Efficient Cloud Gaming Resource Provision Via Multi-dimensional Bin-Packing

Abstract

In order to enable the acceptable level of service quality for cloud gaming services, sufficient resources should be always maintained and optimal resource management is then necessary. However, taking only a limited set of server-related parameters, but ignoring the client-related parameters, in the typical formulation of optimization problem cannot yield for optimal resource allocation in the cloud gaming environment nowadays, where the game server’s computing processors can be driven by both the CPU and GPU, and the client devices’ display resolutions are largely heterogeneous. In this paper, we describe how better cloud gaming resource utilization can be achieved through a formulation of Multi-dimensional bin-packing optimization problem. Based on the experiment results, our proposed mechanism looks promising for realistic cloud gaming services, where the adaptive feature must be taken as a prime consideration for efficient cloud gaming resource management.