摘要
针对工作流在公有云环境执行的费用过高问题,提出基于路径聚簇的费用优化调度算法。算法结合工作流的任务资源、任务依赖等特征,计算任务的最早开始时间、最早完成时间和优先级,聚簇最早完成时间和优先级高的任务进行统一放置,从而减少任务间的数据通信量,降低工作流的执行时间和费用,从而提高机构的执行效率。平台仿真显示改进后算法可有效地降低执行工作流的花费,提高执行工作流的综合性能比。
To solve outrageous cost of workflow execution in public cloud environment, proposes a workflow scheduling algorithm based on path clustering on the same path. Combining tasks' resources requirements and tasks dependence of workflow will be scheduled, computed earliest start time, earliest finish time and priority of tasks, clustered tasks have the highest earliest finish time and priority, thus reduces data traffic between tasks, shortens workf^ow makespan and reduce workfiow running cost, and improves organization efficiency. Platform emulation shows that the improved algorithm can effectively reduce cost of executing workflow, improve the comprehensive performance of workflow execution.
出处
《现代计算机》
2016年第2期8-12,共5页
Modern Computer
关键词
公有云
工作流调度
路径聚簇
费用优化
Public Cloud
Workflow Scheduling
Path Clustering
Cost Optimaztion