SQL Server 2012 – BI Semantic Model, Multidimensional vs Tabular

The landscape of Business Intelligence changes with SQL Server 2012. Microsoft has introduced a new “model” – the BI Semantic Model.
Don’t get too hung up on this word though, the BI Semantic Model is really just an umbrella terminology that says as long as a source follows the BI Semantic Model, it will be supported in the tools. This is the eventual goal anyway; for now there is still a discrepancy which tools support what.

The multidimensional model is the traditional OLAP model. Data source is a cube, that is processed from a data warehouse that usually follows a star schema. Complex calculations and queries using MDX can be done against it. It also allows write-backs.

The new tabular model is not really a “new model” because it is a relational model. Yes the same model we’ve grown accustomed to using when we query our transactional OLTP databases. The difference is, SSAS Tabular is an in-memory database. This is fueled by the xVelocity engine (previously known as Vertipaq), and it also leverages columnstore indexes. All calculations are in memory, and this makes tabular models really fast.

Continue reading

SQL Server 2012 x-Terms

Just finishing up my presentation for this Saturday’s Vancouver Techfest, and was putting a slide for some of the x-terms in SQL Server 2012. Just thought I’d share it:

xVelocity In Memory Analytics Engine

In SQL Server 2008 R2 – this was introduced as Vertipaq for PowerPivot for Excel (see James Beresford blog about Vertipaq)
In SQL Server 2012 this technology was integrated into Analysis Services.
In SQL Server 2012, it was also rebranded into xVelocity. The reasoning is that this is going to become part of a bigger xVelocity family for next generation performance improvements. Simran Jindal has a great diagram that illustrates this family. Thanks Simran, very helpful!

A one liner summary of xVelocity?
This is SQL Server going fast – sometimes really really fast (think 100X faster)! Continue reading