Fabric is composed of multiple workloads, each designed for a specific part of the analytics lifecycle. What makes Fabric unique is that all workloads operate on the same underlying data — no copies, no silos, no synchronization.

The 7 Fabric Workloads

1. Data Engineering (Lakehouses, Notebooks, Pipelines)

Build medallion architectures, transform data, orchestrate ingestion.

2. Data Science (ML Models, Notebooks)

Train models directly on Delta tables without moving data.

3. Real‑Time Analytics (Event Streams)

Ingest and analyze streaming data with low latency.

4. Data Warehousing (Fabric Warehouse)

A fully managed SQL warehouse engine for high-concurrency enterprise workloads.

5. Business Intelligence (Power BI)

Direct Lake models, dashboards, and reports — always current, no refresh needed.

6. Governance (Microsoft Purview)

Lineage, sensitivity labels, access control — built into the platform.

7. Monitoring & Operations

Built‑in observability across all workloads.

Lakehouses — The Heart of Fabric’s Architecture

The Lakehouse is Fabric’s core data engineering construct. It combines the flexibility of a data lake with the reliability of a warehouse.

Key components of a Lakehouse:

• Files — raw data, logs, unstructured assets
• Tables — Delta Lake tables for structured data
• SQL Endpoint — query Delta tables using T‑SQL
• Notebooks — PySpark, SQL, and data science workflows

Why Lakehouses matter:

• They support medallion architecture naturally.
• They unify Spark and SQL workloads.
• They eliminate the need for separate lake + warehouse systems.
• They integrate seamlessly with Power BI via Direct Lake.

A well‑designed Lakehouse becomes the backbone of your entire analytics estate.

Fabric Warehouse — SQL for the Modern Era

Fabric Warehouse is a fully managed SQL engine built for high concurrency, fast query performance, and enterprise workloads. It complements Lakehouses by providing a familiar SQL‑first experience for analysts and BI developers — and connects directly to Power BI via Direct Lake.

Fabric is not a collection of services — it is a single integrated platform.