What you need to know now about the new SaaS platform Fabric
Microsoft Fabric is shaking up the industry. Microsoft boss Satya Nadella describes his company's latest coup as "the biggest launch of a data product from Microsoft since the launch of SQL Server". And he's not exaggerating. We explain why Microsoft Fabric will change your daily data analysis forever.
Because no matter whether you are a data engineer, data analyst, data scientist or business user: Microsoft Fabric brings you, your data team and all employees in your company even closer together when you work with your data. The keyword is collaboration. After all, it has never been so easy to gather all your data in a single place, i.e. a single source of truth, and access it easily and, above all, intuitively. This article clarifies your most important questions about Microsoft Fabric and gives you a solid initial overview of possible areas of application.
What exactly is Microsoft Fabric?
The official documentation states: "Microsoft Fabric is an all-in-one enterprise analytics solution that covers everything from data movement to data science, real-time analytics and business intelligence." All data and analysis products are combined in one product and one data lake: OneLake. Fabric is implemented as a SaaS service for all types of data, analytics and BI functions, including: Data Integration, Business Intelligence, Synapse Data Engineering, Data Warehousing, Data Science and Real-time Analytics. The platform offers comprehensive analytics capabilities designed for seamless collaboration. Each application is tailored to a specific persona and a specific task.
Image source: Microsoft
Why is Microsoft Fabric shaking up the industry?
Although Microsoft Fabric is still in preview and is expected to be officially launched in July, there is already widespread enthusiasm in the industry. The entire data journey, from data collection and storage to transformation and analysis, takes place on a single SaaS platform and on the basis of a new data lake called OneLake. All users work on the same database and save their data as Parquet Delta files, i.e. they use the same data independently of each other for their own tasks in the data journey, whether they are data engineers, data analysts or report consumers. The application is intuitive. Not only can all of a company's employees be involved, they can also work together collaboratively - without technical hurdles or organizational boundaries. This has never been seen before in this form.
What is Microsoft Fabric based on?
Azure Synapse and Power BI are moving even closer together in Microsoft Fabric. Power BI users will quickly get used to Fabric. This is because the SaaS foundation on which Fabric is built is the Power BI service. Fabric workspaces behave like Power BI workspaces, but with more available element types. Navigation works in exactly the same way as in Power BI. Users can collaborate and share Fabric items and Fabric workspaces in the same way as with Power BI. New fabric workloads use the capacity-based calculation model of Power BI Premium. Fabric management also works like Power BI management. The Fabric management portal is the evolution of the Power BI management portal.
The new Fabric workloads are basically updates and improvements to existing Azure data services. These include the established PaaS services Azure Synapse, Azure Data Factory and Azure Data Explorer. They all now run on the common SaaS basis of Fabric. Users who are already familiar with these services will immediately get to grips with the new functions in Fabric. Working with pipelines and notebooks or writing SQL and KQL queries works the same way in Fabric, just in a new context.
What is OneLake?
OneLake is a single, unified, logical data lake for the entire organization. It is reminiscent of OneDrive. OneLake is also automatically deployed with each Microsoft Fabric tenant and is designed to be the single location for all of an organization's analytics data. Where previously data may have been moved back and forth across different services and using separate analytics tools, in Fabric it is all in OneLake. OneLake is the heart of Fabric, so to speak.
OneLake is built on ADLS Gen2. Users can save any file type and use the same APIs they use to connect to ADLS Gen2. All data can be used at any time, wherever and in whatever form it is needed. The data is only recorded and stored once. Users therefore access the central data, process it in the warehouse with SQL, for example, or analyze it as a table data set in Power BI, while the changes to the data are saved as separate Parquet delta files for further processing. Storage and compute are separate - a huge advantage for Microsoft Fabric, as this increases the speed of further processing enormously and simplifies the handling of large data sets many times over, especially when we consider the processing of real-time data.
Can I use Microsoft Fabric for AI applications?
Fabric is fully connected to the Azure OpenAI service. With Copilot, users can use their natural language to create data flows and data pipelines, code or even generate entire functions. AI models can also be created or results visualized here.
Who can work with Microsoft Fabric?
Everyone. Data scientists, data engineers, BI experts, business users, you and me. Until now, every user has worked with their own tools and knowledge, each with their own strengths and weaknesses. This meant that most projects required a considerable amount of integration work to ensure that the output of the respective tools and the linking of the results with each other worked throughout the entire value chain. Errors were inevitable, enormous coordination processes were necessary and compromises had to be made.
In Microsoft Fabric, everyone works together on a single database. Each user uses data sets for his or her own purposes and combines them with the attributes that are important for the respective context. Everyone works natively with the same data in OneLake. This creates added value because no one has to deal with creating integrations so that different tools can work together. Every user and every team can create their own workspaces and share them with others as required. Data from one workspace can be integrated into the other and vice versa. Data is simply saved once in OneLake and can be used as often as required in the subsequent analytics process without duplication.
Do I need Microsoft Fabric?
Perhaps. In any case, the hype is justified. For years, users have been calling for just such a platform that can map a seamless data journey, from integration to data dashboarding. The DataOps Cloud Y42 is currently trying to establish such a status quo in the Modern Data Stack consisting of Snowflake, Fivetran and dbt. Azure Fabric now offers exactly that from Microsoft. There is no longer any need for a third-party provider. The prerequisite is, of course, that everything works as advertised.
My advice: Wait for the preview. Continue to test whether the technological offering makes sense for your use case. Take your cue from the name Azure Fabric: start by loosely combining the first layers of your data analysis and gradually strengthen your basics into a solid structure. Take a look at the development of Fabric so that you don't lose the thread. It can be a game changer.
