State of Data Management in the DOE Research amd Development Complex. Report of the Meeting 'DOE Data Centers: Preparing for the Future' held July 14-15, 2004 in Oak Ridge, Tennessee

摘要

DOE invests billions of dollars in R&D projects that generate and use data as a critical resource integral to science. Examples abound. Through its large-scale experimental facilities, DOE provides its researchers with an ever-growing array of world-class instruments that produce vast amounts of data in probing the fundamental workings of science. Science communities have developed many new ways to exploit data collections. Today, state-of-the-art computer simulations and modeling have become as important to the advancement of science as theory and experimentation always have been. Recently, Ray Orbach, Director of the Office of Science, highlighted the growing emphasis on computational science as a fundamental shift in research. Simulation-based science is dependent on the availability, accuracy, and usability of the data previously generated by scientists and funded by DOE and its predecessor agencies. Increasingly, valuable source data are being stored in digital databases. The charter for the National Science Boards (NSB) Task Force on Long-Lived Data Collections (LLDC) states, the number of digital databases is increasing rapidly, and database sizes are in many cases growing exponentially. Digital databases are now an essential and increasingly large component of the cyber-infrastructure that underpins research and education. Early in 2004 the NSB created the LLDC Task Force to study data management practices across the federal science agencies and determine how they should shape future NSF funding policies. Some R&D agencies already have data management policies, while other R&D agencies are now exploring the needs relating to data preservation and use.