The primary use case is for people who use an embedded engine in their application. They can then deploy the standalone webapp next to it, point it to the same database, and are good to go without extending their application.
Sure, it can be used in combination with a shared engine as well, especially when the shared engine runs on a different application server.
This may work with some limitations. Be aware that access authorizations are managed in the engine database. So if you have two users with the same ID in LDAP 1 and LDAP 2, you cannot give them different user-based authorizations. Same for groups.
Best regards,
Thorben
edit regarding the last point: maybe multi-tenancy can be used to distinguish those users, but I haven’t thought this through.
edit2: never mind, authorizations are not tenant-specific.