What Microsoft Fabric actually is#
Microsoft Fabric is a single SaaS analytics platform covering the whole data lifecycle — ingestion, transformation, stream processing, warehousing, data science, and BI reporting — so you don't stitch together separate Azure services for an end-to-end stack [S1]. Because it's SaaS rather than a pick-your-own PaaS setup, you never touch Azure infrastructure concepts like resource groups, RBAC, Resource Manager, redundancy, or region placement, and you don't need an Azure account to use it [S1].
Fabric organizes its functionality into role-oriented workloads — Power BI, Data Factory, Data Engineering, Data Science, Data Warehouse, Databases, Real-Time Intelligence, Industry Solutions, and the preview Fabric IQ — all in one environment, sharing data and items without duplication [S1]. Underneath every workload sits one shared platform layer used everywhere: OneLake for storage, Copilot for embedded AI, and governance powered by built-in Microsoft Purview [S1].
The vocabulary you need first#
A tenant is one organizational instance of Fabric, mapped one-to-one to a Microsoft Entra tenant. Inside it, workspaces are collaboration containers holding items, drawing compute from a capacity, and controlling access [S2]. An item is any object you create — a lakehouse, notebook, warehouse, eventhouse — and each workload contributes its own item types [S2]. Delta Lake is the standard table format across every workload — data lands as Delta tables by default — letting different engines read each other's output with no conversion [S2].
Getting in and paying for it#
Signing into a Fabric-enabled tenant automatically grants a free per-user license, enough to create and share non-Power-BI items in workspaces backed by an F or Trial capacity [S3]. A workspace's type (formerly "license mode") determines what capacity backs it — an F SKU, a Power BI Premium P SKU, or a shared Pro/PPU pool — which determines which experiences its users can run; only Fabric, Trial, and P-SKU workspace types support the full set [S3].
Security comes built in#
Because Fabric is SaaS, many security, compliance, and governance controls are already built into the platform, needing little or no separate setup — unlike a traditional PaaS data product [S4].
A worked example#
You want to ingest sales data, clean it, and show it to executives. In Fabric, you'd bring the data in with Data Factory, process it in Data Engineering (Spark), and visualize it in Power BI — no manual integration required, because every stage already shares OneLake and the same platform services [S1].
What goes wrong#
- Expecting Azure-style provisioning. There's no resource group or region to configure per service — Fabric is one SaaS surface [S1].
- Confusing workspace, capacity, and tenant. A tenant holds many workspaces; a workspace draws compute from one capacity [S2].
- Assuming the free license covers everything. It only covers non-Power-BI items on F/Trial capacity workspaces [S3].