摘要
针对复杂网格环境下的作业调度,提出了一种新的网格调度模型和作业调度算法。首先,通过对分布式/并行计算环境和网格计算环境下调度算法的不同特点分析,提出了一种适合于复杂网格环境下的新的网格调度模型,模型由客户、节点、客户作业、队列和调度器构成;然后基于新的调度模型,提出了一种基于最小累计启动时间准则的网格作业调度算法。调度算法将到达处理器的下一个作业的累计到达时间与处理器的最小累计启动时间进行比较,如果前者大于后者,则把到达作业调度到属于其队列的处理器资源并执行作业处理;否则,作业必须在队列中等待。采用Turbo C平台实现的仿真结果表明,提出的调度模型和作业调度算法,不仅能够获得良好的调度效果,而且相比于现有的FCFS调度算法,能够减小完工时间。
In this paper,a new grid scheduling model and job scheduling algorithm are proposed in complex grid environment.Firstly,through the analysis to different characteristics of two types of scheduling algorithm used in distributed/parallel computing environment and grid computing environment respectively,a novel grid scheduling model to be suitable for complex grid environment is proposed.It consists of clients,nodes,client jobs,queues and schedulers.Secondly,on the basis of the new scheduling model,a scheduling algorithm for grid job based on minimum cumulative start time criterion is proposed.The scheduling algorithm compares the cumulative arrival time for the next job arriving with the minimum cumulative start time for the processor.If the former is greater than the latter,the arriving job is scheduled to the processor resources that belong to its queue,and the job processing is performed.Otherwise,the job has to wait in the queue.The simulation results implemented by using Turbo C platform show that the proposed model and job scheduling algorithm not only can obtain better scheduling effect,but compared with the existing FCFS scheduling algorithm,can reduce make span.
作者
高卫斌
柳晓龙
GAO Wei-bin(Ningde Vocational Technical College,Ningde Fujian 355000,China)
出处
《长春工程学院学报（自然科学版）》
2020年第4期103-108,共6页
Journal of Changchun Institute of Technology：Natural Sciences Edition
基金
2017年宁德市科技项目(NSK2017137)。
