基于网络流量奇异性特征的LDoS攻击检测方法  

LDoS attacks detection approach based on singularity feature of network traffic

作　　者：张晓瑜[1] 吴志军[2] 岳猛[2] 闫长灿 

机构地区：[1]中国民航大学航空自动化学院,天津300300 [2]中国民航大学电子信息工程学院,天津300300

出　　处：《计算机工程与设计》2016年第1期50-54,共5页Computer Engineering and Design

基　　金：国家自然科学基金面上基金项目(61170328;U1333116;U1433103);天津市应用基础与前沿技术研究计划基金项目(12JCZDJC20900);中央高校基本科研业务费基金项目(3122014D022;3122014B002)

摘　　要：依据低速率拒绝服务LDoS(low-rate denial of service)攻击期间受害端网络流量严重下降且网络流量波动性较强的特征,提出一种基于网络流量奇异性特征的LDoS攻击检测算法。采用高斯小波卷积计算信号奇异点,以时间窗口信号的均值和标准差为检测依据,实现LDoS攻击检测。NS-2仿真结果表明,该方法能有效检测LDoS攻击,检测成功率达90.6%。On the basis of low-rate denial of service（LDoS）characteristic that network traffic flow degrades obviously with strong volatility while being attacked,LDoS attacks detection approach was proposed based on singularity feature of network traffic.Gaussian wavelet convolution was used to calculate signal singularity.The mean and standard deviation of the signal of different time window were used as the detection criterion.Results of simulation in NS-2environment show that the proposed approach can effectively detect the LDoS attack with the accuracy of 90.6%.

关 键 词：低速率拒绝服务攻击 奇异性检测 小波变换 高斯小波卷积 网络攻击检测 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]