Securing Reliable Fog Computing Environments: A Systematic Literature Review and Future Research Directions
TL;DR
Effective protection against stealthy data theft in fog computing requires a shift from isolated security controls toward integrated and adaptive protection, and future fog-security architectures should combine continuous behaviour profiling, dynamic trust evaluation, anomaly detection, and context-aware decoy selection while considering the resource, latency, scalability, and reliability constraints of fog environments.
Abstract
Fog computing extends cloud services toward edge nodes by distributing computation, storage, and networking closer to data sources. This architecture can reduce network latency and bandwidth requirements. However, the distribution of computing resources also introduces reliability challenges and increases the attack surface. Traditional security mechanisms may provide limited protection when attackers use valid credentials or compromise legitimate fog nodes. This systematic literature review examines the development and integration of reliability mechanisms, security frameworks, behavioural analytics, dynamic trust evaluation, and deception-based approaches in fog computing, with particular attention to data-theft prevention and insider threats. The review considers studies published between 2012 and 2025 and uses a structured literature-search, screening, eligibility assessment, quality evaluation, and thematic synthesis process. The findings indicate that existing fog-security approaches are largely based on preventive mechanisms, including authentication, encryption, access control, and trust management. These mechanisms provide essential protection but may be insufficient against insiders or compromised entities using legitimate privileges. Data theft can occur without disrupting service availability. The review further indicates that deception mechanisms are commonly investigated separately from behavioural analytics and dynamic trust evaluation, while adaptive integration remains limited. Effective protection against stealthy data theft in fog computing therefore requires a shift from isolated security controls toward integrated and adaptive protection. Future fog-security architectures should combine continuous behaviour profiling, dynamic trust evaluation, anomaly detection, and context-aware decoy selection while considering the resource, latency, scalability, and reliability constraints of fog environments.