| With the improvement of living standards, people are not limited to the web browser. The video, audio and other large media resources appear gradually, and now video on demand, time shifted TV are extensively spread. Previous methods for web caching have gradually ceased to apply, how to design a cache to improve the media files'Byte Hit Ratio, reduce the server's disk I / O, improve server performance, reduce users'start delay are the focus at this stage of research.In this paper, we study from the video-on-demand service. Our aim is to research and study how to find the common caching strategy or the segmentation and jumping strategy to improve the cache byte hit ratio, reduce the user's wait delay or other related issues. These works mainly include the following three aspects:Firstly, detailly describe the background of video-on-demand, basic principles, system architecture, the future status and other related technologies; the core technology of the principle of streaming media, transfer basis, transmission categories, transfer agreements; also analyze the principle of cache, the classification of cache, and some of core values of the function in common cache.Secondly, design and propose a common caching replacement algorithm, on the one hand make the procedure uniform, easy to implement and error investigation, on the other hand migrate from one segment to another segment gain the greatest degree of recovery. Through revise the uniform and exponential segmentation strategy with the common replacement algorithm, it is easily to see common replacement algorithm can gain a better cache byte hit ratio and effectively reduce the disk I / O bandwidth.Thirdly, through dividing the basic segment and using jumping strategy, we propose flexible segmentation and jumping strategy. This strategy is not only full use of the cache space utilization, in most cases can reduce user's waiting time to zero in VCR operation (forward, rewind, jump, etc), achieving the maximum user's satisfaction. |
PDF Full Text Request