Deep Reinforcement Learning (DRL)-based Transcoder Selection for Blockchain-Enabled Video Streaming

Video transcoding has been widely applied in video streaming industry to convert videos into multiple formats for heterogeneous users. To provide fast and reliable transcoding services, some emerging platforms are leveraging blockchain and smart contract technology to build decentralized transcoding hubs with flexible monetization mechanisms, where any users can rent their idle computing resources to become transcoders in order to receive reward. On these blockchain-enabled platforms, transcoder selection together with resource allocation is a crucial but challenging issue, which should maximize the transcoding revenue to motivate the transcoder candidates as well as handling the dynamic quality of service (QoS) requirements and candidates' characteristics. In this paper, we propose a novel deep reinforcement learning (DRL)-based transcoder selection framework for blockchain-enabled video streaming. Specifically, we propose an evaluation mechanism to facilitate transcoder selection, and design the transcoder selection and resource allocation scheme using the DRL approach. Simulation results demonstrate the effectiveness of our proposed framework.