Forum Discussion
Measures and CalcColumns in Power BI vs Tabular model (replace SSAS with Power BI?)
You need Analysis Services if you have one of the following requirements:
- You need to have row level security
- Your data set is greater than 250MB
- You need to have a fine grained control over how your model is refreshed from the source data (partitioning)
If the aforementioned requirements are not present, then using Power Pivot or Power BI Desktop to build your model and publish it to Power BI is a valid and wonderful choice.
You need Analysis Services if you have one of the following requirements:
- You need to have row level security
- Your data set is greater than 250MB
- You need to have a fine grained control over how your model is refreshed from the source data (partitioning)
If the aforementioned requirements are not present, then using Power Pivot or Power BI Desktop to build your model and publish it to Power BI is a valid and wonderful choice.
Thanks everybody for their input.
Just to clarify: The rationale behind my question is whether I can remove SSAS server and have all the calculated columns, measures relationships and so on created using the Power BI desktop designer instead. This way Power BI would connect straight into the relational DB (the data mart).
My understanding from Andre's answer is that the modelling features of Power BI have some limitations:
- No row level security
- Model cannot be greater than 250MB
- No fine grained control over how model is refreshed from the source data (partitioning)
Am I right to assume that if this limitations are overcome by MSFT, then an SSAS server would no longer be necessary in my case?
Regards,
P.
PS: Apologies for the late response. I got no email alerting me of replies into the thread.