DAMS: Double Deep Q-Network Approach for DDoS Attack Mitigation in SDN

5G represents the most recent and sophisticated generation of mobile communications. It boasts features such as rapid response times, minimal latency, and substantial bandwidth. 5G networks leverage advanced technologies, including Software-Defined Networking (SDN), which plays a pivotal role in designing these networks, enabling network administrators to centralize their control and management. These networks are increasing sharply in size as well as in functionality, especially with the growth of the Internet of Things (IoT), which has given rise to numerous challenges. Nowadays, attackers are developing diverse techniques to exploit vulnerabilities and insufficient security in IoT devices, making them exploitable in large-scale attacks. Consequently, it has become mandatory to define appropriate countermeasures in order to defend these networks against different attack types. Many studies have attempted this using traditional methods, which may be efficient but often prove inadequate or detrimental to legitimate users, highlighting the need for more intelligent solutions. In this article, we use the Double Deep Q-Network (DDQN), a reinforcement learning (RL) algorithm, in conjunction with the programmability of SDN to mitigate DoS attacks in 5G networks. The mitigation process is realized first by supporting remediation selection to maximize the response efficiency while reducing the adverse impact on the network. Second, it involves automating remediation deployment, which minimizes downtime and reduces manual and error-prone incident handling.