Data Virtualization comes with a single ModeShape repository instance that is used for a very specific purpose: query and analysis of the metadata described by DataVirt models and virtual databases. In DataVirt, you can upload models into the repository, and they are automatically sequenced into graphs of JCR nodes that represent the models' relational columns, tables, schemas, keys, indexes, etc. DataVirt exposes this structure via a models that you can embed in a VDB, allowing your application to query the metadata of your other models through DataVirt.
It's a bit of a complex use case, because it really allows your applications to drive their behavior by dynamically analyzing the structure of the virtual databases and the underlying data systems.
Thanks, Randall. Then at this beginner stage, I probably don't need to dive into ModeShape as part of my DataVirt studies but it looks like it warrants further investigation for repository pattern implementations. It's on my very long list of JBOSS products to investigate.
That's a good plan. First master all the data virtualization power that's available for all the many sources, and only then consider how to take it to the next step with accessing all the rich metadata that DataVirt makes available and how to use that within your own apps.