A New Randomized Algorithm for Handling Scheduling Conflicts in Grids

As the scale of Grid platforms grows, the idea of a centralized scheduler loses its efficiency, and it is replaced with the scheme of decentralized schedulers. However, a new problem emerges in distributed scheduling systems, which is how to coordinate the autonomous schedulers to avoid the occurrence of conflicting schedules. In this paper, by exploiting the idea of randomized algorithms, a new scheduling scheme has been proposed, which addresses the problem of scheduling conflicts. The proposed algorithm is thoroughly decentralized in the sense that there is no central point of contact in the system. In addition, our approach is a suitable way toward reaching scalability and autonomy in future Grids. We prove the feasibility and effectiveness of the proposed algorithm through statistical analysis.