Detection of DDoS Attack using Multiple Kernel Level (MKL) Algorithm

Abstract

Software-defined networking (SDN) is a good approach, framework for virtually designing and building hardware network components. In the traditional network domain, fixed automation is made, and it is not possible to change the network connections. SDN has dynamic automation but is still exposed to DDoS attacks. With rising detection accuracy, IDS (Intrusion Detection System) against DDoS still faces provocation in detecting the intrusions and reducing the false alarm rate. In the network, the most efficient way of spotting intrusions is through the deployment of machine Learning (ML) - IDS and deep Learning (DL) - IDS systems. In this paper, our method based on DL proposes an efficient unsupervised level of shallow and deep multiple kernel level algorithms (MKL). To detect the malicious traffic, carry out experiments on DDoS attack databases with the MKL algorithm and correlate the end results with developed methods. Our test outcome reveal that the proposed method provides better accuracy and detection rate.