional method, it was hard detection the detection source. For the fast and accurate attack detection and signal flow separa-tion, and optimal separation and detection of attack flows method was proposed with autocorrelation components analysis al-gorithm based on negative entropy objective function. Through the adaptive adjustment of separation parameters and the RoQ attacking feature was detected and located, finally, the system flow was designed. Simulation result shows that the new method has more perfect detection performance with lower error detection rate. It has prospective application value in RoQ network attack detection.

关 键 词：自相关 网络攻击 成分分析 网络安全 

分 类 号：TN911.23[电子电信—通信与信息系统]