论文成果
A DAG-based scheduling algorithm for dependent tasks in grid
- 点击次数：
- 论文类型：期刊论文
- 发表时间：2012-08-01
- 发表刊物：International Journal of Digital Content Technology and its Applications
- 收录刊物：EI、Scopus
- 文献类型：J
- 卷号：6
- 期号：15
- 页面范围：347-356
- ISSN号：19759339
- 摘要：Scheduling length can measure the performance of grid system. Minimizing scheduling length is important to enhance the performance of grid system. This paper proposes a dependent task priority-based scheduling algorithm (DTPBS). It can schedule task effectively and minimize the scheduling length(makespan). The application of grid is divided into many independent and dependent tasks, and these tasks can be described by a Directed Acyclic Graph (DAG). Firstly, DTPBS calculates the priority of each task according to the DAG, the task with high priority is scheduled first. And then DTPBS allocates the appropriate resource for each task based on the actual completion time (ACT). Moreover, the Load Balancing cannot be ignored. Experiments and performance analysis show that DTPBS has obvious superiority in minimizing scheduling length and Load Balancing compared with Min-min/Max-min algorithm and P-TSA algorithm based on random DAGs and DAGs which are used in the real world.
