摘要
随着越来越多云服务提供商的出现,通过在多个云之间进行任务调度降低总执行时间,从而降低代价,使得云计算更高效和快速反应.为此,本文引入科学工作流的机制,在满足科学工作流的截止日期下,尽可能地减少总时间,即执行代价.首先,根据科学工作流自身的结构特点,尽可能地合并存在有向割边的任务,使所有符合这些条件的任务合并为一个任务,减少总的任务数,而这也就相当于减少了总时间.然后,不断把每个关键父任务作为下次的输入继续寻找关键父任务,直至所有路径覆盖全部任务,即排列好所有有前后关系的任务,使所有任务依次执行.最后,把每个局部关键路径分配到最适合的云上,并执行任务,获得每一个任务的实际结束时间.实验结果表明本文算法的有效性.
With more and more cloud service providers,it can reduce the total time through tasks across multiple cloud scheduling, in order to reduce the cost of cloud computing makes more efficient and rapid response. The algorithm described in this paper is to introduce the mechanism of scientific workflow to minimize the total time, i.e., the execution cost, at the same time, should meet the deadline of scientific workflow. First of all, according to the structure of scientific workflow, the tasks whach mect with all these conditions are combined as a task to reduce the total number of tasks, which is equivalent to a reduction of the total time. Then, based on the iteration mechanism of the critical parent, the local key path is searched repeatedly until all the paths are covered by all the tasks, that is, all the pre and post relations are arranged, so that all the tasks are executed in turn. Finally, the distribution of each local critical path to the most suitable cloud, and perform tasks, obtain actual time of each task. The experimental results show the effectiveness of the proposed algorithm.
作者
徐赛娟
刘耿耿
XU Saijuan;LIU Genggeng(Department of Information Engineering,Fujian Business University,Fuzhou,Fujian 350012,China;College of Mathematics and Computer Sciences,Fuzhou University,Fuzhou,Fujian 350116,China;Key Laboratory of Fujian Network Computing and Intelligent Information Processing,Fuzhou,Fujian 350116,China)
出处
《闽江学院学报》
2018年第5期63-72,共10页
Journal of Minjiang University
基金
国家自然科学基金项目(11501114)
关键词
云计算
多云环境
科学工作流
局部关键路径
cloud computing；multi-cloud environment；scientific workflow；partial critical path
