Adapter

In OBP, an Adapter is an out of process component that sits between OBP and a bank's systems of record (Core Banking System, Payment System, or database) and translates between them.

An Adapter is paired with an Indirect Connector: the Connector inside OBP turns OBP function calls into messages and sends them over a transport (for example RabbitMQ, Akka, or a stored procedure call); the Adapter receives those messages, talks to the CBS, and returns a response in the agreed Outbound / Inbound message format.

i.e. OBP API -> Connector -> Adapter -> CBS

Key properties:

It runs outside OBP, in its own process, typically on the bank's side.
It can be written in any language, as long as it respects the message format published in the Message Docs for the relevant Connector. This is the main advantage over a Direct Connector, which must be written in a JVM language.
It usually contains bank specific integration code: the data access, field mappings, identifier translation, and quirks of that one bank's CBS. As a result, each bank typically has its own Adapter build.
The Adapter is responsible for emitting OBP shaped values where required (for example a UUID shaped ACCOUNT_ID mapped to the underlying core banking account number).

For worked examples of writing an Adapter, see Adapter.Akka.Intro and Adapter.Stored_Procedure.Intro.