Skip to end of metadata
Go to start of metadata

On September 1, 2010, Serge Goldstein of Princeton gave a webinar to Educause Live titled "DataSpace: A Funding and Operational Model for Long-Term Preservation and Sharing of Research Data".  Chris and Patrick M attended. DataSpace is a relatively young service offering at Princeton ( The main part of the presentation covered the sustainability model for DataSpace, which is based on a POSE model (pay once, store endlessly).  The speaker outlined an algorithm for calculating the cost to store endlessly which comes out to between two and three times the cost for initial storage.  However he cautions that this only covers archiving and a permanent URL.  Things like data curation and major reformatting might be additional projects that someone would have to pay for.  Your institution's calculation of a multiplier will vary depending on expected data volatility and infrastructure decisions (e.g., tape backup can add to your multiplier).  There was a large number of comments from the audience about the assumptions built into the model. 

The speaker admitted that convincing researchers to use the service is the more difficult part but argues that having a one-time cost model helps (for funding agencies) and that you have to be patient while you build the case with the research community. 

While Princeton is using DSpace for their research repository, it sounded like the DataSpace was really about the model for sustainability and funding.  Dr. Goldstein concluded by calling others to work with Princeton to build out a network of DataSpace-like research repositories.

It was really good to see a presentation on research computing and repositories on Educause.  There were approximately 330 attendees (including some attendees who obviously represented multiple individuals in a conference room somewhere).  The slides and financial model are attached to a separate wiki page though with access restricted to those in the Data Services team.  Please do not distribute beyond this team.

  • No labels