How Databases are Handled

Databases (DBs) (and the Teams that own them) can only be created at the Hub, and all databases must always exist at the Hub. Ordinary (that is, non-system) databases (and the Teams that own them) are then allocated to the Satellites where they need to exist. (This does not apply to all extracts. Working extracts can only be created at Satellites. Ordinary extracts can either be created at the Hub and then allocated, or the Satellite can request their creation.)

When the teams and databases have been allocated, the Hub Administrator can make each database Primary at the location which needs write access.

Each location will have its own Multiple Databases (MDBs), Users and module definitions, stored in its own System database and propagated to all other locations ‑ see next section for details. MDBs and Users can be created and maintained by the System Administrator for the location: this will either be the local System Administrator or the System Administrator at a remote location ‑ see next section for details. They are used in the same way as MDBs and users in standard projects, although users who need write access to a database will have to be members of the team that owns the database.

Changes are made to databases at their Primary location. The updates are then propagated to all the other locations to which the database is allocated. As copies of all System databases exist at all locations, changes to these databases are propagated to every location.