Detecting Distributed Denial of Service Attacks in Mobile Edge Computing using Modified Extreme Machine Learning

Abstract

Mobile Edge Computing (MEC) is a promising technology which enables 5G and  reduces latency. By bringing cloud computing capabilities closer to end users, MEC enables latency-sensitive applications to perform more efficiently. However, security attacks pose significant challenges to the objectives of 5G with Distributed Denial of Service (DDoS) attacks being a major threat. These attacks can overwhelm target systems with excessive data preventing access to and disrupting network services. Effective mitigation strategies are required to protect MEC technology. Given the high data volume generated by such attacks, this paper utilizes a modified Firefly Algorithm to select relevant features. These selected features are then used to train a proposed variant of Extreme Learning Machine (ELM), where weights are initialized using Neighbourhood-Based Differential Evolution. MATLAB simulations demonstrate that the proposed modified ELM outperforms traditional approaches, providing an effective solution to DDoS attacks in MEC.