Introduction to Microsoft Fabric

Hello All, There a lot of talks, blogs, news, and the hype of fabric in Microsoft and data world. Now all of them are confused and very much lost in the Fabric world. I will try to answer this questions about,

Why Fabric?

Fabric is the simple one stop shop for data enthusiasts that can make life easier. We can control access, using machine learning models, store data with very easy ETL, create measures in fabric workspace, use data warehousing to make the data more readable. Create or manage data pipelines, data flows, and warehouses within the workspace.

How to make Fabric most useful?

What we can expecting Fabric architecture?

1. Power BI Architecture:
   

In Power BI, we have workspaces and we connect data sources directly to Power BI. below image gives an overview of the normal architecture:

Previous power Bi Architecture

In Power BI architecture, we directly connect the data sources to the Power BI. It might be that we are connecting different data sources in the desktop app and then use it to model, create relationship, build DAX measures, and many more operations within Power BI. This architecture is good, but it has some drawbacks within it. Some drawback are:

a. Data Quality: The quality of data matters as we connect to the sources (SQL server, Snowflake, API, Websites, SharePoint. etc). It is very difficult to make the connections to have good connections or credentials. Tis sources can have errors, refresh failure, and different credentials issue.

b. No ML, AI, integration: In this type, there is no integration of the models, AI, and ML. Thus the businesses cant use the models on their data to make insights.

2. Fabric Architecture

Fabric architecture, to get data into fabric first then load to Power BI

In this section, we are going to discuss the Fabric option or architecture. The main aim to introduce the fabric is to simplify and add on the machine learning and analytics capabilities in the Power BI version. There are 3 types of data that can be stored in the Fabric. The 3 types are data warehouse, data lake, and KQL databases. These types of data storage within fabric gives various options to store and flexibility.

Some strategies are:

a. Workspace and access controls: In fabric we can manage the workspace level, item level, and object level (table), and row level.

b. Data access: There are 3 types how we can get data into fabric or access. First is ingestion, where we copy the data using connector or API. Second is mirror, where we mirror the data from SQL server. Third is shortcut, where we use the shortcut to get the data. There is option to get the AWS or on-premises as well.

c. Ensure that the data has quality. The data validation strategies in Fabric is very important features, so that we can sleep at night with peace.