I’m sitting at the bloggers table at the Day 1 keynote of PASS Summit 2015 and I try to recap the important news for the BI community.

  • SQL Server Analysis Services 2016 CTP 3 will be available in a few days here, but Microsoft already announced the news in this release:
    • Previously announced news (most of them already described in this blog and/or available in Power BI Desktop):
      • Extended Events for SSAS available in SSMS
      • Parallel partition processing
      • New DAX functions
      • DirectQuery over SQL Server 2014
      • New DAX functions and syntax (including variables)
    • New SQL Server 2016 compatibility level: it’s no longer based on XMLA (that was created for Multidimenisonal) and we have a new Tabular Object Model (TOM). No problem for existing models if they keep the 1103 compatibility level, but once you migrate to the new model, you have to use the new API (but you also have access to the new features). API for TOM is available in JSON, just as in Power BI APIs.
    • JSON-based tabular scripting language: while there is still a single file for the entire model, the code is more readable and (most important) it’s easier to be managed in a version control system, possibly merging changes applied by different developers to the same data model
    • Bi-directional cross filtering: just as in Power BI Desktop, with an equivalent diagram view layout
    • MDX Support for direct query: this is a *BIG NEWS*. Implication is that you will be able to use Excel pivot tables (generating MDX queries) with Direct Query. But wait, there is more about DirectQuery…
    • Oracle, Teradata, and APS supported in DirectQuery: while I’m still confident that an in-memory database is very useful when you have a certain amount of data, of for certain calculations, the ability to create real-time dashboards connecting your relational database to a thin semantic layer is wonderful. Just be careful, you have to use compatibility level 1103 for this feature in CTP3 (this will certainly change before RTM).
    • SSDT for Visual Studio 2015 available: this is just necessary and expected. The good news is that you don’t need to download 1GB of file if you already have VS2015. Important news here are better performance of SSDT when you design a data model (but I want to test it… again, it’s easy to improve what we have today) and finally we’ll be able to use Visual Studio Configuration manager to create test and production deployment configurations. Really really needed!
    • Improved DAX formula editing: now, it’s easy to do something better than what we had in the past! I’m really demanding here and I hope there will be a way to integrate DAX Formatter (I will investigate on that this week). By the way, the new development environment has formula fixup – you rename a measure and an automatic refactor happens, just as in Excel 2016 and Power BI Desktop.
    • DBCC for Analysis Services: you can check the consistency of your tabular and multidimensional models.

The keynote was more about SQL Server 2016 as a platform, on premises and on cloud. SQL Server is now better than Oracle in Gartner quadrant, in TPC-H benchmark (for data warehouse) and for security (it is the database with the lower number of vulnerabilities in the last 5 years). This would have been a joke 10 years ago. There are also many improvements such as integration with R, many security, scalability, reliability and performance improvements (in-memory technology in SQL Server is always improving).

You can watch PASS keynote on PASStv: http://www.sqlpass.org/summit/2015/PASStv/Keynote.aspx

That’s all for today’s keynote. More news later this week watching sessions, I hope!