Edge-Fog-Cloud Distributed Architecture for Intelligent DDoS Detection and Mitigation

Abstract Cloud and distributed infrastructures face significant challenges from increasingly sophisticated Distributed Denial-of-Service (DDoS) attacks. Real-time efficiency is limited by the latency and scalability issues that affect traditional centralized detection systems. This paper presents a multi-layered DDoS detection and mitigation framework built on the Edge-Fog-Cloud paradigm. Hierarchical intelligence is integrated into the architecture to strike a balance between adaptive defense, resource efficiency, and responsiveness. A threshold-guided lightweight classifier quickly distinguishes malicious, suspicious, and benign traffic at the edge. A compact Deep Neural Network (DNN) verifies anomalies in suspicious flows that are escalated to the fog. For context-aware mitigation, a deep classifier at the cloud layer categorizes confirmed attacks into two main families: reflection/amplification and exploitation. Evaluation on the CICDDoS2019 dataset demonstrates high accuracy, a low false-positive rate, and efficient traffic handling. The modular design ensures scalability and adaptability for modern distributed computing infrastructures.