Starting my Journey in Fabric

This week I’m starting an interesting new journey with Microsoft Fabric. At my company now we have a solid data architecture with Synapse Analytics, dedicated SQL Pool and Power BI. This is what we would like to move to Fabric shortly.

Why Fabric?

Fabric really seems to be the strategic direction for Microsoft for the data domain. Although today we can use Power BI, Synapse and other data services separately based on our choice, but couple of years or months later probably it won’t be the case. Fabric is presented as a platforms that’s ultimately consolidating all those data services in the same place.

Second reason for Fabric is it’s Onelake concept that’s bringing analytical solutions closer to source systems in time. With the current ecosystem the data is mostly running through two steps, one when it is ingested in the data warehouse and the second is when the Power BI semantic model is refreshed. With the Onelake concept we can probably have only one step to be done that would decrease the gap against the source systems.

The third reason looks to be the financials. Synapse Analytics is currently pretty expensive, but it seems that Fabric – although it is not cheap at all – is a bit better option from a biling standpoint. At least for the first site, but I’m really looking forward for the final costs.

How to start?

I’ll start the things with selecting a pilot project, that will be our Salesforce dataset. I’m going to rework the data ingestion from our Salesforce system trying to make it the closest as possible for our current one. More presicely trying to make it dynamic to have it sustainable, and also solving it to move the less data that’s necessary by creating an incremental data load.

Once the data is there, I’ll continue with the modelling piece.

What am I the most excited about?

Definitely the online modelling engine. I’m really looking forward how I can start using the online modelling of the DWH tables and use the model for reporting. I’d like to understand it’s behavior and it’s performance if this solution is really a scalable, enterprise level architecture that can feed 50+ final reports in several different domains.

What can you find shortly in my blog?

In this Fabric section I’ll create short blog posts about the most important differences I’m facing against the former data solutions in Synapse Analytics, Databricks, Data Factory and Power BI.

Let’s start 🙂