Game Theory for Strategic DDoS Mitigation

Mitigating DDoS attacks on the networks is an important security challenge for the Internet Service Providers. Recent work has shown a flexible and elastic DDoS defense mechanism via Software Defined Networking/Network functions virtualization and such systems are showing increasingly efficient performance. However, an intelligent attacker can exploit the system by dynamically adapting the attack profile which a static defense cannot adequately handle. We propose a game theoretic model for a dynamic deceptive defense. We decompose the problem in two phases. The first phase focuses on optimising the defense while trying to deceive the attacker with dynamic adaptations in the defense strategy. The second phase consists of optimising the resource allocation in order to execute the pre-computed defence strategy. We show analytical and numerical results showing the efficiency of our computation algorithms.