If you have been following the PASS summit notes coming from the attendees, you would have noticed an unmatched sense of pandemonium in the SSAS community. The reason is none other than the announcement of Microsoft's decision to move from the traditional multidimensional SSAS database to the new in-memory Vertipaq BISM.
My first source of this news was from Chris Webb’s blog - PASS Summit Day 2 and after reading it, I was in a state of shock for quite some time. It seemed hard to digest that Microsoft could be doing this to one of their most successful tools in the market. This could have been aggravated by the realization that one of my core skills was going to get obsolete and the hesitation or resistance to change from my comfort zone. Even Teo Lachev seemed to mirror the moderate disappointment that seemed to be floating around in Chris’ writeup in his own blog - The Battle of the Models. It wasn’t disappointment everywhere as experts like Boyan Penev (Thoughts on BISM, SSAS and MDX) and Siddharth Mehta (SQL Server Denali - Boost to SSIS career, Start of End to SSAS and MDX career?) have tried to find reason with this development and at least my understanding is that they are more on the happier side.
After this feeling had finally sunk in, I decided to write a pre-mature obituary for one of my favourite tools in my company’s technology page as given below:-
Microsoft has come out with an official reply now in the SQL Server Team Blog - Analysis Services – Roadmap for SQL Server “Denali” and Beyond. The blog has tried to put out the fears by underlining that the new BISM model does not replace the traditional UDM model. This is also evident from the model diagram they have posted (given below also) where we can see the relational and multidimensional data models existing side by side.
Some salient points of the blog are:-