Efficient load balancing scheduling for deadline constrained tasks on grid computing

Grid Computing provides scalable access to wide-area distributed resources. Since, computational grid selects, shares and aggregates wide variety of geographically distributed computing resources and proposes them as a single resource for solving large scale computing applications but there is a need for a scheduling algorithm which takes into account the several requirements of grid environment. Hence, this research proposes a new scheduling algorithm for computational grids that considers user satisfaction, load balancing; fault tolerance based on the resource availability and job characteristics such as user deadline. This algorithm decreases the make-span of the schedule along with user satisfaction means more tasks are successfully completed within user deadline and balanced load means load on all the resources is balanced. The proposed algorithm improves the system parameter such as maximum variation in load; deadline hit count and make-span as compared to BSA.