This image shows an overview of the data federation process.

The serving engine initiates a unified query which is divided into subqueries that are sent to different data sources, including data sharing and cloud store sources, cloud and on-premises databases and other endpoints. The serving engine then stores and provides the federated data to data analysts, analytics engines, and data science and AI services, while harvesting metadata in a data catalog. Analytics engines can federate addition sources and provide insights to business users. The data science and AI services provide insights to data scientists.