We have a Postgres instance hosted in Amazon RDS. The idea is to create some kind of a read-only replica for a separate application. This application requires some modifications of the DB schema - additional indexes, denormalization, etc.
What is the common practice to achieve such goals?