OneLake is Fabric's single shared storage area. All the organisation's data lands here and different engines read the same files.
01. What a Lakehouse Is
An approach combining the flexibility of a data lake with the table discipline of a warehouse. Raw files are stored, but table definitions are placed over them and they become queryable with SQL.
02. The Delta Format
Tables are held in the open Delta format. That means the data can also be read by non-Microsoft tools, which reduces vendor lock-in.
03. The Layered Structure
Common practice uses three layers: raw data, cleansed data and a presentation layer where business rules are applied. Without that separation the lake soon becomes a pile of files nobody can describe.
04. Shortcuts
You can reference data in another location without copying it. A table in Azure storage or another workspace can be used without creating a copy. That reduces both storage cost and freshness drift.
05. Governance
One lake means access management is also done in one place. Row- and column-level permissions for sensitive tables should be defined in the lakehouse layer, not rebuilt in every report.
06. For ERP Data
ERP tables are loaded daily into the raw layer and converted into dimension and fact tables in the cleansed layer. Reports see only the presentation layer. That separation stops structural changes in the ERP from breaking reports.