摘要
网格任务调度的目标是在最短的时间内将任务进行合理分配并实现负载均衡,最终得到实现全局的最优化调度方案。本文提出了一种兼顾任务之间的依赖关系和负载均衡的网格任务实时调度算法。首先,将网格分成若干个簇,每个簇由PC机、任务调度模块、任务控制器、数据集控制器以及数据存储器组成。簇负责接收用户提交的任务,将这些任务存放在本地任务队列中。通过对任务依赖关系的分析,将任务分配给网格的簇,并根据任务的截止期和依赖关系计算任务优先级,优先调度优先级高的任务。仿真实验表明,在Makespan和花费方面,本文算法都优于传统的Min-Min算法。
The goal of grid task scheduling is to achieve global optimization scheduling scheme by allocating tasks in the shortest time with load balancing.In this paper,we propose a real-time grid task scheduling algorithm considering task dependency and load balancing.Firstly,grid is divided into several clusters,and each cluster consists of PC,task scheduling module,task controller,data set of controller and data memory.Cluster is responsible for receiving tasks submitted by users,and these tasks are stored in the local task queue.After analyzing task dependency,tasks are assigned to cluster of grid according to task priority,which is computed by combining task deadline and task dependency.Simulation results show that in the aspect of Makespan and cost,the proposed algorithm is superior to the traditional Min-Min algorithm.
出处
《科技通报》
北大核心
2012年第12期131-133,共3页
Bulletin of Science and Technology
基金
2011年黑龙江省高等教育教学改革工程项目<基于网络环境下高校学生自主学习能力的培养与评价研究>的部分研究成果
关键词
网格计算
实时
任务调度
负载均衡
grid computing
real-time
task scheduling
load Balancing
