Software Defined Networking (SDN) is very popular due to the benefits it provides such as scalability, flexibility, monitoring, and ease of innovation. However, it needs to be properly protected from security threats. One major attack that plagues the SDN network is the distributed denial-of-service (DDoS) attack. There are several approaches to prevent the DDoS attack in an SDN network. We have evaluated a few machine learning techniques, i.e., J48, Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbors (K-NN), to detect and block the DDoS attack in an SDN network. The evaluation process involved training and selecting the best model for the proposed network and applying it in a mitigation and prevention script to detect and mitigate attacks. The results showed that J48 performs better than the other evaluated algorithms, especially in terms of training and testing time.

Additional Metadata
Keywords DDoS, J48, Machine Learning, SDN, Weka
Persistent URL dx.doi.org/10.1109/SERVICES.2019.00051
Conference 2019 IEEE World Congress on Services, SERVICES 2019
Citation
Rahman, O. (Obaid), Quraishi, M.A.G. (Mohammad Ali Gauhar), & Lung, C.H. (2019). DDoS attacks detection and mitigation in SDN using machine learning. In Proceedings - 2019 IEEE World Congress on Services, SERVICES 2019 (pp. 184–189). doi:10.1109/SERVICES.2019.00051