Powering Seamless Data Experiences through Fabric Mirroring

Enterprises today are managing increasingly complex data ecosystems. Teams often use Azure Databricks for advanced analytics and machine learning, while others rely on Microsoft Fabric and Power BI for reporting and business intelligence. However, the lack of seamless interoperability between these platforms creates data silos, duplicate copies, delayed reporting, and governance gaps. This fragmented landscape poses significant challenges to data teams striving for agility, accuracy, and trust in their data-driven decision-making.

Microsoft’s introduction of Mirroring Azure Databricks Unity Catalog in Microsoft Fabric, now generally available, offers a groundbreaking solution to this problem. It enables organizations to connect Databricks and Fabric into a unified lakehouse architecture without moving or duplicating data.

Building a Unified, Open Lakehouse Stack

With this new capability, Microsoft Fabric and Azure Databricks can now interoperate directly through OneLake Fabric’s unified data lake. OneLake allows organizations to store all data in a single, open, Delta Parquet format and make it accessible across multiple tools and platforms. This promotes collaboration, consistency, and security across data workflows.

By leveraging Unity Catalog mirroring, enterprises can now mirror their existing Azure Databricks catalog including all managed and external Delta tables into Microsoft Fabric. This connection does not copy the data; rather, it creates a real-time metadata sync that ensures changes in Databricks, such as schema updates or new tables, are instantly reflected in Fabric.

What Mirroring Enables

The mirroring feature allows organizations to:

• Access real-time Databricks data in Fabric: Business users can instantly analyze live Databricks datasets in Power BI using Direct Lake mode or SQL queries within Fabric.
• Eliminate data duplication and latency: Because the data remains in Databricks, there is no need for time-consuming and costly ETL jobs or storage replication.
• Maintain governance and security: Unity Catalog’s fine-grained access policies are respected in Fabric, and OneLake’s security framework integrates seamlessly.
• Accelerate collaboration across teams: Data scientists working in Databricks and business analysts working in Power BI can collaborate on the same live datasets, governed under a single control plane.

Key Capabilities and How It Works

Mirroring is simple to configure and operate. Users can mirror entire Unity Catalog catalogs, schemas, or specific tables from Databricks into Microsoft Fabric by providing their Databricks workspace credentials. Once linked, Fabric continuously monitors and reflects metadata updates from Databricks.

The mirrored tables are made available in Fabric’s SQL Analytics Endpoint, meaning users can:

• Query the data using T-SQL in notebooks.
• Build Power BI datasets without importing or copying data.
• Create semantic models that remain connected to the live Databricks backend.

There is no requirement for a live Databricks cluster during query execution, as Fabric reads directly from the storage layer using the mirrored metadata.

What’s New in the GA Release

The general availability (GA) release of Mirroring brings several enterprise-grade enhancements:

• Firewall-enabled ADLS access: Allows organizations to restrict data access to trusted networks, supporting more stringent security postures.
• Public APIs for CI/CD automation: Teams can now programmatically create, manage, and update mirrored catalog items, making it easier to integrate mirroring into DevOps pipelines.
• OneLake security enforcement: Deep integration with OneLake ensures that access controls, authentication, and data masking policies are honoured.

These capabilities make Mirroring production-ready for enterprises requiring high availability, robust governance, and secure collaboration.

Real-World Use Case: The Adecco Group

One of the early adopters of this capability, The Adecco Group, has already implemented Mirroring to enable real-time analytics on Azure Databricks data within Microsoft Fabric. By mirroring Unity Catalog data, they eliminated delays caused by ETL jobs and provided both business users and application developers with live access to governed datasets. This significantly improved the speed and reliability of their reporting and data products.

According to Guillaume Berthier, Cloud Solution Architect at Adecco, Mirroring is a “game-changer” that enables both Power BI reporting and GraphQL-based APIs to operate on a single, up-to-date data source.

What’s Next on the Roadmap

Microsoft has outlined several enhancements coming to Mirroring in future releases:

• Support for Delta Sharing and federated tables: Enabling broader collaboration across external platforms.
• Federated views and streaming datasets: Supporting more complex and dynamic analytics scenarios.
• Policy-aware mirroring: Integrating row-level security (RLS), column-level masking (CLM), and semantic inheritance across platforms.

These features will further strengthen the position of Fabric and Databricks as core components of a unified, enterprise-grade data platform.

Final Thoughts

Mirroring Azure Databricks Unity Catalog into Microsoft Fabric represents a fundamental shift in how organizations access and govern data. It eliminates silos, reduces operational complexity, and enables real-time collaboration across teams without compromising on governance or performance.

At Quadrant Technologies, we help organizations simplify data access and governance by leveraging capabilities like Azure Databricks Unity Catalog mirroring in Microsoft Fabric. To learn more or speak with our experts, please reach out to us at marcomms@quadranttechnologies.com