A Novel Interval Caching Strategy for Video-on-Demand Systems

In a Video-on-Demand (VOD) system, in order to guarantee smooth playback of a video stream, sufficient resources (such as I/O bandwidth, network bandwidth) have to be reserved in advance. Given limited resources, the number of simultaneous streams can be supported by a video server is restricted. Due to the mechanical nature, the I/O subsystem is generally the performance bottleneck of a VOD system, and there have been a number of caching algorithms to overcome the disk bandwidth limitation. In this paper, we propose a novel caching strategy, referred to as client-assisted interval caching (CIC) scheme, to balance the requirements of I/O bandwidth and cache capacity, in a cost-effective way. The CIC scheme tends to use the cache memory available in clients so as to dramatically reduce the demand on the I/O bandwidth of the server. Our objective is to maximize the number of requests that can be supported by the system and minimize the overall system cost. Experimental results show the superior of CIC scheme to the traditional Interval Caching (IC) scheme, with respect to request accepted ratio and average servicing cost per stream.