Abstract :
The Biomedical Informatics Research Network (BIRN) is creating, extending, and deploying a production grid to enable groups of biomedical researchers to share complex and large-scale data. In this talk we describe the BIRN, its structure and overall Information Technology Structure. In addition we will describe the details of how the complete software stack is integrated, managed and deployed. The Rocks toolkit, originally designed for clusters has been greatly influenced by the deployment needs of BIRN, GEON, the OptIPuter and other large science-based grids. BIRN is funded by the U.S. National Institutes of Health and integrates science teams across more than 20 institutions to share and federate 3- dimensional medical image data. We will describe, within the context of BIRN, the challenges of everincreasing data set sizes, software stack complexity, integration with other production grids, and overall system reproducibility.
