DDoS Attack in Software-Defined Networks

Software-Defined Networking has emerged as a revolutionary architecture in the field of network management and security.This paper presents a evaluation of advanced machine learning techniques for detecting Distributed Denial of Service attacks in SDN environments.The study aims to provide an in-depth review of SDN and assess the performance of several ML algorithms, including Logistic Regression, Decision Tree, Random Forest, and K-Nearest Neighbors, using a DDoS SDN dataset and aim to find a best model for prevent attacks.The dataset contains 104,345 entries across 23 features, including packet count, flow duration, and network traffic metrics.Evaluation of model performance reveals varied results, with Random Forest and Decision Tree models achieving near perfect accuracy on the test set and our models are generalized, while Logistic Regression and KNN showed promising but slightly lower results then other model.The findings demonstrate the effectiveness of advanced ML algorithms in enhancing DDoS detection capabilities, with potential applications for real-time network security systems in SDN environments and with this we can make a robust structure and secure networks from attacks.