A novel DDoS attack detection method

This paper presents a novel network-wide PCA based detecting method (WPCAD) for detecting the increasing serious Distributed Denial-of-Service (DDoS) attack. Due to the correlation changes of traffic caused by DDoS attack, we construct a model of origin-destination (OD) traffic input matrix from the point of network-wide and analyze the correlations of multiple links with the same destination. We validate our method by applying it to real network traffic with well known and identified anomalies. The experiment shows that the presented method is effective in detecting DDoS attack and has a higher detection rate compared to the current abnormal traffic attacking detection methods. Copyright © 2013 Binary Information Press.