Detecting Low-Rate DDoS Attacks in SDN Using Ensemble Machine Learning Techniques
Abdinasir Hirsi, Lukman Audah, Adeb Salh, Nan Mad Sahar, Mohammed A. Alhartomi, Salman Ahmed · 2024
Software-Defined Networking (SDN), a critical enabler of 6 G and IoT systems, is rapidly evolving due to its ability to simplify network management and enhance flexibility. However, its centralized architecture makes it highly vulnerable to security threats, including low-rate Distributed Denial of Service (LDDoS) attacks. These attacks exploit protocol weaknesses, disrupting SDN operations, increasing latency, and degrading reliability—key challenges for ultra-reliable and low-latency communication. This study addresses these threats by employing machine learning (ML) techniques, using the publicly available CICDoS2017 dataset to ensure reproducibility and validation. An ensemble ML classifier was proposed and achieved a 98% detection accuracy, outperforming existing approaches. This work highlights the importance of securing SDN environments as a foundational step toward enabling robust, low-latency, and reliable communication in future 6G and IoT systems.