The Research Of Load Balancing Policy Based On Grid Computing

With the advancement and development of various technologies, the computing problems we need to solve have become more complicated and larger in size. However, due to the high cost of super computers, they can not be extensively utilized. Thus, people have to resort to distributed systems to resolve the problem of large-scale computing. Grid Computing utilizes the network and combines idle resources scattered in every region for distributed applications. As Grid Computing uses Internet connections, compared with conventional distributed systems, it provides better large-scale sharing resource, improves resource applications and solves these science problems. Load balancing in the grid computing environments is the key technology by way of combining the broad-field internet resource.The thesis has solved load balancing services in the grid computing environment. There are four primary technical innovations in this thesis:(1) Bring forward a new cascading load balancing algorithm combined with many kinds of load balancing algorithms in the grid computing environment, and choosing agilely load balancing algorithms based on the load balancing of first-grade brokers and second-grade brokers in order to improve computing efficiency and system stability.(2) Adopting CORBA-based Architecture in order to make load balancing system transparentized with grid middleware services.(3) Using the primary of dynamic adaptive load balancing algorithm assisted with static load balancing, According to the complexion of computing task, improving retractility and responding time of system by redistributing subtasks.(4) Decentralizing services of grid resource to avoid to be subjected to the unitary unavailable point.The practicability and validity of load balancing policy has been approved with laboratorial simulation in grid computing environment by this thesis.