论文部分内容阅读
随着计算机技术和网络技术的不断进步,视频点播服务已经逐渐变成现实,基于协作缓存的视频点播系统是一个分布式结构,中心集群存放影片数据,本地集群缓存数据并提供视频点播服务,具有很好的可扩展性。该文针对VOD系统的协作缓存,提出了静态调度和动态迁移相结合的调度策略,静态调度能够根据影片数据的缓存分布情况,实时调度服务请求,并同时考虑多个VOD服务器的负载动态平衡;动态任务迁移能对服务流分布进行实时分析,并根据服务数据本地化的原则进行服务迁移,进一步提高了协作缓存的命中率。该文阐述了基于协作缓存的视频点播系统的拓扑结构,对静态调度和动态迁移进行了详细设计,并给出了相应的形式化表示。
With the continuous improvement of computer technology and network technology, video on demand service has gradually become a reality. The video on demand system based on collaborative caching is a distributed architecture. The central cluster stores the video data. The local cluster caches the data and provides the video on demand service. Very good scalability. According to the caching strategy of VOD system, this paper proposes a scheduling strategy based on static scheduling and dynamic migration. Static scheduling can dispatch service requests in real time according to the caching distribution of video data. At the same time, the load balancing of multiple VOD servers is considered. Dynamic task migration enables real-time analysis of service flow distribution and service migration based on the principle of service data localization, further improving the hit rate of the collaboration cache. This paper describes the topological structure of video on demand system based on collaborative caching, details the static scheduling and dynamic migration, and gives the corresponding formalized representation.