July 16, 2007

Sun brightens Purdue's future in managing 'data deluge'

Sun has provided the Purdue Libraries an object-based storage system to help develop its data library to handle the volumes of information being generated by university researchers.

The system is called the Sun StorageTek ™ 5800 System, and Purdue has implemented a storage capacity of 32 terabytes. By comparison, it would take about 6,800 conventional DVDs to store as much information.

Purdue is one of only a handful of universities worldwide to begin work on developing such repositories.  And while Purdue and these other institutions lay the groundwork, it could become a wave of the future as researchers grapple with what many scientists refer to as "the data deluge."

"We feel we're on the leading edge of something that is increasingly crucial - curating science data to make it easily discovered and accessible," said D. Scott Brandt, associate dean for research in the Purdue Libraries and interim head of Purdue's Distributed Data Curation Center, also known as D2C2. "The results of our research will allow scientists to concentrate on their projects while we help them determine how to best archive and preserve their data."

Purdue's digital data collections are not housed in a single building or a on a single server, instead, Brandt refers to the data collections as a part of a "distributed institutional repository." The data is stored in many different places, including researchers' hard drives, departmental servers, and on the TeraGrid, a large-scale national computing grid.

Storing and retrieving data is an increasing problem for scientists as technologies speed the processes of research and create loads of information.  The key is that Purdue librarians will consult with scientists to examine data and produce metadata, which is information that tells how to find and interpret the data.

Michael Witt, the interdisciplinary research librarian and an assistant professor of library science, sought corporate collaborators such as Sun to support his research in data repositories at the D2C2.  He says he has met with several scientists, and they are excited about the prospects.

"By sharing data, other scientists can validate your work," Witt said. "So it's just good science. Plus, some grant funding agencies are beginning to require researchers to archive and cite their data. There’s also no point in re-inventing the wheel.  Once data has been created, it can be reused or repurposed by someone else in the same or another discipline."

An official at Sun said that working with Purdue Libraries will benefit both the university and the company.

"We at Sun are immensely excited about the opportunity to collaborate very early in the rollout of one of our most strategic products, the 5800 system, also affectionately known as 'honeycomb,' with one of the world's leading academic institutions," said Art Pasquinelli, Sun's  director, market development for education. "Purdue's adoption of this new storage technology will be beneficial not only to both our organizations, but also the global archiving and research communities.

"Honeycomb includes embedded metadata attributes and query capabilities to speed up finding the data stored."

Anyone wanting more information can do an Internet search for honeycomb and Sun, Pasquinelli said.

Purdue Libraries and the D2C2 will use the 5800 system to prototype different methods of describing, preserving, and providing access to scientific datasets. The results of their research will be incorporated into the university's new data repository service, Purdue e-Data, that researchers and the general public will be able to see and search when it goes online in the fall.

Writer: Jim Bush (765) 494-2077, jsbush@purdue.edu

Sources: D. Scott Brandt, (765)494-2889, techman@purdue.edu

Michael Witt, (765) 494-8703, mwitt@purdue.edu

Art Pasquinelli, (303) 272-9479, art.pasquinelli@sun.com

Purdue News Service: (765) 494-2096; purduenews@purdue.edu

To the News Service home page