Real-Time DDoS Detection in Cloud-Native Infrastructure Using Ensemble Learning and Adversarial-Aware Datasets

Abstract

Cloud computing environments have become attrac- tive targets for Distributed Denial of Service (DDoS) attacks due to their scalability, elasticity, and shared multi-tenant nature. Traditional security defenses often fall short under evolving attacks. This paper presents a real-time DDoS detection frame- work leveraging ensemble machine learning, specifically Random Forest, Gradient Boosting, and a Soft Voting Classifier, trained and validated on the latest benchmark datasets including CICD- DoS2019 and TON IoT. The solution is containerized and de- ployed on a Kubernetes-based virtualized environment, providing robust defense, scalability, and adaptability. The proposed model achieved an accuracy of 98.6% and demonstrated strong real- time detection and low false-positive rates. In-depth comparisons with recent research methods and datasets are presented.