计算网格中动态负载平衡策略研究

英文篇名：Study Dynamic Load Balancing Strategy in Computing Grid
中文刊名：微电子学与计算机
英文刊名：Microelectronics & Computer
作者：亓雪冬 ; 李霞 ; 梁鸿 ; 仝兆岐
英文作者：QI Xue-dong1 ; LI Xia2 ; LIANG Hong1 ; TONG Zhao-qi1(1 College of Comp and Comm Engineering ; China University of Petroleum ; Dongying 257062 ; China ; 2 College of Information and Control Engineering ; China University of Petroleum ; Dongying 257062 ; China)
中文关键词：计算网格 ; 负载平衡 ; 网格信息服务 ; 地震正演模拟
英文关键词：computing grid ; load balancing ; grid information service ; seismic forward modeling
出版日期：2009-05-05
机构：中国石油大学计算机与通信工程学院;中国石油大学信息与控制工程学院;
年：2009
期：05
出版单位：微电子学与计算机

摘要

负载平衡调度策略对于网格计算系统非常重要,有利于网格子任务和网格资源间作出更合理的匹配,起到优化总计算时间的目的.研究了通用的集中式动态负载平衡策略,基于Globus的网格信息获取和任务分配模型,网格节点硬件类、软件类及文件类信息的获取方法,设计并实现了网格动态负载平衡算法,继而应用该算法在网格中进行了地震波正演模拟实验.实验结果表明:该算法充分利用了网格中每个节点的资源,有效地减少计算密集型任务的求解时间.
Load balancing schedule strategy is very important for grid computing system.The strategy can hand out subtasks to each grid node well-balanced so this strategy plays a role of reducing cumulative computing time.This paper first studied general-purpose dynamic load balancing strategy;the model of information acquirement and task task allocation in grid based on globus;the method of getting grid nodes' parameters of hardware,software and file system.Then grid-specific dynamic load balancing algorithm was proposed and applied in a seismic wave field forward simulation experiment.The results show this algorithm takes full advantage of each grid node's resources and reduces computing time of compute-intensive tasks.

引文

[1]Foster I,Kessel man C,Tuecke S.The anatomy of the grid:enabling scalable virtual organizations[J].Interna-tional Journal of Supercomputer Applications,2001,15(3):200-222.
    [2]Foster I,Kessel man C,Nick J,et al.The physiology of the grid:an open grid services architecture for distributed systems integration[R].Open Grid Services Architecture WG,Global Grid Forum,2002.
    [3]Foster Ian.Globus toolkit version4:software for service-oriented systems[J].Journal of Computer Science and Technology,2006,21(4):513-520.
    [4]Foster I,Kessel man C,NickJ M,et al.The physiology of the grid[M].UK:John Wiley&Sons,2003.
    [5]邹德清,金海,吴松,等.面向网格的协作式网络计算平台[J].计算机学报,2004(12):1889-1893.
    [6]亓雪冬,仝兆岐,何潮观.基于网格的同步计算划分策略研究[C]//全国第18届计算机技术与应用(CACIS)学术会议.浙江:宁波,2007.
    [7]Barry Wilkinson,Michael Alen.Parallel programming[M].北京:机械工业出版社,2005:151-167.
    [8]韩文功,印兴耀,李振春,等.地震技术新进展[M].山东:中国石油大学出版社,2006:73-125.