Research on DDoS detection method based on super logarithm and SVM algorithm
Lili Cheng, Pengbo Xiao, Xinjian Zhang, Yanqi Liu, Lisha Gao, Pan Zhou · 2022 5th International Conference on Data Science and Information Technology (DSIT) · 2022
By analyzing the characteristics of DDoS attacks, this paper designs a primary detection process based on super logarithm algorithm according to the characteristics of attack forgery source IP; In terms of the characteristics of the change of network state metadata caused by the attack, the manual and PCA methods are used to extract detection features from the fine-grained parameters obtained by in-band measurement, and a two-level detection process based on SVM algorithm is designed. The experimental results show that compared with entropy threshold method and KNN method, the detection method in this paper has higher accuracy and occupies the lowest CPU resources.