AI-Based Intrusion Detection and DDoS Mitigation in Fog Computing: Addressing Security Threats in Decentralized Systems

Abstract

Fog computing is a decentralized paradigm designed to bring computation and services closer to the edge of the network. It has emerged as a promising solution for latency-sensitive applications. However, this architectural shift introduces a set of security challenges that are distinct from traditional centralized cloud environments. Traditional centralized security models are often inadequate for fog environments due to their reliance on centralized data processing, which contrasts with the distributed, heterogeneous, and latency-sensitive nature of fog computing. Distributed Denial of Service (DDoS) attacks, unauthorized intrusions, data breaches, malware propagation, and privacy threats are problematic in fog computing due to its decentralized structure, resource constraints, and the heterogeneous nature of fog nodes. This paper focuses on identifying the critical security threats that fog computing environments face, with a special emphasis on DDoS attacks and other forms of intrusion. In light of these challenges, the role of Artificial Intelligence (AI)-driven solutions in mitigating these security risks is also examined. This study discussed how AI-based techniques, including machine learning (ML), deep learning (DL), and reinforcement learning (RL), offer innovative approaches for real-time threat detection, anomaly recognition, and adaptive mitigation strategies. Deploying AI-based models in fog environments presents challenges such as limited computational resources, latency concerns, and energy constraints. This paper also discusses how AI can be used to enhance security in fog computing while addressing the inherent vulnerabilities of decentralized systems.