Reliability-Aware Scheduling Based on a Novel Simulated Annealing in Grid
Wael Abdulal, Sirandas Ramachandram · 2012
Grid technology has emerged as an important implement for solving the computation intensive applications. Grid systems must exhibit high Reliability to continuously provide correct service. Moreover, Grid scheduler, as a core part of Grid systems, must use the available resources efficiently to make mutually conflicting goals satisfied. Therefore, multiple factors such as execution time, flow time, Reliability and time to release task to the user, need to be considered while making a scheduling decision on computational Grids. Achieving these goals will be made more difficult as Grid systems grow in scale and become more heterogeneous and dynamic in nature. This paper provides another way to find a good solution to the scheduling the independent tasks in the Grid environment by introducing a novel Simulated annealing. The novel simulated annealing uses a novel (Random-Minimum Completion Time) heuristic which provides a high-quality solutions at the initialization step. This heuristic can help simulated annealing to find better solutions more quickly than they can from a random start. The experimental results show that the novel simulated annealing is proved to have a better performance than the existing algorithms in the literature specially roulette wheel selection genetic algorithm. Other benefits of the novel simulated annealing include robustness and scalability.