网格计算中作业管理系统的分析及设计

The Analysis and Design of Job Manage System in Grid Computing

随着高性能应用需求的迅猛发展,单台高性能计算机已经不能胜任一些超大规模应用问题的解决。这就需要将地理上分布、系统异构的多种计算资源通过高速网络连接起来,共同解决大型应用问题,作业管理是随着集群技术和网格技术的应用发展而发展起来的一项技术。针对网格计算中的作业管理系统,该文主要分析了作业管理系统的设计、实现及关键技术,以及作业管理系统中的各种组件的工作原理,并在此基础上提出并设计出针对网格计算的作业调度器的框架。

With the rapid developing of applications in high performance computing, single computer with high performance can't competent for some solution of grand scale application. Then, it needs to connect some kinds of computing resources which is distributed in geography and in system through rapid network , solve large scale application together. Job management is a kind of developing technology with the development of technology in cluster and grid .Aiming at job manage system in grid computing ,this paper mainly analysis the design?realization and key technology, as well as the working principle of all kinds of components in job manage system, and on this base bring forward and design the outline of framework of job scheduler in grid computing.