A Priority based Dynamic Load Balancing Approach in a Grid based Distributed Computing Network

Load balancing in grid based distributed computing environment increases the availability and scalability of entire system. Dynamic load balancing has the potential to perform better than static load balancing, but they are inevitably more complex. The overhead involved is much more but one can not negate their benefits. Load balancing strategies try to ensure that every processor in the system performs almost the same amount of work at any point of time. Process migration is one of the important tasks in dynamic load balancing which usually deals with the migration of task from overloaded computing nodes to under-loaded nodes. Although numerous works has been done on the issue of process migration and load balancing. Communication overhead is still a problem which is to be reduced in grid based networks. In this paper, we propose an algorithm that finds an under-loaded node whenever an overloaded node is found, and also takes the idea of assigning a priority to each computing node in the grid system based on their computing power. The proposed algorithm reduces the communication overhead and proves to be cost effective.