论文部分内容阅读
协作式视频缓存系统是近年来兴起的一种视频点播服务解决方案.本文针对协作式缓存服务系统中的请求调度提出了一种基于时间片的统一优化模型,并就最小化延迟和最小化网络代价两个优化目标分别提出了分布式的调度算法.新的调度算法根据当前时间片的链路负载以及实际的用户请求来选择目标服务器,加之对用户请求以数据率为调度单位,从而大幅提高系统的响应度和鲁棒性.仿真实验结果表明,提出的调度方案较之已有的协作式调度机制,在最大链路拥塞率、平均延迟、平均代价以及系统吞吐率等方面都有很明显的改进.
Cooperative video caching system is a video-on-demand service solution that has emerged in recent years.This paper proposes a unified time-based optimization model for request scheduling in collaborative caching service system, and in the point of minimizing the delay and minimizing the network Cost two optimization targets respectively proposed distributed scheduling algorithm.The new scheduling algorithm based on the link load of the current time slice and the actual user request to select the target server, combined with the user request data rate for the dispatch unit, which greatly increased System responsiveness and robustness.The simulation results show that the proposed scheduling scheme has obvious advantages in the maximum link congestion rate, average delay, average cost and system throughput compared with the existing cooperative scheduling mechanism improvement of.