Utility-Driven Joint Caching and Bitrate Allocation for Real-Time Immersive Videos

Abstract

Real-time immersive video demands high network bandwidth and low transmission delay. Limited communication resources make it time-consuming to deliver immersive videos in cloud service scenarios. To overcome this, we design a utility-driven JOint Caching and Bitrate allocation (JOCB) algorithm for the real-time immersive video to better utilize network and caching resources through the Mobile Edge Computing (MEC) technique. Firstly, we coin a concept, the unfreshness indicator, to reflect the obsolescence level of cached tiles in MEC. Secondly, we define the Quality of Immersive videos (QoI) to evaluate the users' experience, including content characteristics, unfreshness levels, and spatial and temporal quality loss. Thirdly, we formulate the system utility that increases effective quality at the cost of transmission loss. The utility optimization problem can be formulated as an integer programming problem and decomposed into the cache update subproblem and the viewing probability-based adaptive bitrate allocation subproblem, which are solved by the branch-and-bound algorithm and the greedy algorithm, respectively. We have implemented an immersive video transmission system to perform experiments. Both simulation and experimental results further imply that JOCB can achieve utility maximization through balancing the transmission cost and the QoI.