Evaluation of a location coverage model for mobile edge computing

The Mobile Edge Computing paradigm shifts the computation back to places where it is required. A traditional MEC architecture comprises a number of Edge Data Centers (EDC) in charge of seamlessly providing services to users with wireless network technologies. In this scenario, it becomes crucial to deploy the EDCs in strategic locations, such as highly visited places. In this paper we focus on the deployment phase of an EDC. In particular, we propose a probabilistic model designed to measure the location converge, namely the probability that a candidate location for an EDC is visited by users. Our model is based on the analysis of user's trajectories and on the probability of detouring towards the target locations for the EDS. The information returned by our model offers the possibility of implementing mobility-aware deployment strategies in urban environments. We test the model with two real-world mobility data sets, evaluating its applicability of realistic settings.