Machine Learning for Securing SDN based 5G Network

Abstract

The fifth-generation (5G) network supports many systems such as reliable communication in potential applications that require maximum security. Advancement in Software-Defined Networking (SDN) is growing with the emerging network architectures targeted from many servers with the various types of Distributed Denial of Service (DDoS) attackers. When malicious users send DDoS attacks, the SDN based 5G networks face security problems and challenges. Despite the security solutions for preventing DDoS attacks in SDN, securing the SDN controller is one of the challenging problems. The purpose of this research is to analyze the suitable machine learning (ML) for securing the SDN controller targeted by DDoS attacks. This paper proposes a security scheme that includes the ML algorithm, adaptive bandwidth mechanism, and dynamic threshold technique. Therefore, the main focus is on the mitigation scheme of DDoS attacks considered in SDN controller through the ML trained model. In this scheme, the proposed approach uses the best ML as a method for finding security solutions that enhance the security of the SDN controller and network performance. In this method, the Extreme Gradient Boosting (XGBoost) and other ML algorithms were used, which not only enhance the accuracy of the security solutions but also improve the overall network performance.