The Study Of Grid And Ant Algorithm-Based Task Scheduler Simulation Design

Along with the development of modern information requirement, it is very difficult to solve more and more complex question. At the same time, we need faster hardware and more sophisticated software to deal with the fact that a great deal large scale computing is emerging. So we put out the Grid. Grid is a new concept, a new infrastructure of technology and a new technology advance after Internet.Task scheduling and load balancing are very important parts in Grid. Because Grid resources lie in different geographical positions, managed by different organizations, and some resource may fail sometimes, task scheduling and load balancing are very complex.There are two parts in the contents of this thesis. The first part mainly understands the knowledge of Grid infrastructure and introduces the study of Grid system in existence, applied globus based on linux7.2 system. The second part mainly introduces the study of task scheduling on Grid. Base on ant algorithm, we design task scheduler simulation with Java which give a good method for task scheduling on Grid.The first part of this thesis describes the theory of Grid infrastructure. Applied with globus, we know the grid architecture. Subsequently, we can create and implement ogsa service on globus3.0 and know the advantage and disadvantage on Grid system in existence.The second part introduces the task scheduling on Grid resource. We design a method of ant algorithm-based distributed resource management and scheduling for Grid computing. In this part, we apply the classification of tasks and add a load balance factor. We submit tasks to Grid resource relatively idle and make the task dispatcher and load balancing effectively.