Edge Computing And Internet Of Things Security Frameworks: A Comprehensive Survey Of Architectures, Threats, And Countermeasures

Abstract

The rapid proliferation of Internet of Things (IoT) devices has fundamentally transformed the computing landscape, generating unprecedented volumes of data at the network periphery. Edge computing has emerged as a critical paradigm to address the latency, bandwidth, and privacy limitations inherent in traditional cloud-centric IoT architectures. However, the convergence of edge computing and IoT introduces a complex and expanded attack surface that demands rigorous security frameworks. This paper presents a comprehensive survey of edge computing architectures for IoT security, encompassing Multi-access Edge Computing (MEC), fog computing, and cloudlet-based paradigms. We systematically analyze the threat landscape across all IoT layers—device, network, edge/fog, and application—and evaluate existing security frameworks addressing authentication, encryption, and intrusion detection at the edge. A comparative analysis of machine learning-based, deep learning-based, federated, and hybrid intrusion detection systems deployed on edge infrastructure is presented, demonstrating that hybrid ensemble approaches achieve superior detection performance with F1-scores exceeding 95%. We identify critical research gaps in lightweight cryptographic protocols, privacy-preserving computation, and standardized security benchmarks for resource-constrained edge environments. The findings provide actionable guidance for researchers and practitioners designing secure edge-IoT ecosystems.