Cloud Strife: Expanding the horizons of Cloud Gaming Services

Abstract

Cloud gaming where computer games are executed on cloud based resources is a new service which can be extremely challenging for service providers. Unlike other web and video streaming applications it requires both a relatively high constant downlink bandwidth and low latency simultaneously. Hence, cloud gaming providers recommend that clients be relatively close to data centers to ensure reasonable Quality of Service. Latency, however, is also dependent on the compression type used for the video streaming and there is a trade-off between compression efficiency and complexity which affects client computer capability. In this paper we propose the Strife system which examines the processing capabilities of the client as well as the network conditions between the client and data center to select the compression type that maximizes the size of the client base for cloud gaming services for which reasonable Quality of Service can be provided. We model a cloud gaming service using data which reflects gaming hardware in common use, the bandwidth available in the regions and optimize the number of users that can receive reasonable Quality of Service. Finally, we compare other schemes with the Strife system to evaluate its performance.