Files

Abstract

Video streaming is one of the increasingly popular, as well as de- manding, applications on smartphones today. In this paper, we con- sider a group of smartphone users, within proximity of each other, who are interested in watching the same video from the Internet at the same time. The common practice today is that each user down- loads the video independently using her own cellular connection, which often leads to poor quality. We design, implement, and evaluate a novel system, MicroCast, that uses the resources on all smartphones of the group in a co- operative way so as to improve the streaming experience. Each phone uses simultaneously two network interfaces: the cellular to connect to the video server and the WiFi to connect to the rest of the group. Key ingredients of our design include the follow- ing. First, we propose a scheduling algorithm, MicroDownload, that decides which parts of the video each phone should download from the server, based on the phones’ download rate. Second, we propose a novel all-to-all local dissemination scheme, MicroNC- P2, for sharing content among group members, which outperforms state-of-the-art peer-to-peer schemes in our setting. MicroNC-P2 is designed to exploit WiFi overhearing and network coding, based on a local packet broadcast framework, MicroBroadcast, which we developed specifically for Android phones. We evaluate MicroCast on a testbed consisting of seven Android phones, and we show that it brings significant performance benefits without battery penalty.

Details

Actions

Preview