Network Job Management System--Implementation On Windows 2000 Platform Study On High Performance Cluster

Within the past few years, a number of technological trends and events have begun a convergence that has a very strong potential of change the face of scientific computing. Intel-based computers, which have long dominated the PC market, have grown strong enough to compete with the more traditional framework or workstation. On the other hand, gigabytes net substantially promote the step of, to say, parallel computing. And cluster techniques connect geographically distributed PCs as a overall somewhat "supercomputer" to meet the ever-growing need of faster computing capability with lower price. Then modern job managing system face up the point and grow further in the fields of high performance computation.In this thesis, I'm trying to present the system structure and principle of a work I took part in from 2000 on job managing system, which is the first commercial product of this field. During this course, mathematic methodology and graphic tables got from the real emulating environment are employed to put my point of view through. I try to upgrade some of the function in this system by adding new ideas into the opening subsystem NQS, including load distributing and resource management among the machine group.To meet these ends, there are seven chapters included in: Chapter 1:Introduction to job managing system and hot topic Chapter 2Implementing of a job managing system. Chapter3Implementing of jobcenter cluster techniques in windows 2000 advanced server platform Chapter 4Load balance on job managing system Chapter5Resource management in job managing system Chapter6Evaluation of the job managing system Chapter7An example.At the end of this thesis, I foresee the future of job managing system. Future will see a splendid scene in the further step of job managing system's application with the upgrading of hardware in all walks of our country.