Caching policy assisted multicast patching for video streaming on-demand server

Video streaming on-demand servers are characterized by stringent real-time constraints, as each video stream requires isochronous data playout. With heavily skewed access patterns to the stored video content, the VOD system is often disk arm-bound. In such cases, serving video streams from a memory cache can result in a substantial reduction in server cost. With the falling prices of the memory, buffering and caching becomes an efficient method to reduce the required disk bandwidth. Most of the existing caching policies for video streaming on-demand server are designed in the context of unicast; however, using unicast will overload both network and server. Caching policies assisted with multicast will be more efficient and more scalable. Thus almost all extant caching algorithms are not suitable to be applied in actual multicast systems. In this paper we advance a new caching policy, which has fully combined the characteristics of an efficient multicast algorithm, for VOD server. This caching policy is very efficient, owing to a novel concept of entity and a unique caching gain function.