Towards cinematic internet video-on-demand

Video-on-demand (VoD) is increasingly popular with Internet users. It gives users greater choice and more control than live streaming or file downloading. Systems such as MSN Video and YouTube deliver content at low bitrates. This may suit short clips, but great films and 5-minute bloopers are as different as symphonies and jingles. For cinema, poor quality and high jitter are less acceptable. Combining user control with high bitrate is compelling, but technically challenging. VoD is expensive due to the load it places on video source servers. Many researchers have proposed using peer-to-peer (P2P) techniques to shift load from sources to peers (peer-assistance), yet none have implemented and deployed a system with the first purpose of openly and systematically evaluating this approach. To fill this void, we have built and deployed GridCast1. GridCast doubles the bitrates of current popular internet VoD systems, provides a full set of VCR2 operations, and employs peer-assistance to improve scalability and continuity. GridCast has been live on CERNET3 since May of 2006. In peak months, GridCast has served videos to approximately 23,000 users. From the beginning, we have gathered information to understand GridCast and improve its algorithms. This paper introduces and evaluates GridCast. In May of 2007, we deployed multivideo caching, a major change to the caching algorithms. This paper analyzes scalability and continuity before and after this change. Our results contain several surprises and underline the importance of deployment to validate simulation results. We discuss what improvements can be developed beyond multivideo caching.