Ensemble Based Detection Model for DDoS Attacks in SDNs Using Advanced Feature Selection
Abdussalam Ahmed Alashhab, Aisha Edrah, Mohd Soperi Mohd Zahid, Md. Siddikur Rahman · 2024
Software-Defined Networking (SDN) enhances flexibility, scalability, and innovation by decoupling the control plane from the data plane, managed through streamlined controller operations. However, Distributed Denial of Service (DDoS) attacks pose significant cybersecurity threats to SDNs, disrupting services by flooding targeted systems with traffic from multiple sources. Real-time detection of these attacks remains challenging, as traditional methods often lack the ability to accurately identify complex attack patterns due to limited feature sets. To address this, we propose an ensemble-based model that combines three classifiers (SGD, EBM, and MLP) for effective DDoS attack detection and mitigation in SDNs. Our approach integrates a reliable feature selection methodology, leveraging Principal Component Analysis (PCA) to identify the most informative features for attack detection. This enhancement significantly improves the accuracy and efficiency of the ensemble model. Evaluated using the CIC-DDoS2019 and InSDN datasets, our model demonstrates substantial improvements in detection rates and computational efficiency. Results show that the proposed approach achieves 99% accuracy, underscoring its potential as a resilient solution for DDoS attack detection in SDNs.