Deployment of Multi-Tier Fog Computing System for IoT Services in Smart City

Abstract

One of the assuring technologies today is Fog computing, focus on the extensive use of the computational and storage capacities of the end devices in a network. In present day there due to the enormous amount of data from the number of IoT devices, a centralized cloud system is quite inadequate. This challenge can be addressed by deploying Fog devices neighbouring to these IoT devices so as to provide with real-time response. Thus for the creation of a smart city we advance towards an architecture in which the cloud data centre at the top followed by SDN controllers, fog controllers, fog devices and smart sensors. We propose an integer programming model in our problem formulation of deploying the fog nodes, fog controllers, SDN controllers which results in minimization of latency, traffic and cost with constraints such as device capacity, offloading workload, range etc. Further on, our work solve this NP-hard problem by weighted sum method and the two meta-heuristic algorithms Genetic Algorithm (GA), Particle Swarm Optimization (PSO) compared to Randomized Algorithm. Thus a network planner with a cost efficient fog network in mind can relate to the simulations illustrated in the paper for their existing computational and storage configuration. We verify our proposed model and algorithms through simulation which help to design efficient fog network.