基于MFOPA算法的LDoS攻击检测

LDoS attack detection based on MFOPA algorithm

LDoS攻击隐蔽性强,结合KPCA算法提取并融合的TCP流量均值、变异系数、信噪比、能量熵、TCP流量和总流量相关性5种网络流量特征,结合WSOS算法进行离群概率分析,提出基于MFOPA算法的检测方法。仿真结果表明,所提方法在NS2平台、test-bed平台、LBNL数据集和WIDE2018数据集上均可有效检测LDoS攻击,相较于AEWMA、Multifractal、KPCA network等其他检测算法,具有更高的检测准确率与精确率,且误报率和漏报率更低。

LDoS attack is strong,Combined with TCP traffic mean,coefficient of variation,signal-to-noise ratio,energy entropy,TCP traffic and total traffic correlation extracted and fused by KPCA algorithm and five network traffic characteristics related to the total traffic,this paper analyzes the outlier probability with WSOS algorithm,and proposes a detection method based on MFOPA algorithm.Simulation results show that the proposed method can effectively detect LDoS attacks on NS2 platform,test-bed platform,LBNL data set and WIDE2018 data set.Compared with other detection algorithms such as AEWMA,Multifractal and KPCA network,it has higher detection accuracy and accuracy,and lower false positive rate and false negative rate.