Machine Learning Approaches for Mitigating Distributed Denial of Service Attacks: A Systematic Review of Advanced Security System

Abstract

Distributed Denial of Service (DDoS) attacks are among the most destructive cyber threats. They push so much fake traffic to these sites that they can no longer be reached by real users. It is hard for legacy intrusion detection systems to catch new threats as soon as they arise. This review focuses on how ML helps find DDoS attacks, mainly in SDNs, IoT systems and those used in Agriculture 4.0. Using both supervised, unsupervised and hybrid ML models greatly boosts the accuracy and expands the applications of detection technology. We also review important datasets, different feature engineering solutions and how models perform, giving a full overview of both existing research and upcoming trends.