What Is Microsoft Fabric? My Point of View

Bringing the Data Platform Workspace and the Power Platform workspace Together

Over the last 18 months or so I’ve been watching and supporting the development of Microsoft’s new Software as a Service (SaaS) offering that has now been named at MSBuild 2023 as Microsoft Fabric. As a Microsoft MVP I’ve be aware of this offering and had access to the private preview for a long time. So, it’s great to see the public preview announcements finally made. If you missed them, check out the MSBuild content here: https://build.microsoft.com

Moving on and assuming you have seen the event sessions, I want to give you my point of view to help explain what Microsoft Fabric is. Firstly, lets clear up call out was terminology to support this understanding. Is this software offering a resource, service, platform, or solution? To answer this question, perspective is key, perspective with a timeline (2018 to 2023). We could simply say that Microsoft Fabric is all these things. All things to all data professionals and beyond. But, to understand this, let’s consider the journey Microsoft has been on and how this technology has evolved. I believe this journey is the best way to help explain what Microsoft Fabric is, rather than focusing on all the new and shiny bits.

Note: to support this understanding and evolution I am going to be very deliberate with my graphics, to follow in Microsoft’s footsteps and hopefully clear up some confusion by displaying things a certain way. To be clear, these graphics aren’t perfect and far from what I would consider architecture diagrams. They are simply to inform awareness of the tooling.

2018 – The Core Four

As a data engineer in 2018 and for several years prior, when working in Microsoft Azure to deliver a data platform or to deliver data analytics the following core resources were and are still used.

Storage – Azure Data Lake
Compute – Azure Databricks
Orchestration – Azure Data Factory
Visualisation – Power BI

This combination of Data – Lake, Bricks, Factory gave us a stable architecture to scale and grow use cases. Also, with these core resources we had/have the technical ability to plugin lots of other Azure products to support many different requirements. Including near real-time data processing, dimensional model data warehousing and data science to name a few.

I am more than happy to debate the position of other products like Azure DevOps and Azure Analysis Services. But let’s keep things simple for now, with a data engineering focus.

Also, be mindful on this timeline, Delta Lake was introduced in February 2019 and made open source by Databricks in April that same year.

A data professionals view of the tools in 2018:

2020 – A Unified Platform

At the end of 2019 (November) Microsoft announced the public preview of Azure Synapse Analytics, which a full year later in 2020 became generally available. This offering unified our core resources into a single product, or at least it tried to. This unification wasn’t perfect, and we lacked feature parity compared to the mature equivalent resources listed above. However, the vision Microsoft had was sound. A vision that I think started with a unified set of services in 2005 that we grew to know and love called Microsoft SQL Server. I have previously said, Synapse Analytics is to data processing in Azure, what SQL Server is to data processing on premises. That unification of resources into a single platform offering. Providing this integrated experience for all the technical capabilities we require is certainly desirable. It makes development effort easier and accelerates time to insight for business users.

A data professionals view of the tools in 2020:

2023 – Data Software as a Service (FABRIC)

Now, continuing that vision (started with SQL Server or Synapse). The unification of technical capabilities continues as Microsoft Fabric takes the stage. Our data architecture can be fully integrated with compute and storage simplified into a single offering. Including data serving and visualisations. This is different from the abstraction provided in Synapse because now the workspace is truly unified. Previously the Power BI workspace had to be integrated into the Synapse workspace and the experience was “clunky”.

In addition, we see Delta Lake become the standard format for all data within the Data Lake (One Lake), regardless of the transformation capability you choose to use. This is definitely a game changer for both processing and consumption, reducing all the friction and data movement previously required.

Microsoft Fabric gives us everything we need to deliver data insights in a complete offering. Furthermore, it plays nicely into the industry concepts of Data Fabric and Data Mesh.

A data professionals view of the tools in 2023:

Summary

To summarise, the goal and the problem faced by data professionals has not changed. But now we have another “new” unified platform to deliver insights to our customers. Supporting business growth and informing decisions with the potential time to insight greatly reduced. Using a software as a service offering and feeding on the popularity of the Power Platform workspace to enpower business users.

Hopefully, Microsoft learnt from the less unified experience of Synapse Analytics and this time everything across the product suite works well. Including DevOps and future capabilities such as Data Activator.

An evolution of the tooling over circa 5 years: