Advances in the Security of SDNs: Exploration of Recent ML-Based Approaches

In contrast to conventional networks, Software-Defined Networking has revolutionized network control by separating data and control layers, introducing a dynamic and programmable infrastructure. Despite its benefits, SDN faces various security challenges. This study investigates specific attack vectors, such as DDoS/DoS attacks, topology poisoning, and malicious applications, along with machine learning applications that exhibit several benefits in addressing SDN-specific vulnerabilities and attacks. The study emphasizes the need for investigations and contributions directly addressing unique security challenges across SDN planes, interfaces, and attack categories with the help of ML techniques.