首页> 外文会议>IEEE International Conference on Cloud Computing and Big Data Analysis >Fairness scheduling with dynamic priority for multi workflow on heterogeneous systems
【24h】

Fairness scheduling with dynamic priority for multi workflow on heterogeneous systems

机译:异构系统上具有多个工作流的动态优先级公平调度

获取原文

摘要

In Heterogeneous Computing Systems, completion time and overall fairness are both crucial to multi workflow scheduling. Unfortunately, most dynamic workflow scheduling algorithms fail to take fairness into account or adapt unreasonable fairness policy, which will probably result in some DAGs' scheduling failure in certain scenarios when they can't complete all the tasks before deadline. Aiming at the need to solve the flaws mentioned above, we propose a novel dynamic workflow scheduling algorithm named FSDP (Fairness Scheduling with Dynamic Priority for Multi Workflow). The algorithm focuses on the deadline of each workflow to achieve a more reasonable fairness when allocating the processors so that those urgent workflows can acquire a higher priority. Also, a new metric, urgency, is proposed to update priority. Experimental results show that it is possible to meet different requirements of dynamic workflows.
机译:在异构计算系统中,完成时间和整体公平对多工作流程调度至关重要。不幸的是,大多数动态工作流程调度算法未能考虑到公平或适应不合理的公平政策,这可能会导致某些方案中无法完成所有任务的某些方案的调度失败。旨在解决上述缺陷的需要,我们提出了一种名为FSDP的新颖动态工作流程调度算法(具有多工作流的动态优先级的公平性调度)。该算法侧重于每个工作流程的截止日期,以在分配处理器时实现更合理的公平性,以便这些紧急工作流程可以获得更高的优先级。此外,提出了一种新的度量标准,以更新优先级。实验结果表明,可以满足动态工作流的不同要求。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号