At the Open Repositories 2010 conference, DuraSpace will be recommending that the DSpace community begin to work more closely with the Fedora community to allow the DSpace software to run "on top of" the Fedora platform. Essentially, this would mean that the DSpace application would have an option to install it with "Fedora Inside". What follows is an initial FAQ around this proposal.
Please note that this FAQ is still an early draft – feedback and comments are welcome! Contact Tim Donohue or Valorie Hollister at DuraSpace if you have additions / changes you'd like to see in this FAQ.
DSpace & Fedora Integration Questions
What are the plans for integrating the DSpace application with the Fedora repository platform?
For some time there has been interest expressed among DuraSpace communities in having Fedora act as either a robust storage layer or archive layer behind the DSpace application while maintaining the out-of-the-box DSpace experience including workflows, and look and feel. There is work underway than may eventually enable the option to run the DSpace application on top of a Fedora repository. To achieve this requires a series of developments:
- A way to extract the items from DSpace's current internal store, along with all of their relationships (communities, collections, and Epeople).
- Related Project(s): DuraSpace is currently developing this as a general, backup/restore AIP capability for DSpace.
- An implementation of the DSpace data model using the Fedora repository.
- Related Project(s): There is prior art (which would need to be updated) for this from GSoC 2008 and GSoC 2009. There is also a current GSoC 2010 project to backport DSpace 2.x Storage Services to 1.x – this work may enable the ability to use prior Fedora integrations in the DSpace 1.x platform.
- A conversion, crosswalking, or import utility for the above two items.
- Related Project(s): At this time, there is neither a schedule nor a commitment for this work.
- Designing an integration of the DSpace business logic, User authentication and authorization with Fedora.
- Related Project(s): At this time, there is neither a schedule nor a commitment for this work.
- Porting the DSpace user, administrative, and web services interface applications to run against the resulting environment.
- Related Project(s): At this time, there is neither a schedule nor a commitment for this work.
Not all of these steps need to be complete for some models of operation involving running DSpace and Fedora as distinct, cooperating applications.
What is the proposed end result of this integration?
The end result would be a family of packages that you would install together for the full DSpace experience. One of these packages would offer a Fedora-based storage layer. So, when you install DSpace, there would be an option to use Fedora as its storage layer.
Essentially, DSpace will still retain very similar interfaces and workflows, but it will now have "Fedora Inside".
Will the end result still look like DSpace and be called DSpace?
Yes. We are committed to maintaining the DSpace out-of-the-box experience. The end result of this integration may be an evolved version of DSpace (with new or slightly changed features), but it will still provide the same out-of-the-box institutional repository software.
To put it another way, the goals of DSpace software will not change. DSpace will continue to provide the same type of experiences you've come to expect. However, we hope the end result will provide even more useful features and will help to evolve the DSpace platform.
What are the perceived benefits? Why do this integration?
We feel this tighter integration between the two platforms provides the following benefits:
- Less redundancy between DSpace & Fedora code and development.
- Ability to combine development work/projects – a larger pool of developers working on common technologies.
- DSpace will also be able to take advantage of features offered by Fedora (which are not available in DSpace). Some examples may include:
- File versioning feature of Fedora, which DSpace could utilize to support versioning of DSpace Objects.
- More flexible architecture, which may allow DSpace to eventually support hierarchies other than "Community -> Collection -> Item"
- Flexible relationships model, which may allow for better defined relationships between DSpace Objects (e.g. linking together all articles within a Journal Issue, or all Issues within a Journal)
How is this work being supported?
DuraSpace cannot support this work on its own. That's why we are making a strategic recommendation that the DSpace and Fedora communities work towards tighter integration between the software platforms.
DuraSpace is committed to helping with strategic development work that helps to further this initiative. For example, we feel our ongoing AIP Backup/Restore work will be beneficial in upgrading existing DSpace instances to an eventual DSpace with "Fedora Inside". However, we do not have the technical staff to do this work alone.
Who will do this work? When will it be complete?
We currently don't have an exact timeframe for this work to be complete.
As stated, DuraSpace does not have the technical staff to do this work alone, nor would we want to. As with any open source development initiative, the commitment and support of the broader community will build a better 'end product', than if one institution were to implement it on their own. We are currently looking for DSpace and Fedora community volunteers to help make this initiative happen.
How can I help? How do I volunteer?
Once this work is complete, how hard will the upgrade be for current DSpace users? Will I have to do all the work myself / how much expertise will I need to perform this upgrade?
At this time, we don't know the complete answer. Our intention is the make this upgrade as easy as possible and come up with common solutions that everyone in the community to utilize. One of these tools we are currently working on is a way to export all your DSpace contents (Communities/Collections/Items) into a series of AIPs (Archival Information Packages). It's possible that this tool could be used to ease the migration of your DSpace contents from the current DSpace data model into a Fedora data model.
Will I also need to learn how to install/upgrade Fedora in order to install/upgrade DSpace?
In the early stages, you may need to perform some basic steps to install or upgrade Fedora (how many steps?) whenever you install or upgrade DSpace. We anticipate that in the future we will be able to simplify this process. The final goal is to be able to install DSpace with "Fedora Inside" without too much effort.
Will my current DSpace User Interface / Add-On customizations still work?
We currently don't know. We hope that basic user interface customizations (CSS / Javascript / etc) will work immediately. However, any customizations to the internal API or complex Add-ons will likely need to be rebuilt. As we move closer towards figuring out the internal API changes required for this integration, we will be able to better answer this question.
What happened to the DSpace 2.0 work? Is this replacing it?
This initiative is actually an extension of the DSpace 2.0 investigation work that took place in 2009. The goal of the DSpace 2 work was primarily to modularize DSpace, and Fedora as one of those modules just makes sense. The DSpace 2 work, along with recent Google Summer of Code projects (in 2008 and 2009), have helped us to more closely realize how DSpace and Fedora could be integrated.
