Evaluation of MLDL algorithms in detecting DDOS attacks for subset features in SDN environment

The growing complexity and scale of modern network infrastructures have made them more vulnerable to Distributed Denial of Service (DDoS) attacks, posing significant risks to service availability. Software-Defined Networking (SDN) has emerged as a promising architecture to enhance network management and security. This research focuses on evaluating various machine learning and deep learning (MLDL) algorithms for detecting DDoS attacks in an SDN environment using a subset of network traffic features. The proposed approach includes a hybrid model combining Random Forest (RF) and Decision Tree (DT), along with individual models such as Support Vector Machine (SVM) and XGBoost. These models aim to classify network traffic into benign and DDoS-affected categories based on feature subsets like packet flow, traffic patterns, and anomaly indicators. The performance of each model is assessed in terms of accuracy, precision, recall, and F1-score, with the goal of identifying the most effective algorithm for real-time detection of DDoS attacks in SDN-based networks. The results demonstrate that the hybrid RF-DT model outperforms the other algorithms, offering a robust solution for mitigating DDoS threats in SDN environments. This study provides valuable insights into the integration of MLDL techniques for improving network security and resilience against evolving cyber threats.

Keywords: DDoS attacks, Software-Defined Networking (SDN), Machine Learning, Deep Learning, Hybrid Model, Random Forest, Decision Tree, Support Vector Machine (SVM), XGBoost, Network Traffic, Anomaly Detection, Classification, Cybersecurity,  Network Security.