Fabric is the new data analytics platform launched by Microsoft that combines all the analytics tools organisations need. It is currently in the preview stage.

It combines Azure Data Factory, Synapse Analytics, and Power BI into a single product, but it’s not just a rebranding of the existing Synapse Analytics offer. Instead, it builds on the vision of a unified platform at the UX level in Synapse Analytics. However, the core of the compute engines was also refactored tightly to integrate at the data level.

Fabric is a SaaS solution that aims to lower adoption barriers, reduce setup complexity, and empower every team member with the role-specific experiences they need. Like Power BI, you don’t need to spin up a cluster or decide the size of a database. Everything you need is just there. Azure powers it, but you don’t even need to open the Azure portal to use it.

Fabric is built around OneLake and provides seven core workloads: Data Factory, Synapse Data Engineering, Synapse Data Science, Synapse Data Warehousing, Synapse Real-Time Analytics, Power BI and Data Activator.

The cool thing is that all these engines will work with the same data format: delta parquet, an open standard, meaning no vendor lock-in.

This also means that data written by one engine can be read by any other engine, eliminating the need for multiple copies of the same data. This will reduce complexity, integration and storage costs, and the time from data ingestion to presentation.

Expect to have the Azure OpenAI service tightly integrated with Fabric at every layer to assist users in developing or extracting insights from their data.

“Whom is Fabric intended for?”

On paper, it caters to everyone, but in practice, it depends on your specific use cases. Do you already have an analytics platform? What technologies are you using? If your analytics platform is already in Azure, chances are that some migration paths will soon exist. But it’s still early days for a definitive answer.

“I’m about to start a new analytics project. Should I wait for Fabric General Availability?”

We don’t know how long it will take for Fabric to reach GA. If your analytics project is in Azure and follows the previous answer, all the work will likely be ported to Fabric. If the project doesn’t involve Azure, then the reasons that worked against Azure will most probably also work against Fabric.

“How much does it cost?”

There will be two types of SKU: you can reserve capacity with a monthly or annual subscription, like Power BI Premium. Or opt for Azure’s “pay-as-you-go” approach, with no commitment and billed by the second.
To date, the capacity reservation figures have not been released. The “pay-as-you-go” prices start at 0.36USD/hour or 263.80USD/month for a capacity of 2CU’s (capacity units) and go up to 2048CU’s, which corresponds to 368.64USD/hour or 269,107.20USD/month.
The storage fee is paid separately and is equivalent to the Azure ADLS fee.

“What if I need to scale the capacity?”

With Azure SKUs, you can scale and pause/resume capacity as needed. Not sure how it will work with the subscription SKUs.

As mentioned, Fabric is still in preview, and some of the announced functionalities still need to be made available or are buggy. However, it’s an exciting product, and we are eager to see how it develops. Each Fabric piece deserves a dedicated article (or more), and we’ll try to do that as we explore the platform further.

If you have the opportunity, test the product or contact BI4ALL to see if it meets your needs.