I think cultural heritage organizations need research on how to work with multiple databases. We have many systems such as DSpace, Omeka, ArchivesSpace, and Collective Memory. Each of these have very good use cases, but none is a complete solution. Instead of trying to create a one-size-fits all database, we need research into building systems of systems. We also need education for institutions to understand how and why to do it.
I'd also be interested in comparisons of what digital preservation actions these systems and what needs to be provided by another system, e.g. Archivematica.