Leveraging SDN for early detection and mitigation of DDoS attacks

Abstract

Distributed Denial of Service (DDoS) attacks being one of the most challenging security issues in the current network requires a lot of attention from the research community. Detection and mitigation of DDoS attacks at early stages could reduce the impact of the attack on legitimate users. Software Defined Networking (SDN) has emerged as a technique to aid the resolution of DDoS attacks effectively. This paper proposes one such detection scheme that utilizes Radial Basis Function networks optimized with Particle Swarm Optimization for early detection of DDoS attacks in SDN networks. A feature set for training and testing of detection module is also proposed that allows the identification of DDoS attacks. The proposed detection scheme is efficient enough to classify the heavy load of network traffic from that of DDoS attacks. Not only detection is important in such scenario, but the mitigation technique also needs to be selected very carefully in order to meet the desired network requirements as well as to secure the legitimate users. For the purpose of identification of suitable mitigation scheme an analytical comparison of possible controller based mitigation techniques is presented. These techniques are further compared based on several parameters governing the effect of mitigation on network users and processing.