While integrating any enterprise system (cloud or on-premise) decisions are made whether data or process flow should be real time or in scheduled intervals using batch mode or polling mode.
Of the two approaches, obviously, the real time data feed is much more beneficial as there will be no delay in data availability and also notifying on any integration errors.
Most of the iPaaS tools provide API management using which you can create/consume API for various integration requirements in which data/process flow will happen in real time.
Boomi has an excellent feature of API management using which you can create rest API end points that can help you achieve various tasks from reading/writing into databases, processing messages and operations.
As Boomi can connect to both cloud apps and on-premise systems, your API in Boomi can be very helpful in interacting with legacy on-premise systems. This is a common use case of Boomi where it connects to an internal system like SAP or JD Edwards and exposes API to get/update data into those systems.
In my post “Cloud App to On Premise Legacy System Integration” you see how API can be leveraged to communicate with a server lying internally in a company network.
Also in my post “Build Cheaper Call Center Solution using AWS and Dell Boomi” you can find out how Rest API can be used to set up a system of truth that can be used to retrieve information from various systems in the organization.