ABSTRACT
With the fast development of the Internet, computer applications need to solve the
more complex problem, more and more large-scale. Shared, heterogeneous and dynamic
resources widely distributed on the Internet, with the attendant will have a new Internet
technology which utilizes the Internet to make geographically dispersed resources full
share—Grid Technology. Grid Technology will make geographically dispersed
resources, heterogeneous resources and dynamic resources interconnected or
re-integration by Internet or some private network, which aims at realizing resources
comprehensive sharing and collaboration.
Task management, task scheduling and resource management are three basic
functions in grid system. Task scheduling is a reasonable solution to find the optimal
and dynamic match between tasks and a variety of resources, which is one focus of
studying Grid and a core technology in Grid System. In the grid system, resources are
dynamic, heterogeneous and autonomous, different scheduling algorithms may have
different goals, some tasks are to make some priorities, some to make the task
completion time at least, some to make the task with minimal economic cost to deal
with, due to the characteristics of grid resources and the complexity of the task, so the
task scheduling in the grid environment is extremely complex. Task scheduling is facing
a NP-complete problem. In the grid system, task scheduling has become a bottleneck
restricting the grid.
In this paper, firstly introduce overview of the concepts of grid computing and the
development status, describes the importance of task scheduling. Finally, projects the
task scheduling algorithm which considers the computing power of an integrated,
dynamic scheduling and load balancing.
Key Words :Grid Technology, Task Scheduling, Load Balance,
Dynamic Optimal Scheduling Match