基于负载均衡的任务调度算法  

Algorithm for Tasks Scheduling Based on Load Balance

作　　者：张腊[1] 刘淑芬[1] 韩璐[1] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012

出　　处：《吉林大学学报（理学版）》2014年第4期769-772,共4页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金(批准号:60973041)

摘　　要：基于传统任务调度算法,通过在调度过程中增设服务器最佳期望序列及负载均衡指数,采用任务连接数约束服务器选择,解决了传统任务调度算法对于集群负载均衡性能的忽略问题.实验结果表明,该改进算法能达到较高负载均衡度,同时缩短任务完成时间,并提高了集群对于批量级任务的处理能力,从而达到了优化的目的.On the basis of traditional tasks scheduling algorithm, we established the optimal sequence of servers and load balance index for tasks in the process of scheduling. Meanwhile, we used a number of connections to realize the constraint of servers, which has improved the performance of load balancing that has ignored in the traditional task scheduling algorithm. The experimental results show that our improved algorithm can achieve the high degree of load balancing and shorten the task completion time. In addition, this algorithm improves the ability of batching task processing for server cluster, thereby achieving the goal of optimization.

关 键 词：负载均衡 任务调度 服务器集群 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]