截止日期约束的云中资源调度成本优化算法

Deadline Constrained Cloud Resource Scheduling Cost Optimization Algorithm

随着越来越多的工作流应用程序部署在云端,如何在满足工作流截止期限约束的前提下优化资源调度成本成为一个热门研究领域。本文提出了一种截止日期约束的成本优化(CODC)算法。首先,合并工作流任务以减少不同实例之间的数据传输开销。其次,关注父任务和子任务对当前任务优先级的影响,并考虑任务的子截止日期未被满足的情况,以选择最早完成任务执行的实例。最后,在5种工作流上与现有算法进行对比。与3种对照算法相比,CODC算法具有更低的工作流执行成本。

As more and more workflow applications are deployed in the cloud,how to optimize resource scheduling cost while satisfying workflow deadline constraints is a popular research area.In this paper,we propose a cost optimization with deadline constraint(CODC)algorithm.Firstly,workflow tasks are merged to reduce the data transfer overhead between different instances.Secondly,paying attention to the impact of parent and subtasks on the priority of the current task,the case of unmet task sub-deadlines is considered to select the instance that completes task execution earliest.Finally,the experimental results are compared with the existing algorithms on five workflows.CODC algorithm has a lower workflow execution cost compared to other three comparison algorithms.