Workbook on Digital Private Papers > Digital repositories > Comparing repository software for preserving personal digital archives
Comparing repository software for preserving personal digital archives
Managerial concerns
System administration concerns
System administrators are responsible for administering the infrastructure. They will want to know the answers to a number of questions, such as those listed below:
- What environment - the hardware and software - is needed by the repository?
- How is the environment and repository installed, configured and patched?
- How is the repository to be backed up?
- How is the repository secured? What authentication and authorisation services are available?
- How will the repository fit with existing infrastructure?
- Are upgrades designed to be minimally disruptive?
| Name |
Provides installation instructions including hardware and software required |
| Detail |
System administrators must prepare the environment for installation. |
| DSpace |
Supported |
| Detail |
The DSpace wiki has links to the installation documentation in various formats. See the latest DSpace documentation at Sourceforge. |
| Fedora |
Supported |
| Detail |
Fedora provides clear links to technical documentation for the current release, which includes:
- Release notes.
- Installation guide.
- Update and migration.
|
| Name |
Provides patching information |
| Detail |
System administrators must know when patches are available, and whether they should be installed immediately or whether they may adversely affect the repository. |
| DSpace |
Partial support |
| Detail |
Patches are communicated via the mailing list. There is the implicit expectation that systems administrators are familiar with the requirements of the repository, monitor the bug-tracking systems and are able to assess the implications of upgrading or patching the system. |
| Fedora |
Partial support |
| Detail |
Patches are communicated via the mailing list. System administrators are expected to monitor the bug-tracking system and mailing-list and make decisions about the necessity of patches offered there. As Fedora evolves into an open source community this is unlikely to change. |
| Name |
Provides information on backing-up |
| Detail |
System administrators need documentation of any special requirements to be aware of in planning backup and disaster recovery strategies for the repository. |
| DSpace |
Supported |
| Detail |
Information on how to organise backup for a DSpace repository is available. |
| Fedora |
Supported |
| Detail |
Fedora 2.2 provides a journaling module that allows a repository to be mirrored, or to restore a Fedora repository to the exact state before failure, rather than the state at last backup, in case of corruption or failure of the repository. The Fedora repository can also be completely rebuilt by crawling the digital object XML source files that are stored on disk. |
| Name |
Provides information on securing repository |
| Detail |
What authentication and authorisation services are available? Are they compatible with Shibboleth? How are these documented? |
| DSpace |
Supported |
| Detail |
DSpace associates application sessions with specific users and/or groups via a mechanism called Stackable Authentication (permitting custom authentication methods to be stacked on top of the default DSpace username/password method). DSpace also maintains authorisation policies that allow it to understand which credentials are required (if any) to undertake actions on particular resources. Shibboleth can be used with DSpace. |
| Fedora |
Supported |
| Detail |
Fedora 2.2 supports an Access Control and Authentication module that includes the ability to enforce fine-grained access control policies expressed using XACML at the level of the Fedora web service APIs and down to the object/datastream/dissemination level. Release 2.2 also introduced configurable authentication (Tomcat realms and login modules) with out-of-box support for multiplexing multiple authentication sources, including Tomcat-users and LDAP. Shibboleth can be used alongside Fedora. |
| Name |
Are upgrades designed to be minimally disruptive? |
| Detail |
How is updated software supplied to the community? Are update and migration instructions provided? |
| DSpace |
Supported |
| Detail |
Upgrade and migration instructions are provided. The nature of the DSpace architecture has led some users to make some local changes to the software that may require some work to migrate to newer versions. |
| Fedora |
Supported |
| Detail |
Upgrade and migration instructions are provided and Fedora's service architecture should minimise upgrade problems. |
Developer concerns
Developers are responsible for creating new system functionality. They will want to know:
- Is there good quality developer documentation which explains how to obtain source code, the development conventions for the creation and submission of source code, how to report bugs, etc.?
- What is the quality of the source code and APIs?
- Is there a healthy developer community?
| Name |
Provides developer guidance |
| Detail |
Guidance to get developers started should be available. |
| DSpace |
Supported |
| Detail |
Information for developers is available. |
| Fedora |
Supported |
| Detail |
The documentation is distributed at present. Now the software is to become community driven a developer's guide has been promised. |
| Name |
Good quality developer documentation |
| Detail |
Published APIs, data models, etc., are essential. Explicitly documenting these aids third-party development of decoupled applications. |
| DSpace |
Supported |
| Detail |
System documentation is packaged with the source code and Javadocs can be generated from the sourcecode after download. There is also developer documentation for coding with the DSpace system. |
| Fedora |
Supported |
| Detail |
System documentation and Java docs for v. 2.2 are available. |
| Name |
Active development community |
| Detail |
Both DSpace and Fedora are open source communities with small teams of developers distributed across organisations. Provision of a developer support network is important. |
| DSpace |
Supported |
| Detail |
An active and open development community, but the code base is complex. |
| Fedora |
Supported |
| Detail |
Limited to the development team at Cornell and Virginia until relatively recently. This is beginning to change as the original funding for the Fedora project expired in September 2007 and the project directors have put measures (e.g. governance, working groups, opening of SVN repository on Sourceforge, coding rules, etc.) in place to smooth the transition from funded project to the Fedora Commons open source community. |