One of the main themes from my SQL Saturday presentation in December was the ongoing enhancements to SSAS Tabular and how this impacts the decision to use Tabular or Multidimensional.
My slides from this presentation are available here:
On Slide #10 of this document there’s a Word document (top right) that contains a feature matrix – the rows representing Multidimensional features and the columns representing versions of Tabular since SSAS 2012 (1100 CL).
The main takeaways with several slides and this document was A) Tabular is now (2016, 1200 CL) ready to handle projects of high scale and complexity and B) given the roadmap, Power BI, and Azure Analysis Services there are other longer term advantages to choosing Tabular for almost all new projects.
The following feature matrix document has been updated to include SSAS Tabular 2017 features through CTP 2.1:
SSAS Tabular 2017 adds drillthrough and ragged hierarchy support among other features that further close remaining gaps with Multidimensional. At this point, even if a project requires or benefits from the few features exclusive to Multidimensional mode such as Writeback or Unary operators I would generally look to work around and offset this given the many other advantages of Tabular such as M expressions and TMSL.
Further, even if a fairly rich multidimensional model is in production and is highly utilized, I would still evaluate developing a Tabular model to migrate to over time. There are no signs that Multidimensional will be deprecated but the main conversation is now Tabular Import vs. Tabular DirectQuery and investments in Power BI flow to Tabular.
*I wouldn’t be surprised if one of the main announcements at Data Insights Summit is the ability to import a Power BI Desktop model to SSAS 2017 as a 1400 CL model.