电信科学 ›› 2013, Vol. 29 ›› Issue (12): 87-91.doi: 10.3969/j.issn.1000-0801.2013.12.014

• 云计算专栏 • 上一篇    下一篇

截止期限约束的实例密集型云服务流调度算法

刘炜,李陶深,黄汝维   

  1. 广西大学计算机与电子信息学院 南宁530004
  • 出版日期:2013-12-20 发布日期:2017-07-04
  • 基金资助:
    国家自然科学基金资助项目;广西自然科学青年基金资助项目

A Deadline-Constrained Scheduling Algorithm for Intensive-Instance Serviceflows in Cloud Environment

Wei Liu,Taoshen Li,Ruwei Huang   

  1. School of Computer, Electronics and Information, Guangxi University, Nanning 530004, China
  • Online:2013-12-20 Published:2017-07-04

摘要:

针对商业云计算中存在大量实例密集型服务流的问题,提出一种新的云环境下两阶段服务流调度算法。该算法先将用户自定义的全局截止期限分配到系统中的每个实例,再将每个实例的截止期限分配到实例中的每个任务中,最后在服务流执行阶段,动态调整后续任务的截止期限,解决了任务可能存在的未能在其截止期限内完成的时间异常问题。CloudSim仿真结果表明,与现有的算法相比,该算法能满足用户定义的截止期限,节约了执行成本,并减少了资源的竞争率,提高了调度的成功率。

关键词: 云计算, 服务流调度, 时间异常, 截止期限

Abstract:

The problem that there are many intensive-instance serviceflows in business cloud computing environment was aimed to solved, so a new two-phase serviceflows scheduling algorithm was proposed. In the first phase, the global deadline to each instance was assigned, and then the deadline to the each task of instances was assigned. In the serviceflows execution phase, in order to solve the possible problem of time exception which task was not meet its sub-deadline, the algorithm adjusted deadlines of subsequent task dynamically. The experimental results in CloudSim show that the algorithm compared with others algorithm saves the execution cost with meeting the user-defined deadline, avoids the fierce competition of cheaper services and improves the success ratio of the scheduling.

Key words: cloud computing, serviceflows scheduling, ime exception, deadline

No Suggested Reading articles found!