Fabric and Databricks Dual Deployment Architecture

An option for organizations interested in both Azure Databricks and Fabric is deploying both services and sharing data between them, or assisting during migration. This illustration shows an environment with lakehouses created and managed in Azure Databricks (left) where the Databricks Unity Catalog provides user access control and a metastore, a central location that gathers and shares schema details (tables and columns) for each lakehouse.

Data Sharing

Key to a dual-deployment environment are data-sharing features (in preview) that provide security controls and in-place query features that remove the need to duplicate data.

OneLake shortcut to Databricks connects Fabric users to Databricks Unity Catalogs, providing access to Databricks lakehouse schemas. Fabric users can build queries in Fabric Data Engineering and other Fabric services that honor the Databricks security model and query data in-place. This feature will likely interest customers with high-end security requirements who are more comfortable with Databricks advanced security features and existing Databricks customers who want to retain their Databricks environment but use Fabric’s other services, such as Power BI.