Fog Node Optimum Placement and Configuration Technique for VANETs

Abstract

Intelligent Transportation Systems (ITS) are nowadays considered very important applications of smart cities. One of the most important technologies that are utilized to support ITS is Vehicular Ad-hoc Networks (VANETs). In VANETs, vehicles communicate with each other (V2V) or with the infrastructure (Roadside Units) (V2I). Roadside Units (RSUs) collect data from vehicles in the coverage area and send it to cloud servers through the Internet. Cloud servers have high performance computational and storage capabilities that ITS applications require for data processing. However, due to the real-time requirements of the ITS applications, cloud approach alone cannot be guaranteed to satisfy the strict time constraints due to long latency access of the centralized cloud server. Fog Computing is an emerging approach that extends the services of cloud computing to the edge of the network. Fog Computing can be utilized in VANETs through deployment of fog nodes into RSUs. One of the major challenges is identifying the optimum number, locations and computational capabilities of the RSUs particularly in urban regions where obstacles exist heavily inside the coverage area of the RSUs. In this paper, we consider the optimization problem of fog-based RSU placement where the objective is to maximize the achieved level of service quality in a cost-effective way. The problem is formulated as a Satisfiability Modulo Theories (SMT) problem and solved using Microsoft Z3. The proposed approach is able to generate a set of solutions as Pareto front. We obtained data from OpenStreetMap for Cairo city. Our approach outperforms other solutions in the literature in terms of cost.