Detecting DDoS attacks in Software Define Networking: A Machine Learning Based Approach

Software-Defined Networking (SDN) is a type of network strategy that enables the separation of a network's control plane and data plane, thereby allowing centralized management and control of network traffic. Although SDN has many remarkable advancements, it encounters several security concerns like DDoS attack. The Distributed Denial-of-Service (DDoS) attack is a fast-growing threat that endangers the future of internet technology. In this paper, some machine learning classifiers namely support vector machine (SVM) using linear based kernel and radial basis function (RBF) based kernel, random forest (RF), decision tree (DT), logistic regression (LR), and k-nearest neighbors (KNN) are evaluated to identify and analyze DDoS assaults within an SDN environment. A comparative performance study according to the results from the experiments is described to determine the best suitable techniques for DDoS detection. The results show that some classifiers reached up to 100^% accuracy and the performance is validated using ROC-AUC curve. By utilizing proper classifiers derived from machine learning and taking the benefits of SDN, the controller can be safeguarded against DDoS attacks.