Over the (roughly) last 6 months we all had the ability to play with Microsoft Fabric, when it went Public Preview at Microsoft Build.
Now that the keynote of Microsoft Ignite by CEO Satya Nadella has been delivered, and also the amazing in-depth Fabric session by Amir Netz, Arun Alagaratchagan: Make your data AI ready with Microsoft Fabric and Azure Databricks. Back then at Build, Satya called it:
"...the biggest launch of a data product from Microsoft since the launch of SQL Server!"
Generally available
But now, Satya went one step further and announced GENERAL AVAILABITIY of Fabric!
Also, Copilot in Fabric is now in public preview.
More updates
There's a ton of updates inside Fabric on existing features on the latest Fabric blog here.
And there are also some exciting new features:
- Seamlessly connect your data sources to Fabric
The ability to create shortcuts was already there, where you virtualize data in OneLake without having to move or duplicate that data.. You can create shortcuts to another Lakehouse and Warehouse, but also to files on ADLS or even Amazon S3 or Google storage.
The newest feature just announced is called Mirroring, where you can add and manage existing cloud data warehouses (and databases) in Fabric's Data Warehouse experience. The way this works, is like replication in SQL Server. Fabric replicates a snapshot of that database to OneLake in Delta Parquet files and keeps that in sync in near real time, which relies on the Change Data Capture feature of the underlying source. Initially it's supported for Azure Cosmos DB, Azure SQL DB and Snowflake, more sources will follow next year. - Copilot in Power BI (public preview)
Just be aware that Copilot will be rolling out in stages, - Smart Narrative is an existing visual in Power BI Desktop, now rebranded to Narrative with Copilot
- The November Desktop update let's you generate synonyms for your fields, measures and tables using Copilot.
- In the future, there's also going to be:
- a report creation experience
- a DAX writing experience
- Direct Lake support on Data Warehouse
There's also an update on the size limits of your Fabric capacity and when it will fallback to Direct Query. - Stored credentials for Direct Lake semantic model
You can now specify a fixed identity (like a service principal) for a Direct Lake mode semantic model - Pricing on Fabric is updated!
Reserved pricing is now available, with a rough discount of around 41% off of Pay-As-You-Go pricing. - OneLake integration for Import-mode semantic models is coming!
This allows for a seamless (at least that's what Microsoft claims 😄) integration for your import Power BI Desktop models into OneLake. I wonder if this also implies that you can convert your import report to a Direct Lake mode model afterwards. It's not totally clear to me at this point.
[update on November 16th]
- Power BI Desktop November is also out!
Power BI November 2023 Feature Summary - DAX query view
- Stored credentials for Direct Lake
- Explore
Conclusion
Will I now go all-in on Fabric? "It depends", but probably not 😀
I think it still depends on a case per case basis. Is it a newish customer with not too many investments in other data platform services like Synapse/Data Factory. Then it makes sense to evaluate the requirements and see if it's worthwhile to start with Fabric, considering a lot of features are still in preview, and others are not there yet.
On the other hand, if it's an existing customer with real estate in Synapse and databricks for example, where they implemented a medaillon structure and have infrastructure running there, I'd seriously reconsider if it's worth moving to Fabric. I'm leaning towards a no for now.
Keep a look out on the official Fabric and Power BI blogs and Microsoft Learn for all new content:
I'm updating this post live while the updates are rolling out, so come back later for more updates!😀
