فرهاد پاکدامن

In recent years, cloud gaming (CG)'s market share has grown in both academic research and adoption by the industry. Despite its advantages, CG has its own challenges. Delay has one of the most detrimental effects on the QoE of players. This is more intense in multiplayer cloud gaming (MCG), where players interact, and the game conditions should be the same for all players to be able to play fairly. Hence, network resources should be allocated to players based on their requirements and current network conditions, so that inter-delays of interacting players would be relatively close. To remedy this and to improve the QoE of CG, this paper combines the two concepts of edge computing and software-defined networks (SDN). Edge Computing brings servers closer to clients to decrease end-user delay. SDN enables dynamic management of network, based on user requirements and network states. We propose a suboptimal dynamic server and path selection method that allows players to achieve their required delay and bandwidth. Moreover, this method minimizes the inter-player delays for players in the same group. Experimental results show that our resource provisioning reduces inter-player delay and total delay by 94%, 31%, respectively, compared to competing methods.