Mitigating Data Imbalance in DDoS Detection for SDN Through Machine Learning Methods
Wei Song, Zakaria Alomari, Xiaoyu Zhang, Benxin Xie · 2025
With the widespread adoption of Software-Defined Networking (SDN), Distributed Denial of Service (DDoS) attacks pose significant threats to network security. Machine learning-based detection methods suffer from data imbalance, where normal traffic significantly outweighs attack traffic, leading to biased models. This study proposes an optimized Voting Classifier that combines Decision Tree and Random Forest algorithms with resampling techniques to improve minority class detection. Experimental results show that the proposed method achieves 99.99% accuracy, 100% recall, and an AUC-ROC of 100%, outperforming baseline classifiers such as Random Forest and Gradient Boosting. Additionally, we evaluate the model's deployment feasibility in SDN using Mininet and the RYU controller. The findings demonstrate the practicality of integrating machine learning-based DDoS detection in real-world SDN environments.