Stochastic Playback Delay Upper Bounds of Vehicular Video Content Delivery Networks with Cache-Enabled RSUs

Abstract

In the future vehicular communication scenarios, real-time video streams should be provided to drivers and passengers with significantly strict delay guarantees. With cache-enabled roadside units (RSUs), video contents can be obtained from local caches instead of remote servers, which can help to avoid large playback delay resulting from limited backhaul capacity. However, the existing works mainly focus on the average delay with cache-enabled RSUs in steady scenarios where the content delivery rates are assumed to be constant. In this paper, a stochastic network calculus (SNC) based method is proposed to obtain the stochastic playback delay upper bounds of vehicular video content delivery networks with cache-enabled RSUs. Consider the stochastic characteristics of wireless channels, the stochastic strict service curve of RSU to vehicle (R-V) channel, and equivalent base station (BS) to RSU to vehicle (B-R-V) channel are derived, respectively. With this framework, the stochastic playback delay upper bounds are obtained. Simulation results prove that the upper bounds are reasonably tight. Compared with the conventional RSUs, the superiority of cache-enabled RSUs to guarantee stochastic playback delay performance is validated.