An Artificial Intelligence Resiliency System (ARS)

Security is one of the principle concerns confronting the improvement of new networking and communications systems. Another challenge is to verify network consistency, such that it is functioning precisely as planned. On the other hand, propels in Artificial Intelligence (AI) innovation are expected to play a key role in offensive and defensive measures to ensure a fast reaction to respond to the landscape of developing threats. Software Defined Networking (SDN), being centralized by design, offers a global perspective of the network. Adaptability and robustness are part of a package offered by programmable networking, which drove us to consider the integration between both AI and SDN. The hybrid security techniques, inspired by the fascinating tactics of the human immunity system, directed the design of our general hybrid Artificial Intelligence Resiliency System (ARS). ARS aims to strikes a good balance between centralized and distributed security techniques that may be applicable to various network environments. Another objective is to investigate and leverage the state-of-the-art AI techniques to enhance network performance in general and resiliency in particular. A second objective of ARS is network consistency verification as a resiliency measurement alongside our security system. Throughout this work we discuss a new architecture that integrates both, a multilayer layer security and consistency establishment techniques, while preserving data privacy. Our results show, on one hand, the architecture tests along with the accuracy of multiple AI techniques for both security layers, and on the other hand, the attack mitigation and consistency tests. Finally, we propose a new distributed AI based security enforcement technique based on a state-of-the-art distributed Neural Network overlay.