ContentRepository

Why a content repository for Collaborilla?

• Apache Jackrabbit for Java
• No outside connections to other systems
• WebDAV can be (or is?) implemented in Jackrabbit as well
• Versioning is native
• Access control
• Full text searching
• Event monitoring
• Compact deployment
  • No LDAP server, no patching
  • No additional DB (Apache Derby can be used as backend)
  • No dependencies
  • No network traffic between Collaborilla service and repository: faster data access
• LDAP alone is not sufficient
  • difficult (impossible?) to index
  • hard to adapt to new features (discussions, ...)
• Jackrabbit's JAAS (Java Authentication and Authorization Service) can be combined with Shibboleth

Questions

• Which backend should we use?
  • JDBC, Derby, XML, ...
  • Performance?
  • Is the size of blobs limited?
  • Do big data blobs impact the performance?
• Does it make sense to publish directly into the content repository?
• Should we switch to RMI?

Implementation - What Jackrabbit can be used for

• Content repository for Collaborilla
• RDF-information and the dependencies
• Publishing of containers directly to the repository
  • Conzilla to Jackrabbit or
  • Conzilla to Collaborilla to Jackrabbit
  • Usage of native Jackrabbit WebDAV
    • Is there an implementation?
    • How good is the implementation?