With SMILA, it is possible to do the data import (crawling) on one SMILA instance, and push the resulting records to (the REST API of) another SMILA instance.
That means, that you can e.g. have one (local) SMILA for processing and indexing, and other SMILA instances on other (remote) machines that import the data of several data sources and push it to the local SMILA.
The remote crawling is configured like a normal crawl job, it only has a different parameter in the UpdatePusherWorker that describes where the imported data should be pushed to.
Hint: You also have to enable the remote access in the (indexing) SMILA instance, otherwise SMILA won't be accessible from the remote machines.