Abstract:
Software Defined Networking (SDN) is an emerging network architecture that has many advantages such as programmability, flexibility, cost effectiveness and dynamicity compared to traditional networking. Software-defined network technology is a physically distributed but logically centralized network that is centrally controlled by controller software. SDN allows network administrators to manage the network service through decoupling into layers to ease network management. The separated layers make SDN vulnerable to different attacks. Distributed Denial of Service (DDoS) is a major threat on SDN and to identify DDoS attack traffic classification on SDN is more difficult because of the fake widely used address (Spoofed IP). Hence, there arises a need to accurately and efficiently detect DDoS attack on the SDN. This paper proposes a machine learning technique known as eXtreme Gradient Boosting algorithm (XGBoost) is used for detecting distributed denial of service attack on a software defined network. XGBoost is an efficient, scalable and computationally fast, however it is regarded an excellent classifier. The performance of the algorithm was compared with some machine learning algorithm such as Naïve Bayes, Support Vector Machine and Logistic Regression, experimental result shows that XGBoost has higher Accuracy rate and Detection rate.
Keywords: DDoS Attacks, SDN, XGBoost, Detection, and Networking,
