Managing Large Distributed Data Sets Using the Storage Resource Broker

摘要

The Storage Resource Broker (SRB) data grid enables the formation of shared collections that span administrative domains and multiple storage resources. A typical collection may have over a million files of both simulation and observational data. Uniform access mechanisms, uniform global names and uniform access controls are applied to the data, no matter where they are located. This makes it possible to automate data analyses across the distributed data, removing the need to aggregate data into a single location. The SRB is generic infrastructure that supports the requirements of multiple data management applications, ranging from data grids for shared collections, to digital libraries for data publication, to persistent archives for data preservation, to real-time data collection federation.