Power BI vs (Azure) SSAS Tabular: when/why do we need SSAS?

[Rubenvw]

Hi,

I am really curious about your opinions.

For me arguments for using SSAS Tabular were mostly based on performance, the ability to create a semantic model based on the DWH for business users and adding RLS. When Power BI is becoming more and more mature, and features are 'connect to Power BI Service' are being released, I do not see so much the need for SSAS Tabular models aymore. In fact, Power BI datasets are technically SSAS Tabular models.

So i was wondering: when/why do we need SSAS Tabular? 

Hopefully you are able to give me some information that will be to useful to make a decision.

Thanks.

[vickyd]

I was about to ask the same question and found this thread. Even I am looking for an answer on why I would need SSAS anymore if refresh frequency is not an issue. 

Few things I am aware of are:-

1. Partitioning

- For massive (>10GB) datasets, SSAS allows partitioning for processing. Don't think Power BI has that yet. 

2. SSRS connectivity

- Within our organization, we still leverage SSRS for a lot of reporting that currently Power BI can't really do. Not sure if SSRS reports can connect to Power BI Data models so if SSRS is needed then SSAS is a better option.

I think in general for Enterprise grade solutions, one should go with SSAS.  

[tringuyenminh92]

Hi @Rubenvw,

As you know, PBI has 2 connection modes (Import mode or Direct query), and DQ could be called Live Connection if your sources are cubes (Tabular or MD). Import mode is in-memory that imports all your data to PBI cloud but there is restriction of data size (10gb for uncompress and 1gb after compressed) and refreshing 8x/day. So when reaching the limitation of data size or need higher refreshing rate, you could think about switch to DQ/live connection.

• Live connection (cubes): will be good choice as high-performance sources for query requests but has latency for processing of cubes and not public your db/dw to MS cloud to compliance specific policies in your company (your cubes could be built on-premise or in private cloud).
• Direct Query (databases): near real-time for line of business applications.