Harvard Builds its Dataverse Network on GlassFish
By woodjr on Mar 06, 2007
Our project, the Dataverse Network, is an on-line archive for sharing data within and across universities and other institutions conducting research. Stability and scalability have been a concern in developing this software from the beginning. The application serves not only as an online archive for storing research data, but also provides social science researchers with the means to cite their own data and allow others to replicate the results through an extensive on-line analysis tool...
What application server did they choose to power the Dataverse Network? GlassFish, of course. Early in 2006, the team decided upon using Java EE 5 and "inquiries at the JavaOne 2006 conference made it clear that GlassFish would be the only fully compliant Java EE 5 implementation in our initial development timeframe."
The benefits of using GlassFish didn't end with its first-to-market Java EE 5 support, though. Remember their stringent stability and scalability needs? Not a problem. As Merce notes, "GlassFish's stability enabled us to concentrate on the code and not worry about the server environment."
Integrating other technologies with GlassFish also went well. The team uses NetBeans, Java Studio Creator, the Lucene search engine, PostgreSQL database, Shale standalone tiles, and AWStats web analytics tools. Integrating and using so much new technology required that the team "forge new ground" in a few cases, but overall they found the GlassFish server environment to be "very configurable and easy to use." For example, integrating the Awstats web analytics tool "was as simple as locating the HttpServices window and modifying the access log format all in an easy-to-use, uncomplicated interface."
The Dataverse Network will be live in the next few weeks, "serving all social science data to all Harvard and MIT faculty, students and staff." Want more info in the meantime? Here are some additional resources:
- The full response to the GlassFish adoption questionnaire from the Dataverse Network project's Senior Manager of Software Development, Merce Crosas.
- An Introduction to the Dataverse Network as an Infrastructure for Data Sharing (academic paper) by Gary King, principal investigator of the Dataverse Network Project.