In order to perform functional decomposition of a "big blob" of a database the first step we wanted to take is to separate different modules of the database into separate schemas on the same database instance. We think we want to divide the single schema (180 tables) into probably 8 schemas. Over time each table will be moved to the right schema "where it belongs".
Are there any performance implications we should be aware of, related to foreign keys "crossing" schemas?
We are planning to run the entire database within the same instance/database; we only want to segregate by schemas at this point in time.
