Multi-SDN Based Cooperation Scheme for DDoS Attack Defense

Abstract

Distributed Denial of Service (DDoS) attack is one of the most severe threat in current internet. Software Defined Network (SDN) is novel network structure based on the idea of separation of control plane and data plane. SDN allows us to program and monitor networks, and decide how to forward a packet, so it provides a new solution to defend DDoS attack. This paper proposes a multi-SDN Based cooperation scheme to defend DDoS attack. We adopt machine learning to detect DDoS attack, and design a protocol to enable communication among controllers. This protocol can achieve two goals, one is to build and maintain an independent network among controllers of different SDN, and the other is to enable attack information exchange among controllers, so they can find attacker and mitigate DDoS attack. The experimental results show that the proposed protocol can achieve high detection accuracy, find attackers accurately and mitigate DDoS attack traffic effectively with a relatively low cost and latency.