通信学报 ›› 2013, Vol. 34 ›› Issue (6): 146-155.doi: 10.3969/j.issn.1000-436X.2013.06.018

• 技术报告 • 上一篇    下一篇

基于商空间的层次式数据网格资源调度算法

夏纯中1,2,宋顺林1   

  1. 1 江苏大学 计算机科学与通信工程学院,江苏 镇江 212013
    2 江苏大学 信息化中心,江苏 镇江 212013
  • 出版日期:2013-06-25 发布日期:2017-07-20
  • 基金资助:
    十一五国家科技支撑计划基金资助项目;国家自然科学基金资助项目;江苏省自然科学基金资助项目;江苏省普通高校研究生科研创新计划基金资助项目

Hierarchical data grid resource allocation based on quotient space theory

Chun-zhong XIA1,2,Shun-lin SONG1   

  1. 1 College of Computer Science and Communication Engineering,Jiangsu University,Zhenjiang 212013,China
    2 Information Center,Jiangsu University,Zhenjiang 212013,China)
  • Online:2013-06-25 Published:2017-07-20
  • Supported by:
    The Eleventh Five-Year National Science and Technology Support Program;The National Natural Science Foundation of China;The Natural Science Foundation of Jiangsu Province;Jiangsu College Graduate Research and Innovation Plan

摘要:

为了解决传统数据网格调度算法在对层次式数据网格调度过程中出现的极易陷入局部最优值和收敛速度过慢的问题,将粒计算的思想引入到网格调度中,提出了一种基于商空间的层次式数据网格资源调度QSHDGRA (quotient space theory based hierarchical data grid resource allocation)算法。首先分析了层次式数据网格的特点,接着提出一种基于业务请求平均等待时间和网络与节点资源利用均衡度的调和函数的调度问题模型,随后设计了基于商空间的层次式最优资源调度算法。该算法的特点是可以在不同粒度上由粗至细地对网格业务进行调度,从而保证不同业务的QoS,并实现系统全局最优资源分配。仿真实验表明,算法可以显著地提升系统整体的吞吐率,具有更快的收敛速度,并具备线性扩展能力。

关键词: 数据网格, 资源调度, 分布式系统, 商空间, 粒子群算法

Abstract:

In order to solve the problems of falling into local optimum value and converging too slowly when allocating resources in hierarchical data grid using traditional algorithms,the granular computing was introduced and a quotient space theory based hierarchical data grid resource allocation (QSHDGRA) algorithm was proposed.Firstly,the charac-teristics of hierarchical data grid were analyzed.Secondly,a reconciling model of minimum average waiting time and maximum network and node resource utilization was defined,and then the QSHDGRA algorithm was designed.The al-gorithm can allocate resources from coarse granularities to fine ones,so it can guarantee the QoS of different businesses and make global optimal resource allocation.Simulation results show that QSHDGRA can improve overall system throughput with faster convergence speed and linear scalability.

Key words: data grid, resource allocation, distributed system, quotient space theory, particle swarm algorithm

No Suggested Reading articles found!