A Cognitive Platform for Mobile Cloud Gaming

Abstract

Mobile cloud gaming provides a whole new service model for the video game industry to overcome the intrinsic restrictions of mobile devices and piracy issues. However, the diversity of end-user devices and frequent changes in network quality of service and cloud responses result in unstable Quality of Experience (QoE) for game players. A cognitive cloud gaming platform, which could overcome the above problem by learning about the game player's environment and adapting the cloud gaming service accordingly, does not currently exist. To fill this void, we design and implement a component-based gaming platform that supports click-and-play, intelligent resource allocation and partial offline execution, to provide cognitive capabilities across the cloud gaming system. Extensive experiments have been performed to show that intelligent partitioning leads to better system performance, such as overall latency.