ARTIFICIAL INTELLIGENCE AND SOFTWARE-DEFINED NETWORKING: INTEGRATING MULTI-PRECAUTIONARY MEASURES

Software-defined networking (SDN) separates the network control from the data plane in order to allowprogrammability and flexible management. But this separation, in a way, is responsible for new vulnerabilities that amalicious actor may exploit. To this end, AI (artificial intelligence) techniques provide adaptive detection anddecision-making capabilities that eventually reduce these types of risks. This article reviews the literature on the useof AI by SDN for multi-layer precautionary measures. It identifies issues in the integration of AI and SDN, looks atthe research examples in the fields of intrusion detection, traffic optimization and fault management, and comparesthe performance metrics of these methods. In the end, it features frameworks and subsequent research orientationsfor resilient AI–SDN systems.