Cost-based scheduling for data-intensive applications on global grids

We present an algorithm for scheduling distributed data intensive bag-of-task applications on data grids that have costs associated with requesting, transferring and processing datasets. We evaluate the algorithm on a data grid testbed and present the results.