Admission Control And Caching Scheme In Streaming Media System

Video-on-Demand system has gained much attention since its first appearance and has become the hot spot in the field of multimedia.With the development of network and computer technique,computing,storage and network bandwidth are much cheaper these days.However,designing a large scale VOD system is still a challenge,including challenges to the video server architecture,service strategy,transmission and storing techniques.Solving these problems is important and has practical significance.The work in this dissertation is mainly about important issues that affect the service ability of the VOD system,including admission control algorithm in streaming clusters and caching scheme in streaming servers.For each kind of issues,this dissertation gives an analytical model to describe the problems and finally proposed a corresponding policy.The main work is as follows:1.we propose a CAC policy called Call-rate-based Dynamic Threshold Admission Cotrol Policy(CDT) for a video-on-demand system with shared storage subsystem. The policy iteratively estimates the call rate,which improves its adaptive learning capability.A fast algorithm for searching the optimal threshold is also presented,and it reduces the computation complexity of the admission control policy.The results of simulation show that CDT can effectively improve system utility compared to other policies.2.Based on discussing the behavior of existing multimedia streaming caching schemes, a new cache management algorithm AHIC(Adaptive Hybrid Interval Cache) using the classic Interval Cache scheme was proposed.The new scheme considers the influence of user activities as well as the popularity of multimedia objects, proposed a hybrid scheme which blend some speciality of segment cache and interval cache.Caching admission and replacement algorithm based on the segment popularity estimated periodically was implemented.The results of simulation show that AHIC scheme can effectively improve the byte hit ratio of multimedia streaming systems compared to the traditional interval cache policy.3.In distributed streaming media server clusters,by adopting cooperative caching technique,the free cache memory of all the servers in the cluster can be used more efficiently.It will help raise the hit rate of the cache and reduce disk accesses, resulting in the improvement of the overall throughput of cluster.In this dissertation, a novel distributed caching model called Scalable and Cooperative Caching (SCC) is presented.SCC makes use of each server to perform the caching task and redirects the request to the proper server in the cluster to make the cache scheme more efficient.This way,SCC can provide high scalability and throughput in clustered servers with a reasonably small amount of communication among hosts.