Characteristic Approximation for Resources on Computational Grids

This paper proposes a statistical method to approximate the distributions of computational and communicational abilities of resources on computational grids. The well-known gamma and normal distributions are employed to approximate the distribution of computational ability and communicational ability for those computational resources. The result of the preliminary experiment shows that this study can lead to a better approximation of computational and communicational abilities in the real world. Our method is useful for the grid simulation tool because it provides the computational and communicational abilities with realistic characteristics for a grid resource model.