From your perspective, why would someone choose MsFabric over DBX? Stored procedures, the quickstarts and ADF? Seems like costs would be higher in MsFabric from having cluster specific servers.
Good question. 1) Fabric is a end user self service platform whereas Databricks is more of a Data Engineering and AI solution building platform. 2) Fabric is extremely well suited to supporting Power BI reporting and can even eliminate the need to load a Tabular Model., 3) the t-shirt Fabric capacities means you select a fixed compute level that all the Fabric objects and services share so the cost is contained. However, you could find that you need to increase the Fabric capacity at some point. 4) with Fabric, you don't need to manage or even explicitly create the underlying Azure resources, Fabric does it for you and even hides the details. This is extended with the OneLake and shortcuts concepts. Make sense?
@@BryanCafferky Would you then say, that Fabric is more akin to a Modern Data Stack and more focused on the Analytics Engineer role rather than to a dimensional modeled warehouse?
@@mubar2 I think Fabric is aimed at the end user, not the engineer. It gives business users tools to do analysis, build dashboards and build solutions but they will not have the skills to build a a Kimball styled data warehouse in my opinion. They will get something up and running fast.
This is brilliant, thank you!
You're welcome.
From your perspective, why would someone choose MsFabric over DBX?
Stored procedures, the quickstarts and ADF?
Seems like costs would be higher in MsFabric from having cluster specific servers.
Good question. 1) Fabric is a end user self service platform whereas Databricks is more of a Data Engineering and AI solution building platform. 2) Fabric is extremely well suited to supporting Power BI reporting and can even eliminate the need to load a Tabular Model., 3) the t-shirt Fabric capacities means you select a fixed compute level that all the Fabric objects and services share so the cost is contained. However, you could find that you need to increase the Fabric capacity at some point. 4) with Fabric, you don't need to manage or even explicitly create the underlying Azure resources, Fabric does it for you and even hides the details. This is extended with the OneLake and shortcuts concepts. Make sense?
@@BryanCafferky Would you then say, that Fabric is more akin to a Modern Data Stack and more focused on the Analytics Engineer role rather than to a dimensional modeled warehouse?
@@mubar2 I think Fabric is aimed at the end user, not the engineer. It gives business users tools to do analysis, build dashboards and build solutions but they will not have the skills to build a a Kimball styled data warehouse in my opinion. They will get something up and running fast.