Context-Driven IoT Management: A Step Closer Towards A Self-Governing IoT

Despite the adequate computing power available at the edge, fog, and cloud layers, IoT management is still performed very hastily.IoT actuators may make the hosting devices autonomous but not the IoT.IoT which are meant to be self-governing are still far from getting there.This paper, by proposing a contextdriven IoT management process, brings us a step closer to a self-governing IoT.IoT contexts are envelopes of data from various IoT devices that can be evaluated by devices or humans to recognize one or more events that call for actions that are performed by humans or which are automatically evoked.An IoT comprises operational contexts at the edge, functional contexts at the fog, and strategic contexts at the cloud.IoT contexts, whether they are written in terms of auditable deterministic scripts, or belief structures, are however always tainted with a great deal of uncertainty that, due to the presence of ambiguities and inconsistencies, cannot be managed in a Bayesian manner.We hence apply Dempster and Shafer theory to perform context-driven IoT management.In order to do so, IoT contexts are written into polynomial contexts in a hierarchic manner from cloud to edge.Context-driven IoT management is modeled as polynomials of strategic contexts that are written as polynomials of functional contexts which are in their turn written as polynomials of operational contexts for which belief structure are constructed based on edge data.This paper provides a simple and detailed example to demonstrate the working of our contextdriven IoT management model.