Microsoft OneLake is the unified logical data-lake storage layer underlying Microsoft Fabric, a SaaS analytics platform spanning data engineering, data science, real-time analytics, and business intelligence. It centralizes access while retaining compatibility with open table formats and external storage. section 5.4.5
OneLake’s architecture combines interoperability with virtualized access:
- Supports Apache Iceberg and Delta Lake, allowing external engines and tools to interoperate with Fabric data. section 5.4.5
- Creates shortcuts to ADLS Gen2, Amazon S3, Google Cloud Storage, and S3-compatible services without physically moving data. section 5.4.5 section 8.6.1
- Pushes filters and column projections toward remote storage while keeping most joins and aggregations local to the querying engine. section 8.6.1
Virtualized external dataset
An S3 dataset can appear as a native OneLake table through a shortcut. Queries can request only required columns or filtered data, avoiding a full data copy while limiting cross-system query expressiveness. section 8.6.1
