Momentum One Academy · glossary
OneLake
Definition
OneLake is Microsoft Fabric’s tenant-wide data foundation: each Fabric tenant has one logical lake for storing, organizing, governing, and accessing analytics data. It is built on Azure Data Lake Storage and uses open table formats such as Delta Parquet, so Fabric workloads can work with the same underlying data rather than maintaining separate copies. 1
Where it sits in Fabric
OneLake sits beneath the Fabric workspace and item model. Workspaces provide distributed ownership and access boundaries; items such as lakehouses, warehouses, eventhouses, and KQL databases use OneLake for their data. A lakehouse exposes Tables and Files over that storage, with Spark, its SQL analytics endpoint, pipelines, and Power BI working from the same governed foundation. 1 2
A shortcut is a reference in OneLake to data in another OneLake location or an external store. It can make data available without ingesting another copy. Direct Lake is a Power BI semantic-model storage mode that reads Delta tables in OneLake into memory for fast interactive analysis; it refreshes table metadata rather than rebuilding a full imported copy. 3 4
When to use it
Use OneLake when a Fabric solution needs a common, governed place for data engineering, SQL analytics, BI, and other workloads. Prefer a shortcut when the source should remain in its current location but Fabric users need live access. For large curated Delta tables, evaluate Direct Lake when fast reporting and low-latency synchronization matter and upstream preparation can be done in the lakehouse or warehouse. 1 3 4
Practical example
A team lands CRM extracts and clickstream files in a bronze lakehouse, transforms them into curated Delta tables, and builds a gold layer for reporting. A second lakehouse creates a shortcut to an approved dataset in another workspace instead of copying it. A Power BI semantic model then uses Direct Lake over the gold tables, allowing reports to use the OneLake data foundation without a separate full import.
Production caution
Treat a shortcut as a live pointer, not an ownership or backup boundary. Moving or deleting the target path can break the shortcut, and deleting files through a shortcut can delete the source data when the caller has target write permission. Test target permissions, identity behavior, and recovery procedures before exposing shortcut paths to production consumers. 3
Workshop connections
In Module 1, place OneLake beneath Fabric’s tenant, workspace, and item architecture. In Module 2, connect the lakehouse’s Tables and Files experience to the shared storage foundation. In Module 8, use shortcuts to discuss access without unnecessary data movement. In Module 10, use Direct Lake to connect curated OneLake Delta tables to performant Power BI semantic models.
Ready to apply these patterns in a real delivery? Explore the Momentum One Academy Fabric Production Workshop.
References
Plan your workshop