Microsoft Introduces Iceberg Support in Fabric and Collaborates with Snowflake for Enhanced Bi-Directional Data Access

Microsoft is raising the bar in simplifying data management for enterprises. At the ongoing Build 2024 conference, the company announced an expansion of its partnership with Snowflake to enable bi-directional data access for joint customers.

This initiative, which includes support for Apache Iceberg, aims to enhance interoperability between Microsoft Fabric—its comprehensive data analytics platform—and Snowflake’s data cloud. By eliminating the need to maintain multiple data copies across various locations, this collaboration promises significant cost savings for customers.

“We are enhancing this experience by embracing open standards, partnering with industry leaders, and facilitating rapid access to our market innovations,” said Arun Ulagaratchagan, Corporate Vice President of Azure Data at Microsoft.

Understanding Microsoft and Snowflake’s Interoperability

Last year, Microsoft introduced Fabric, a unified platform capable of managing an organization’s data and analytics workloads seamlessly—from data integration and engineering to visualization. Featuring OneLake, a data lake that utilizes open formats like Apache Parquet and Delta Lake, Fabric allows data from Microsoft’s ecosystem and external sources to be consolidated effectively.

“The decision to support open formats, rather than proprietary ones, stems from customers' desire for complete data openness. With formats like Apache Parquet and Delta Lake, hundreds of compatible tools can work with our customers' data,” Ulagaratchagan explained.

Now, Microsoft is further advancing its collaboration with Snowflake by adding support for Apache Iceberg. This integration enables Snowflake to treat OneLake as a native data store, allowing joint customers to store and access a single copy of their data in the Iceberg format on Fabric. Consequently, Iceberg data in OneLake will be accessible through Snowflake.

Previously, the two platforms did not interoperate, necessitating the building of pipelines and maintenance of separate data copies, which consumed time and resources.

With bi-directional storage and access, users will seamlessly integrate their Snowflake data with applications within Microsoft's ecosystem, such as Teams, Excel, Power BI, and Azure AI Studio. Simultaneously, data in Fabric’s OneLake can be utilized in Snowflake to enable cross-cloud AI, application development via Cortex, collaboration, and advanced analytics.

“What customers want is to manage their data, control it, and mix and match technologies to suit their specific needs, and that is what we are delivering,” stated Christian Kleinerman, Executive Vice President of Product at Snowflake.

Rollout Timeline

While the announcement is made, the interoperability experience between Snowflake and Microsoft Fabric is under development and will launch in preview later this year.

“We’ve noticed substantial interest from customers. Development is active, and while we haven't begun testing yet, there’s considerable anticipation from enterprises eager to leverage this innovation,” Arun added.

Microsoft Build is taking place from May 21 to May 23, 2024.