1.5.2. Why This Pattern Matters for the Exam

The exam frequently presents scenarios requiring you to identify which layer handles a transformation:

• "Deduplicate records" → Silver layer
• "Calculate YTD revenue" → Gold layer
• "Ingest raw JSON from API" → Bronze layer

In Fabric each layer is a lakehouse (or a warehouse for Gold) in OneLake—Microsoft recommends one lakehouse per layer, ideally each in its own workspace, but doesn't require it. There are no storage accounts to provision or integrate: Bronze can use shortcuts to source data in OneLake, ADLS Gen2, S3 or Google Cloud Storage instead of copying it, and Spark, SQL and Power BI all read the same single copy of each layer's Delta tables.

⚠️ Exam Trap: The medallion architecture is a logical pattern, not a Fabric feature. Fabric doesn't enforce it—you implement it through naming conventions and workspace organization. Questions about "configuring bronze layer" are testing your understanding of the concept, not a specific setting.