Player Assignment in MEC Gaming for Social Interactivity and Server Provisioning Optimization

In Mobile Edge Computing (MEC), edge servers are deployed on Base Stations (BSs) to provide high accessibility to nearby users, alleviating the computational burden of the remote cloud and significantly reducing network delays. Still, for distributed interactive applications (DIAs), such as cloud games, where players team up and form social communities, one fundamental challenge is the discovery of the most suitable player-to-server assignment, that achieves a balanced tradeoff between low-delay social interactivity and low-cost server provisioning, under strict proximity and capacity constraints. In this paper, we formally model the problem via mixed-integer linear programming. However, its solution is complex falling under the category of existing capacitated facility location problems. Thus, we present the “Interactivity-based Edge Server Allocation” (IESA) algorithm, a novel heuristic approach to efficiently assign and then iteratively improve the player-to-server allocation. IESA is evaluated using trace-driven simulations that demonstrate how it outperforms the baseline and state-of-the-art assignment alternatives.