Deep Learning based Ensemble Convolutional Neural Network Solution for Distributed Denial of Service Detection in SDNs

Abstract

Software defined networks (SDNs) are considered to be the future of networking as it decouples the control plane from the forwarding logic and fulfils the escalating demand of faster and more proficient networks. However, emergence of SDNs also bring security challenges to its centralized architecture such as Distributed Denial of Service (DDoS) attack. Therefore, the need for a timely detection of large-scale sophisticated DDoS attack is of paramount concern for subsequent countermeasures. This paper presents a deep learning (DL) based CNN (Convolutional Neural Network) ensemble solution for efficient detection of DDoS in SDNs. The proposed framework's performance is evaluated through standard evaluation parameters with state-of-the-art Flow-based dataset (ISCX 2017). Empirical results of the proposed framework demonstrate high attack detection accuracy: 99.48% in minimum time with conducive computational complexity.