Joint adaptive rate and scheduling for video streaming in multi-cell cellular wireless networks

We consider adaptive-rate scheduling for downlink unicast transmissions of video streams over cellular wireless networks. We study a service under which each mobile client receives requested video streams at a variable Quality of Experi- ence (QoE) level, based on its experienced communication quality condition. We employ a proxy-video manager at the base station node. The manager classifies users into two groups, based on their reported experienced CQIs (Channel Quality Indicators). The manager intercepts a client’s video stream request destined to the HTTP server, determines its group classification, and proceeds to transmit a properly encoded version of the requested stream. To effectively regulate inter-cell signal interference, we examine a number of different spectral reuse and fractional frequency reuse scheduling schemes. We calculate, for each scheduling and user group classification scheme, the average bandwidth per stream that is required to provide user groups with their targeted QoE levels. We demonstrate a user classification process that aims to optimize a utility metric based on the definition of a will-to- pay utility function. The derived optimal configuration of FFR based schemes are shown to significantly enhance the system’s performance behavior.