No abstract

No abstract

1.IntroductionComputational Grids are a new trend in distributed computing systems. They allow the sharing of geographically distributed resources in an efficient way, extending the boundaries of what we perceive as distributed computing. Various sciences can benefit from the use of grids to solve CPU-intensive problems, creating potential benefits to the entire society. With further development of grid technology, it is very likely that corporations, universities and public institutions will exploit grids to enhance their computing infrastructure. In recent years there has been a large increase in grid technologies research, which has produced some reference grid implementations.Task scheduling is an integrated part of parallel and distributed computing. Intensive research has been done in this area and many results have been widly accepted. With the emergence of the computational grid, new scheduling algorithms are in demand for addressing new concerns arising in the grid environment. In this environment the scheduling problem is to schedule a stream of applications from different users to a set of computing resources to minimize the completion time. The scheduling involves matching of applications need with resource availability. There are three main phases of scheduling on a grid [5]. Phase one is resource discovery, which generates a list of potential resources. Phase two involves gathering information about those resources and choosing the best set to match the application requirements. In the phase three the task is executed, which includes file staging and cleanup. In the second phase the choice of the best pairs of tasks and resources is NP-complete problem [5]. A related scheduling algorithm for the traditional scheduling problem is Dynamic Level Scheduling (DLS) algorithm [17]. DLS aims at selecting the best subtask-machine pair for the next scheduling. To select the best subtask-machine pair, it provides a model to calculate the dynamic level of the taskmachine pair. The overall goal is to minimize the computational time of the application. In the grid environment the scheduling algorithm no longer focuses on the subtasks of an application within a computational host or a virtual organization (clusters, network of workstations, etc.). The goal is to schedule all the incoming applications to the available computational power. In [2,10] some simple heuristics for dynamic matching and scheduling of a class of independent tasks onto a heterogeneous computing system have been presented. There are two different goals for task scheduling: high performance computing and high throughput computing. The former aim is minimizing the execution time of each application and latter aim is scheduling a set of independent tasks to increase the processing capacity of the systems over a long period of time. Our approach is to develop a high throughput computing scheduling algorithm.The organization of the paper is as follows. In section 2 the simulated annealing method is discussed and it basic structure. In se...

No abstract