Detecting DDoS attacks against data center with correlation analysis

Abstract

Distributed denial-of-service (DDoS) attacks pose a great threat to the data center, and many defense mechanisms have been proposed to detect it. On one hand, many services deployed in data center can easily lead to corresponding DDoS attacks. On the other hand, attackers constantly modify their tools to bypass these existing mechanisms, and researchers in turn modify their approaches to handle new attacks. Thus, the DDoS against data center is becoming more and more complex. In this paper, we first analyze the correlation information of flows in data center. Second, we present an effective detection approach based on CKNN (K-nearest neighbors traffic classification with correlation analysis) to detect DDoS attacks. The approach exploits correlation information of training data to improve the classification accuracy and reduce the overhead caused by the density of training data. Aiming at solving the huge cost, we also present a grid-based method named r-polling method for reducing training data involved in the calculation. Finally, we evaluate our approach with the Internet traffic and data center traffic trace. Compared with the traditional methods, our approach is good at detecting abnormal traffic with high efficiency, low cost and wide detection range.