基于负载感知和QoS的多中心作业调度算法  

A Job Scheduling Algorithm in Multi-computing Centers Based on Load-aware and QoS

作　　者：冯鸣夏 伍卫国[1] 邸德海[2] FENG Ming-xia;WU Wei-guo;DI De-hai(School of Electronic and Information Engineering,Xi’an Jiaotong University,Xi’an 710049,China;School of Management,Xi’an Jiaotong University,Xi’an 710049,China)

机构地区：[1]西安交通大学电子与信息工程学院,陕西西安710049 [2]西安交通大学管理学院,陕西西安710049

出　　处：《计算机技术与发展》2018年第12期1-6,共6页Computer Technology and Development

基　　金：国家重点研发计划(2016YFB0201402);国家自然科学基金(91630206);海外及港澳学者合作研究基金(61628210)

摘　　要：在国家高性能计算环境中,针对分布在不同地域的计算中心存在计算资源忙闲不均,缺少统一调度和管理的问题,提出了一种基于负载感知和QoS的多中心作业调度算法(Laq)。首先根据各个计算中心的负载情况将需要处理的作业分配给负载较轻的计算中心,然后充分考虑到用户的QoS需求,将作业尽量调度到符合用户QoS需求的计算中心,从而提高计算中心的服务水平。实验结果表明,与Random和RoundRobin算法相比,在Laq算法的作业调度策略下,多计算中心作业的完成时间、平均响应时间、平均等待时间和slowdown等性能指标都显著减少。各个计算中心保持了计算资源负载的相对均衡,提高了资源利用率。In order to solve the problem of uniform scheduling and load imbalance in the national high-performance computing environment,we propose a job scheduling algorithm in multi-computing centers based on load-aware and QoS(Laq).First,the job that needs to be processed is assigned to the lighter computing center according to the load of each computing center.Then,the QoS requirements of users are fully taken into account,the job is dispatched to the computing center that conforms to the user’s QoS requirements to improve the service level of the computing center.The experiment shows that the performance indicators,such as the makespan,the average response time,the average waiting time are significantly reduced in comparison with the Random and the RoundRobin.Each computing center maintains the relative balance of computing resource load and improves resource utilization with Laq.

关 键 词：高性能计算 负载感知 QOS 作业调度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]