Fuzzy logic and Fog based Secure Architecture for Internet of Things (FLFSIoT)

Motivated by the recent explosion of interest around the Internet of Things (IoT), this paper is focused on the dynamics of its security. It is a fact that at present, there is a dearth of consolidated and methodical visions for assuring the security of IoT. This owes to its unique characteristics like interdependence of IoT devices and services, device heterogeneity, mobility, affinity to owners, pervasiveness, and unattended nature. The constraints of energy, memory, bandwidth and computability further complicate this picture. The strategies available in the literature today do not entirely consider this intricate nature of IoT environments. Moreover, uncertainty is not adequately acknowledged by the researchers while designing the security procedures for IoT, even though most of the IoT related security problems stem from an uncertain environment. This paper focuses on addressing this uncertainty, along with the other major security issues of IoT. It gives detailed categorization of the attacks in IoT. To this end, a Fuzzy Logic and Fog based Secure Architecture for IoT (FLFSIoT) has been proposed in this paper that works in real-time. In FLFSIoT, fuzzy logic has been used to alleviate the uncertainty of belonging to one crisp cluster of an edge node and for detecting various classical attacks. The Fog supported IoT architecture has been used to make FLFSIoT intrinsically more secure in comparison to the cloud-supported IoT by omitting the latency and other issues. As such, this work feeds the benefits of fuzzy log and fog computing into the carving of a generic solution to the IoT's security problem. The efficiency of FLFSIoT has been evaluated against the Distributed Denial of Service (DDoS) and Collusion attacks, and it has been observed that it gives more accurate results than the existing benchmarks.