Boomi DCP runs on a separate server from the Hadoop nodes. It should reside on the same network segment as the Hadoop clusters, to minimize latency. The server requires both edge and cluster network access in order to accommodate requests and access the data.

The Boomi server can be added to an existing Hadoop cluster or in a new installation. Adding it to a cluster can be done using the Hadoop user interface, and is as easy as adding any other node. A few connectors are added to the Boomi interface.

Boomi can discover the data sources in the cluster through the Unifi portion of the stack. It can connect to the existing Hadoop cluster using one of the provided plug-ins, such as the HDFS plug-in. Data discovery and ingestion can occur once the connection is established. As the data source is added, Boomi can walk through the data and generate catalog information for later use.

The robust security model enables fine-grained control over which portions of data users can access.