Download Now Free registration required
Distributed and dynamic nature of grids causes the probability of failure is great in such systems. So fault tolerance has become a crucial area in computational grid. In this paper, the authors propose a new genetic algorithm that used RFOH for having reliable job scheduling in computational grid. This strategy maintains the fault occurrence history of resources in Grid Information Server (GIS). Genetic algorithm with RFOH finds a near optimal solution for the problem. Furthermore, it increases the percentage of jobs executed within specified deadline. The simulation results demonstrate that proposed strategy decreases the probability of failure and therefore increases reliability. Also, it reduces total execution time of jobs. So they will have a combination of reliability and user satisfaction.
- Format: PDF
- Size: 280.06 KB