Data Cube : A relational Aggregation

Many scientific computing tasks are too large to be processed quickly on one computer. Instead, the grid has become popular for a wide range of disciplines. Efficiently using the all the resources of the grid is a challenging, NP problem though. The authors introduced a new method for scheduling that accounted for data dependencies, a topic often ignored in job scheduling algorithms. They also analyzed the method extensively, in terms of tasks that were compute bound, data bound and tasks where job completion was estimated incorrectly.