流媒体系统中基于请求迁移的任务调度算法  

Request migration based task scheduling algorithm in Vo D system

作　　者：李军[1,2] 倪宏 王玲芳[1] 陈君[1] 

机构地区：[1]中国科学院声学研究所国家网络新媒体工程技术研究中心,北京100190 [2]中国科学院大学,北京100190

出　　处：《吉林大学学报（工学版）》2015年第3期938-945,共8页Journal of Jilin University:Engineering and Technology Edition

基　　金：'863'国家高技术研究发展计划项目(2012AA011703);国家科技支撑计划项目(2012BAH02B01);中科院重点部署项目(KGZD-EW-103-4;KGZD-EW-103-2)

摘　　要：提出了基于请求迁移的任务调度策略(Request migration based task scheduling,RMTS)算法,当有新请求到达系统时,运行RMTS算法的调度器根据当前负载状态和迁移路径长度决定是否对请求进行迁移,并选择最优迁移路径。当某个媒体服务器出现故障无法提供服务时,调度器将该服务器正在服务的请求迁移到其他存储有相应内容的媒体服务器以保障点播服务不间断。采用请求成功率和服务时延度量系统的性能,仿真结果表明:RMTS算法优于传统迁移算法和REM(Random early migration)算法,服务时延和请求成功率分别提高了14%和15%。A Request Migration based Task Scheduling algorithm （RMTS） is proposed. When a new request is received, the scheduler running RMTS calculates a probability based on the current system load and request migration path length to determine whether migration is needed. If migration is needed, the optimal migration path is selected. With one video server failure, the scheduler transfers all the requests in service to another server to ensure service uninterrupted. The request success rate and service delay are used to measure the system performance. Experiment results demonstrate that the proposed RMTS performs better than traditional request migration algorithm that, with RMTS, the request success rate and service delay are improved by 14% and 15% respectively.

关 键 词：信息处理技术 任务调度 请求迁移 服务不间断 请求成功率 服务时延 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]