February 27, 2002
Conference Paper

Data Grid Implementations

Abstract

Data grids link distributed, heterogeneous storage resources into a coherent data management system. From a user perspective, the data grid provides a uniform name space across the underlying storage systems, while supporting retrieval and storage of files. In the high energy physics community, at least six data grids have been implemented for the storage and distribution of experimental data. Data grids are also being used to support projects as diverse as digital libraries (National Library of Medicine Visible Embryo project), federation of multiple astronomy sky surveys (NSF National Virtual Observatory project), and integration of distributed data sets (Long Term Ecological Reserve). Data grids also form the core interoperability mechanisms for creating persistent archives, in which data collections are migrated to new technologies over time. The ability to provide a uniform name space across multiple administration domains is becoming a critical component of national-scale, collaborative projects.

Revised: February 15, 2005 | Published: February 27, 2002

Citation

Moore R.W., R.S. Studham, A. Rajasekar, C. Watson, H. Stockinger, and P. Kunszt. 2002. Data Grid Implementations. In Proceedings of the 4th Global Grid Forum Conference, edited by Charlie Catlett and Ian Foster, 26 pp. Glen Ellyn, Illinois:G.L. Anderson & Associate, Inc. PNNL-SA-36034.