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 KP-CA network,it has higher detection accuracy and accuracy,and lower false positive rate and false negative rate.
MFOPA algorithmLDoS attack detectionnetwork traffic characteristicsjoint characteris-ticsoutlier probability analysis
NETL
NSTL
万方数据
