Achieving DDoS resiliency in a software defined network by intelligent risk assessment based on neural networks and danger theory

Distributed Denial of Service (DDoS) attacks are becoming a very versatile weapon. Unfortunately, they are becoming very popular amongst cyber criminals, and they are also getting cheaper. As the interest grows for such weapons on the black market, their scale reaches unimaginable proportions. As is the case of the Spamhaus attack, which was mitigated by CloudFlare through null-routing techniques. This paper presents a way of mitigating DDoS attacks in a Software Defined Network (SDN) environment, by assessing risk through the means of a cyber-defense system based on neural networks and the biological danger theory. In addition to mitigating attacks the demo platform can also perform full packet capture in the SDN, if the central command component deems it necessary. These packet captures can be used later for forensic analysis and identification of the attacker.