Dynamic ACO-based fault tolerance in grid computing

Scheduling jobs in distributed conditions of grid computing is nearly impossible to have a completely fault-free system. It is important to integrate fault tolerance capability in the system so that the system can continue to run even in the presence of failure in addition to improving the scheduling process as well as reducing the possibility of faults. Typically, load balancing is not considered in the presence of failure and this may lead to an inefficient scheduling process despite having a good fault tolerance strategy. This paper presents an ant-based fault tolerance algorithm that used checkpoint and resubmission techniques with consideration of execution history in the pheromone updating process to enhance fault tolerance capability. Experimental results showed that the proposed algorithm has better performance as compared to other relevant algorithms in terms of execution time, success rate, and average turnaround time per job.