Lower bound of mean user waiting/watching time ratio in wireless video streaming networks

We consider centralized wireless networks for users of video streaming services. If the network is congested, wireless channel resources can be insufficient to satisfy all user requests. In this case QoE with some users may heavily degrade (video playback may periodically interrupt due to rebuffering effect, or startup buffering can take an intolerably long time). These negative effects can be partly reduced by appropriate scheduling of radio resources of the network. In the paper we propose a lower bound for the mean user waiting time spent on startup buffering and rebufferings in all possible schedulers. Through asymptotic analysis and simulation, this bound is compared with the results of a few generic scheduling policies. The proposed bound can be used as a basic performance test for existing and new algorithms of radio resources management in wireless video streaming networks.