Understanding Microsoft Fabric - Video 2

[@sushilmarati]

I appreciate the insightful information you've shared. I kindly request that you consider creating a video comparing when to choose Databricks versus Fabric. Such a video would be very beneficial. Thank you for your consideration.

[@ChrisUK70]

Thanks Bryan for such a helpful explanation I was so confused with Fabric then there is Synapse. I always wondered what happened to Sybase, I worked on a Sybase database 24 years ago!

[@sathwikdhanya5400]

Hey! Does it have data catalog and governance ability like unity catalogue

[@BryanCafferky]

Yes. Microsoft Purview serves that function across Azure.

[@chad_the_stud]

Does Google have something similar in GCP ?

[@BryanCafferky]

Unlikely since Fabric is centered around Microsoft products like Power BI.

[@B-Luv]

Great job! The table with the Microsoft speak and translation is very helpful. I have a Medallion architecture question for you. If Power BI is your primary subscriber of all of the data...does it make sense to have a Gold Layer? Today we have been doing a lot of connecting directly to source systems in Power BI and creating the Semantic Model. These are discoverable and reusable. I feel the Semantic Model and Gold Layer serve the same purpose. Would love your thoughts. Also, since you mentioned it...this self service stuff is laughable. It's a full-time job just learning and staying up on all this stuff. The citizen analysts don't have time to think about end to end solutions. Self service in my mind is I the citizen analyst discovering a Semantic model and using that to build a new reports. Maybe add to the Semantic model now that they have mixed query modes.

[@BryanCafferky]

Thanks! The Gold layer is really just the final Semantic layer ready to use for reporting. Usually, its better to have an ETL process build this than to just make Power BI do all the work at reporting time. Bear in mind the Semantic layer is often a Dimensional Model which can be pretty complex to load but offers features like surrogate keys and dimension history. Not sure I understood your question. Does this make sense?

[@B-Luv]

​@BryanCafferky Understood. Currently we have the tradional SQLDW (PDW) with facts and dims(no history) but in practice, I find myself usually needing to do a bunch of additional transformations in Power Query no matter what. I have this maybe crazy notion that the Power BI Semantic model is the true Gold Layer and we should build many small ones to fit different use cases. And each Semantic model should support many reports.

[@BryanCafferky]

@@B-Luv Yeah. I like to think of the Semantic Layer as the DW and not part of PBI b/c it may be used by other tools inc. AI, other reporting or analysis tools, etc. PBI is specific use so I think of PBI as just another reporting user of the DW Semantic Layer. Admittedly MS blurs this line in their product promotion in my opinion.

[@daithisad]

Hero