If you live in the SQL Server world, you'll eventually hear someone say 'We’re moving to Fabric.'
For a lot of DBAs and data folks, that sentence is immediately followed by:
- Is Fabric just Power BI with better marketing?
- Is it replacing Synapse? Data Factory? Warehouses?
- Where does SQL Server fit in all of this?
... or, Fabric? What is that?
This post is the first in a short series where I’ll walk through Fabric in simple terms, from a data-person’s point of view: what it is, where it lives, and where it may actually useful.
At a very high level, Fabric is Microsoft’s unified analytics platform – a single, SaaS-based environment that pulls together things we used to think of as separate:
- Power BI (BI and reporting)
- Data Factory (pipelines, data integration)
- Data engineering (Spark, notebooks, lakehouses)
- Data warehouse (SQL over analytical storage)
- Real-time analytics / streaming
- Data science / ML
- Databases (operational data in a Fabric-first world)
All of those land on top of a single storage foundation called OneLake. Call it OneLake + a bunch of analytics engines + Power BI, delivered as a SaaS service.
OneLake: OneDrive for data
Microsoft likes to call OneLake OneDrive for data. I think of it more like the place where all the analytics engines are supposed to meet.
If you’re used to juggling multiple storage accounts, subscriptions, and resource groups, Fabric tries to hide that. Every Fabric tenant gets OneLake – a single, logical data lake for analytics data. It is a unified, SaaS-based, multi-tenant storage layer that supports every Microsoft Fabric workload.
Think of it this way:
- Instead of managing separate data lakes, you get one logical lake.
- Different engines (ie., Spark, SQL, Power BI, Real-Time) all read from that same lake.
- You don’t have to spin up storage accounts, containers, or file systems – Fabric does this for you.
LAKEHOUSE, WAREHOUSE, AND WHY YOU SHOULD CARE
Once you have OneLake, Fabric gives you different shapes of compute and metadata on top of it. Two you’ll hear a lot:
- Lakehouse – files and tables in OneLake, with Spark and a SQL endpoint over Delta tables.
- Warehouse – a more traditional, SQL-first analytics experience, still backed by the same storage layer.
Both live in the same OneLake, and both are accessed through Fabric workspaces, alongside Power BI reports, dataflows, pipelines, and more.
Where does SQL Server fit in?
If you spend your days with SQL Server on-prem or Azure SQL, Fabric doesn’t replace that overnight. Instead, think of it as a place to:
- Land data from your SQL Server / Azure SQL / other sources into OneLake.
- Shape that data in a lakehouse or warehouse.
- Serve it back out through Power BI, direct lake, and other Fabric workloads.
Over time, you’ll see more patterns like:
- Mirroring – continuously replicating data from Azure SQL, Cosmos DB, or other systems into Fabric.
- Zero-ETL style integration within the Fabric world.
- Less movement between random storage accounts, more consolidation in OneLake.
For now, Fabric is more about where your analytics land than where your OLTP system lives. Your SQL Servers aren’t going anywhere – but the reporting, analytics, and data science layers may shift into Fabric.
When would I even consider Fabric?
A few situations where Fabric starts to make more sense than just adding another SQL box:
- You already live in Power BI, and your reports depend on many different data sources.
- You have multiple competing data platforms – some Synapse here, some Databricks there, some random data lake in one subscription.
- You are spending more time copying data around than actually analyzing it. Yeah - this is totally you.
- You’re being asked for real-time or near real-time analytics, not nightly batch only.
Fabric doesn’t magically fix modeling, governance, or bad schema design -- but what it can do is give you one place for analytics data, with multiple engines sharing the same lake.
What's next in this series?
In the follow up Fabric posts, I’ll get out of the marketing slides and focus more on practical views and usage:
- How to picture OneLake + lakehouse + warehouse if you’re coming from SQL Server.
- How Fabric thinks about items, workspaces, and capacities.
- Where Fabric can compliment (not replace) the SQL Servers you're already managing.
- SQL Database in Microsoft Fabric.
- What Microsoft isn't telling us.
More to Read:
No comments:
Post a Comment