Federated Learning, Attack Detection and Distributed Denial of Service: A Cloud Computing Environment Study

Abstract

Cloud computing has transformed data storage, processing, and service delivery by offering highly scalable and cost-efficient infrastructures for a wide range of applications. However, its distributed architecture increases exposure to security threats, particularly Distributed Denial of Service (DDoS) attacks that can overwhelm resources, disrupt services, and cause significant operational and financial losses. Traditional centralized detection mechanisms often prove insufficient in such environments due to latency in response, performance bottlenecks, and challenges in maintaining data privacy across dispersed nodes. In this review, federated learning, attack detection and distributed denial of service: a cloud computing environment study was investigated as a decentralized, privacy-preserving, and scalable framework for detecting and mitigating DDoS attacks within cloud computing environments. FL enables distributed cloud nodes to collaboratively train detection models without transmitting raw data, thereby improving detection accuracy, safeguarding sensitive information, and reducing risks associated with single points of failure. The paper discusses the architecture of an FL-based detection framework—encompassing local nodes, central aggregation servers, secure communication protocols, and adaptive learning algorithms—and highlights how their integration enhances real-time anomaly detection, scalability, and resilience to evolving cyber threats...