A generic and lightweight security mechanism for detecting malicious behavior in the uncertain Internet of Things using fuzzy logic- and fog-based approach

Inspired by the massive surge of interest in the Internet of Things (IoT), this work focuses on the kinetics of its security. By automating everything, starting from baby monitors to life-saving medical devices, IoT brought convenience to people's lives and rapidly became a trillion-dollar industry. However, the future of IoT will be decided on how its security and privacy concerns are dealt with. It is a fact that at present, the security of IoT is lacking in coherent and logical perspectives. For example, the researchers do not adequately accommodate the uncertainty and insider attacks while developing the IoT security procedures, even though most security concerns related to IoT arise from an insider and uncertain habitat. This paper provides a critical analysis of the most recent and relevant state-of-art methods of IoT security and identifies the parameters that are crucial for any security posture in IoT. Considering all the intricate details of IoT environments, this work proposes a Generic and Lightweight Security mechanism for detecting malicious behavior in the uncertain IoT using a Fuzzy Logic- and Fog-based approach (GLSF(2)IoT). It is developed on the principle of "zero trust," i.e., trust nothing and treat everything as hostile. While Fuzzy Logic has been used to remove uncertainties, the Fog-IoT architecture makes GLSF(2)IoT inherently better than the cloud-IoT. Once the malicious activity is detected, GLSF(2)IoT automatically limits the network access against the IoT device that initiated this activity, preventing it from targeting other devices. We evaluated GLSF(2)IoT for blackhole, selective forward, collusion and DDoS attacks, i.e., attacks which can invalidate any IoT architecture. Besides yielding better accuracy results than the existing benchmarks, we found that GLSF(2)IoT puts extremely low pressure on the constrained nodes, is scalable, supports heterogeneity, and uncertainty of the IoT environments.