A systematic review on security aspects of fog computing environment: Challenges, solutions and future directions

The dynamic and decentralized architecture of fog computing, which extends cloud computing closer to the edge of the network, offers benefits such as reduced latency and enhanced bandwidth. However, the existing fog architecture introduces unique security challenges due to the large number of distributed fog nodes, often deployed in diverse and resource-constrained environments. Further, the proximity of fog computing nodes to end-users and the open, distributed nature of the architecture make fog environments particularly vulnerable to unauthorized access and various types of cyberattacks. Therefore, in order to address these challenges, the study presented a detailed systematic review that aims to analyze existing security technologies in fog computing environments, identify current security gaps, and propose future research directions. The comprehensive literature review uses quality databases, focusing on articles published within the last four years, i.e. from 2020 to 2024. Further, the review followed a systematic methodology with clear inclusion and exclusion criteria to ensure relevance and quality with respect to security in fog computing. Consequently, key research questions are also formulated and answered for addressing various security concerns, such as architectural security, IoT integration vulnerabilities, and dynamic security management. Finally, the detailed review summarizes the key findings through MTGIR analysis to give valuable insights on the existing security framework of fog computing systems. The result analysis further revealed that 16% of the research is focusing on blockchain and elliptic curve cryptography, alongside the utilization of artificial intelligence and machine learning, which is around 13.2%, specifically for dynamic threat detection. Furthermore, there are few technologies which require attention are federated learning, secure key management, and secure communication mechanisms, as these technologies are less considered in literature, i.e. around 3% only. Finally, the analysis underscored the necessity for real-time security monitoring and adaptive threat response to manage the dynamic nature of fog computing environments effectively.