Platform Comparison - Game Over for Databricks?

[@AleksiPartanenTech]

Things have improved with Fabric's pricing model and managed private endpoints are now supported for all capacities and not just F64

[@adammessina1414]

I’m just getting started in my analytics career and I’m investing heavily in learning Fabric!

[@AleksiPartanenTech]

That's a very good plan! Also, it is very important to learn some more foundational skills like SQL, Python, data modelling etc. if you haven't learned the basics of those yet. :)

[@mfayed19]

yes compare them on technical and feature terms especially after the Data + AI summit

[@AleksiPartanenTech]

Ok! I might do that video during upcoming weeks :)

[@LukeGarton-gi4ui]

What would your top 3 Enterprise features that you believe Fabric is missing from the maturity standpoint you discussed?

[@AleksiPartanenTech]

Thanks for your question! In my opinion these are the top 3 issues currently: 1. Preventing public access to OneLake meaning that it should have private endpoints like storage accounts have in Azure. 2. Deployment from one environment to another is still very problematic and will in many cases require manual work. For example, currently deployment parameter overrides are missing. 3. Secret management has many features missing and support for using azure keyvault for secrets is still very limited in many ways I am very certain that all of these issues will be fixed within one year but for now Fabric is not truly ready to be used properly in enterprise context

[@davidlean8674]

While we liked the innovation & ability of Databricks, we found our people were much faster at delivering in a no-code/low-code environment. When you include your people costs ( inc contractors, consultants, Project mgrs) into the equation the dial swings heavily into Microsoft's favour. Especially when you factor in all the additional AWS/Microsoft platform costs; (storage & compute). The other challenge with Databricks is their contract. (a) You need to estimate your annual consumption, which is hard to do. Unused credits don't roll over into the next year. (b) They add 20% support on top of usage charges. (c) some other SAAS tools you may be using base their pricing off the Databricks contract price. I'm not saying Databricks is bad. They are great. They are also more expensive than their direct costs would suggest.

[@AleksiPartanenTech]

Thanks for you comment! I agree with you mostly. My main issue with Microsoft Fabric is that it is not ready yet and many features needed for secure enterprise use are still missing but should be coming at some point.