Follow-me Prefetching for Video Streaming Over Mobile Edge Computing Networks

摘要

Mobile video streaming services have increased exponentially in recent years due to the popularity of mobile devices, the advancement of mobile networks and the availability of a variety of video contents over the Internet. Mobile Edge Computing (MEC), in connection with the backend cloud computing, has been used to bring contents close to the end user in order to reduce transmission latency. However, the quality of video streaming services suffers from degradation when an end user moves from the coverage of one node to another or when the condition of a mobile network degrades. In this paper, we propose a novel follow-me Edge Node Prefetching (ENP) scheme to prefetch appropriate video segments in advance in the following-on mobile node to avoid video quality degradation during video streaming. We set up a test bed consisting of a back-end cloud (OpenStack), two edge nodes (LXD Containers) and a mobile device, and implemented the ENP algorithms on the cloud server and client sides. Extensive experiments for Dynamic Adaptive Video Streaming over HTTP (DASH) services were carried out based on dash.js from the DASH Industry Forum. Preliminary results show that the ENP scheme can achieve better video quality (in terms of provisioning of the average video bitrate per segment) and less service migration time between mobile nodes when compared with existing approaches. The scheme might be useful in supporting video streaming services over MEC, and/or in future smart city applications.