Flooding DDoS mitigation and traffic management with software defined networking

Mitigating distributed denial-of-service attacks can be a complex task due to the wide range of attack types, attacker adaptation, and defender constraints. We propose a defense mechanism which is largely automated and can be implemented on current software defined networking (SDN)-enabled networks. Our mechanism combines normal traffic learning, external blacklist information, and elastic capacity invocation in order to provide effective load control, filtering and service elasticity during an attack. We implement the mechanism and analyze its performance on a physical SDN testbed using a comprehensive set of real-life normal traffic traces and synthetic attack traces. The results indicate that the mechanism is effective in maintaining roughly 50% to 80% service levels even when hit by an overwhelming attack.