Multi-SP Profit Maximization and Game Frame Time Reduction Scheme for Multi -Access Edge

Multiple Service Providers (SPs) may operate over an area. When one of them does not have enough resources, it may use resources of another instead of cloud resources to reduce expenditure and latency. We consider the scenario of a multi-player AR/VR game, where an area is divided into regions. The task taking user inputs is compute-intensive and latency-sensitive and may be offloaded to the server at the edge. The service processing the task output and generating action data is also hosted at the edge. A Region Coordinator (RC) generates a common response to all the players by processing the action data of players belonging to its region. The Game Frame Time (GFT) is the maximum total time of all players, from the time a player gives its input to when it receives the response from its RC. We formulate the problem of maximizing the profit of each SP while placing a constraint on the GFT and considering the task migration and service migration costs due to the mobility of players, as an optimization problem. We implement and evaluate the problem and compare it with a scenario where players do not move.