An AI Based IDS Framework For Detecting DDoS Attacks In Cloud Environment

Abstract

ABSTRACT Cloud computing makes it easier for users to access resources from anywhere at any time. This is for as long as they have access to the internet connectivity by employing a “pay-as-you-use” model. Despite its merits, cloud computing faces shortcomings, notably the escalating security concerns linked with it. Distributed Denial of Service (DDoS) attack is a primary and biggest concert to the availability of the services offered by cloud. DDoS attacks use numerous machines to flood consumers with packets with high data overhead, flooding the network with unwanted traffic. Due to the obsolete datasets, many deep learning (DL) models are processing-intensive or may not successfully address new DDoS threats. This paper seeks to address this issue by proposing FEwDN, an AI-based DDoS detection framework that employs a hybrid approach, integrating machine learning and deep learning algorithms. The framework optimizes feature selection via ensemble techniques, enhancing accuracy by leveraging deep neural networks for traffic classification. The proposed framework is experimented on the CICDDoS2019 dataset and demonstrates superior performance over benchmark techniques across multiple metrics. The FEwDN outperforms well with other models against various performance metrics. This research strengthens cloud security and DDoS detection in modern clouds.