Enhanced SDN Defense: Analysis of DDoS Attack Detection Techniques Using Machine Learning

Software defined network (SDN) is drawing attention to handling next-generation traffic through centralized management. However, the centralized aspect of the network makes it more vulnerable because the central controller might be targeted by malicious actors looking to interfere with network operations and data security. Typically, the SDN controller defense system uses machine learning (ML) techniques to inspect data from individual IP to detect distributed denial-of-service and intrusion attacks. As soon as an anomaly is detected, the system enables rapid response to mitigate attacks and ensure smooth system operations. In the accessed model, the CICDDoS-2019 public dataset is used to evaluate and compare the performance with various ML algorithms. Ultimately, an efficient mitigation approach is suggested and utilized for each detection method. The results suggest guaranteed detection rates and efficient throughput. Therefore, the multi-layer perceptron approach is suitable for the SDN defense system.