Performance Comparison of Different Cache-Replacement Policies for Video Distribution in CDN

Content Distribution Network(CDN) is one of the key components of the Internet nowadays. Many users, served by same ISP/node, request same data resulting in higher Internet traffic. Caching of data has been popular mechanism to avoid fetching the same data repeatedly from a source. There exist different caching algorithms and strategies and much of work has been done to measure their performance for general data. Although both CDN and cache replacement policies have been actively looked into by the research community, the interplay of them is still less well-understood. Moreover, performance comparison for the data as a video content has not been done much in depth. In this paper, we investigate the interaction between CDN and cache replacement algorithms for video/movie distribution in a CDN. Issues covered include 1)comparative impact of different cache replacement policies on a CDN and 2) the effect of cooperative caching versus hierarchical caching schemes. By extensive simulations, we observed that LFU is the most efficient cache-replacement algorithm and cooperative caching is better than hierarchical caching in terms of user experienced latency and CDN system cost.

[1]  Brad Cain,et al.  A Model for Content Internetworking (CDI) , 2003, RFC.

[2]  Peter Druschel,et al.  Measuring the Capacity of a Web Server , 1997, USENIX Symposium on Internet Technologies and Systems.

[3]  David Mosberger,et al.  httperf—a tool for measuring web server performance , 1998, PERV.