Purdue Libraries and Information Technology at Purdue (ITaP) are collaborating on an initiative that includes innovative software developed on campus to help researchers store, sort, archive, retrieve and manage large-scale data and information.

The Distributed Institutional Repository is a Web-based data portal that provides tools and systems to manipulate large data sets and to help users understand the origins of data and learn about additional research applications using the same data.

“The DIR is an architecture Purdue Libraries developed which utilizes a unique approach to pulling access together from a number of distributed repositories,” said Scott Brandt, professor of library science and associate dean for research for Purdue Libraries.

Most institutional repositories use a single system within a single software environment. In addition to data repositories built by ITaP and the Libraries, the DIR will access a wide variety of information systems, including electronic dissertations e-prints and archival special collections.

Purdue Libraries' role in the repository is to define organizational structure, access to and the retrieval of data deposited and archived from three broad categories: traditional sources, such as journals and books; current digital sources, such as Web pages, digital video and electronic documents; and future sources, such as remote sensing data and large-scale visualization and perceptualization.

“The Libraries have always supported research through building and organizing collections,” Brandt said. “With a distributed repository, we hope to enhance discovery and use of data across campus.”

An institutional repository is necessary to meet the growing need to house massive amounts of data and to make it usable for users across the academic environment, Brandt said.

“Projects like this are another example of how we are bridging discovery and learning in non-traditional ways,” said Krishna Madhaven, research scientist and project leader with ITaP's Rosen Center for Advanced Computing. “We're starting to see the big picture of how large data computation, learning, research and network are coming together.”

The ITaP data repository is a multi-disciplinary data management system that was built on top of Storage Resource Broker (SRB) and deployed on the TeraGrid network, said Lan Zhao, senior analyst and programmer in the Rosen Center. She also developed the SRB-based data management architecture that is connected to the Distributed Institutional Repository.

SRB is an open-source data-management middleware developed by the San Diego Supercomputer Center.

Purdue added its own twist through software developed by Michael Witt, senior research systems administrator for Purdue Libraries, that can interface between SRB and standard OAI-PMH software (open archives initiative-protocol for metadata harvesting), which allows repositories to talk to each other.

This new interface, called OAISRB, allows connectivity between the traditional provisions of both SRB and OAI-PMH.

“It's an innovation,” Brandt said. “It's one of the unique things that Purdue has done.”

Through Purdue's repository, three data sets are currently available and being measured for accessibility and reliability. Each one uses a different storage/retrieval vehicle: climate data, stored on the TeraGrid shared by nine universities; remote sensing data available via the Laboratory for Applications of Remote Sensing; and water-quality data, available through the Web portal at Purdue.

“We need to be able to access this data wherever it resides,” Brandt said. “This gives us different ways to test that access and discovery.”

Part of the overall initiative is to work with Purdue's Cyber Center to address issues related to interoperability of data — representing and combining data in new ways to generate knowledge.

In addition to providing a home for researchers' massive data sets, the repository project also will spur decisions about data-rights management, sharing of intellectual property and length and format of data archival storage.

“Library science is stepping in to fulfill this role for modern research,” Brandt said.

A gift of a 32-terabyte Centera storage system from Indianapolis-based EMC Corp. in September provided the impetus for ITaP and the Libraries to launch the repository project. The project will lead to future collaboration with the Purdue Cyber Center in Discovery Park as well.

For more information about the May 2 presentation, visit www.rcac.purdue.edu/rcac/events/library.htm.

The Rosen Center for Advanced Computing is a research-computing center located at Purdue University's West Lafayette campus. The Rosen Center is a component of Information Technology at Purdue, the central IT organization at the university. The Rosen Center, through its role in the TeraGrid project, provides advanced computing resources and services to support computationally intensive research nationwide. The center also conducts its own research and development to enhance and extend the capabilities of these resources.

—–

Source: Purdue News Service