Greening the Video Transcoding Service with Low-Cost Hardware Transcoders

Abstract

Video transcoding plays a critical role in a video streaming service. Content owners and publishers need video transcoders to adapt their videos to different formats, bitrates, and qualities before streaming them to end users with the best quality of service. In this paper, we report our experience to develop and deploy VideoCoreCluster, a low-cost, highly efficient video transcoder cluster for live video streaming services. We implemented the video transcoder cluster with low-cost single board computers, specifically the Raspberry Pi Model B. The quality of the transcoded video delivered by our cluster is comparable with the best open source software-based video transcoder, and our video transcoders consume much less energy. We designed a scheduling algorithm based on priority and capacity so that the cluster manager can leverage the characteristics of adaptive bitrate video streaming technologies to provide a reliable and scalable service for the video streaming infrastructure. We have replaced the software-based transcoders for some TV channels in a live TV streaming service deployment on our university campus with this cluster.