We no longer have a built in configuration repository and we rely on container scans to do deploys, so we have lost the ability to internally control the notion of a deployed vdb version.

We should strive for the following:
1. Overwrite of a deployed vdb should not invalidate or otherwise corrupt current connections.
2. Switching to a different vdb as the new default for a particular name should not require two deployments of the same vdb.