BI Key note - A lecture from Tech Ed

BI Key note

Lecturers: Ori Weinroth and Frank Sidi

What's new in SQL Server 2011 BI Services (Code name "Danali")

Integration Services:

• Better developer experience - simply adding and removing columns.
• Manageability - easier control (didn't explain what exactly)
• Zoomx was bought by Microsoft - Integrated data cleansing tool.

Analysis Services

• In-memory mode using the Vertipack engine. Now BI Developers can use MOLAP, ROLAP or Vertipack to implement cubes. This will improve our competitions against Qlikview and similar products. You can think about is as real PowerPivot server (PowerPivot uses vertipack engine). We expect many migrations from MOLAP to Vertipack in the near future. More about the subject in a different post.

Reporting Services

• Code name "Crescent" is a new project for Silverlight self-service BI tool. This was one of the major problems comparing SSRS 2008 / 2008 R2 to the competitors.
• Teaser movie here: http://blogs.msdn.com/b/bi/archive/2010/11/09/data-visualization-done-right-project-crescent.aspx

Relational Engine

• Juneau - SQL Server IDE. Not sure I fully understand it. I'll ask for more info.
• Always On - next generation HA / DR. There are two sessions for this subject.
• Apollo - In memory query processing. New types of indexes (Column based).
• Sequences (Oracle like)
• Up to 256 Logical Processors

Some Interesting numbers - the largest DB / Single DB / table / Cube

• SQL Application of 88 PT
• One DB of 70 TB
• One table of 20 TB
• One cube of 12 TB

Project "Atlanta"

http://www.microsoftatlanta.com/

Configuration Assessment Could service for Microsoft SQL Server. It will check your configuration and alert you if something is wrong.

Parallel Data Warehouse

A lot was said and written about the solution(http://www.microsoft.com/sqlserver/2008/en/us/parallel-data-warehouse.aspx ).  At the demo Frank showed some reports. For ex. processing 1.2 B records in 26 seconds. It is a very impressive solution but as you know, designed for at least 50 TB of data. Else, probably it is overkill. Please don't forget this is a solution for Data Warehouse, NOT for  OLTP solutions.

Self Service BI

We saw already many demos of PowerPivot (http://www.powerpivot.com/) . Frank demoed how easy it is to join data from the local DB with data from SQL Azure (for ex. statistics about geographic regions my customers live in). Interesting, but not sure how relevant it is to most of my customers.

The second demo was of http://www.getpivot.com/