Application of Artificial Intelligence to Software Defined Networking: A Survey

Background/Objectives: This paper surveys the application of Artificial Intelligence (AI) to the Software Defined Networking (SDN) paradigm which is a part of previous efforts to give the computer networks the ability of being programmed based on the separation between the control and forwarding planes. In SDN approach, the controller represents the central brain of the network which leads to an advanced level of flexibility and network intelligence. Methods/Statistical Analysis: Different artificial intelligence-based techniques have been applied to achieve an enhanced load balance, network security and intelligent network applications in the SDN approach. Findings: Ant colony algorithms were successful in increasing the maximal Quality of Experience (QoE) by 24.1% compared with the shortest path routing approach. Neural network based intrusion prevention system has shown a scalable performance with low false positive rate. Applying reinforcement learning based technique in adaptive video streaming system compared with the shortest path routing and greedy-based approaches has shown decreasing of the frame loss rate by 89% and 70% respectively. Applications/Improvements: This study highlights the first attempts for applying artificial intelligence in SDN paradigm. However, hybrid intelligent techniques could be the key for achieving more advanced behaviour in SDN-based networks.