Security and Privacy Issues in Fog Computing for the Internet of Things

Abstract

In recent decades, the world has witnessed a significant increase in connectivity due to the widespread adoption of the Internet of Things (IoT), cloud, and fog computing. These technologies have revolutionized data collection, processing, and storage. Fog computing, designed to reduce latency and enable local data preprocessing, addresses some limitations of centralized cloud computing. However, it faces significant challenges in ensuring data privacy and security, being vulnerable to cyber-attacks. To fully benefit from fog computing, robust security and privacy techniques must be implemented. In this thesis, we propose two contributions that utilize various technologies. Given the diverse privileges of users accessing fog networks, our first contribution involves the adoption of an Adaptive Neuro-Fuzzy Inference System (ANFIS) within Software-Defined Networking (SDN). This approach aims to detect and mitigate Syn flood Distributed Denial-of-Service (DDoS) attacks in fog computing networks. In our second contribution, we present a Federated Learning-Based Intrusion Detection System (IDS) approach specifically designed for IoT-enabled smart healthcare systems. This approach uses decentralized identifiers (DID) and verifiable credentials (VC) to facilitate user authentication. Through experiments conducted on securing the fog-enabled IoT network, Our proposed solution combines IDS and SDN capabilities, using machine learning to provide robust security and preserve user privacy in fog computing environments. Experimental findings demonstrate the effectiveness of this integrated approach in balancing strong security measures and user privacy preservation