There is a lot of business intelligence news coming out of the Microsoft camp.  First, sooner than expected, they have released the SQL Server Fast Track Data Warehouse, which is the development of the Datallego acquisition and Microsoft's entry into the appliance marketplace.

Some time ago, they also announced Kilimanjaro, the next version of SQL Server, and Gemini, the next version of SQL Server Analysis Services.  I didn't rush out and blog this though because the release is not expected until 2010.  Still, it's exciting and worth thinking about now.  And who knows, maybe these will emerge sooner than expected.

I thought SQL Server 2005 was BI focused.  Now, Microsoft says Kilimanjaro will really be BI focused.  And they could be right.  With the addition of optional column-oriented storage capabilities and in-memory storage, they are addressing 2 very obvious ways the market has recently addressed performance.  Both options were built from scratch by Microsoft.  I look forward to seeing how the optimizer determines when to go columnar vs. not.  To-date, solutions have addressed it with disparate systems and obvious suboptimal performance of some of the queries within each system. 

I'm a fan of SSAS as long as the MOLAP option is not overused, but that's another story.  Stronger connectivity is promised between Gemini and Excel and that may really get the casual user's attention.  Gemini will be publishing data to SharePoint Server as well.  SharePoint is becoming a centerpiece of Microsoft BI activity.  With the publishing options,Gemini will have the Microsoft enterprise covered.

What will be the next game changer?

Pivotlink is a pure SaaS play for midmarket or enterprise departmental needs.  They make a clear claim to target companies with BI shelfware and have a number of case studies where customer's BI needs were met with PivotLink after failures with other solutions. 

Pivotlink delivers data in various forms, but mostly based around the concept of a pivot table, providing ad-hoc data access across a high level of dimensionality.   

Interestingly, I have learned that PivotLink uses a proprietary columnar in-memory database management system.  This DBMS is not sold separately.  As with any columnar, there is a high compression ratio.

And here you can see how the Pivotlink data model is customizable through an abstracted view.

From a technology perspective, Pivotlink is interesting from its combination of in vogue approaches like columnar DBMS and SaaS.  It claims access to a variety of source systems, supporting customers with billions of rows of data like REI.