DDoS Attacks Detection Based on Sequential Change Detection
Liu Qingwei · Jisuanji gongcheng · 2005
This paper presents an effective method to detect DDoS attack by formulating and solving the task as a sequential change point detection problem. It uses non-parametric CUSUM algorithm to complete the detection task of DDoS attacks. The method has several attractive features. First, it has manageable computational complexity. Second, it minimizes the average delay of detection for a given false alarm rate. Third, it is effective in various network conditions. At the end , it demonstrates the efficiency and robustness of the method in various attack scenarios by using simulations.