Gaurav Aggarwal, Gauarv Kumar, Divyashikha Sethia
Abstract:
Recently, there has been a tremendous growth in accessing media content like video for both mobile and computer platforms. This trend has resulted in increase in bandwidth requirement on client and server side. Many solutions like Caching video have been suggested to reduce the backend bandwidth requirement, but none have been efficient enough to be implemented in real scenario. In this paper, we propose a Dynamic Prefix Caching (DPC) algorithm with Lazy Update mechanism that utilizes past request patterns of video to compute the length of Prefix that should be cached on the proxy server. Our primary objective is to cache a video from starting till a calculated point, depending upon its popularity i.e. for a more popular video a larger Prefix length video will be cached on the proxy server. As the popularity of the video will change, the length of the cached video Prefix will be updated. Each time we are about to exhaust allotted Caching space, criteria of popularity is made more stringent. The experimental results indicate significant improvement in bandwidth requirements and client waiting time.
Publication:
“Dynamic Prefix Caching of Videos with Lazy Update” has been accepted in the conference of CCSEIT- 2012 to be held in Coimbatore, India, October 26 ~ 28, 2012 and proceedings will be published by ACM- ACM International Conference Proceeding Series (ICPS) (Confirmed). ACM International Conference Proceeding Series (ICPS) will be available in ACM Digital Library.