Fairness-Oriented Edge Allocation for Interactive Group Gaming in Edge Computing

In recent years, edge computing has emerged as a highly promising paradigm for the extension of cloud multimedia systems, enabling network latency reduction by pushing services and data from remote clouds to edge servers. For immersive distributed interactive applications, such as virtual reality and multiplayer online games, this shift has revolutionized the way nearby mobile users interact with one another to jointly participate in collaborative virtual events. However, because the virtual world is shared among many users, one fundamental decision problem for the game providers relates to: (i) how the servers will be deployed, given the limited budget in edge infrastructure investment; and, subsequently, (ii) in what manner the users will be assigned to maintain group fairness across all participating parties, given the lag variance in game view consistency. In this paper, we call this the “Fairness-oriented Edge Allocation” (FEA) problem, and formally formulate its properties and discuss its complexity. To provide solutions efficiently, we present a heuristic algorithm, namely the h-FEA, and theoretically analyze its time complexity. We, then, evaluate h-FEA against several benchmark alternatives using comprehensive simulations on a real-world topological trace. The results clearly demonstrate, under different scenarios, the superiority of our proposed algorithm in minimizing the fairness loss for users engaged with the same games, while achieving high edge user admission rates.