Dimensional Modeling Fundamentals
A well designed dimensional model provides the best performing, most flexible, and most easily understood data architecture for the DW/BI system. The dimensional model underlies every successful Microsoft-based DW/BI project. This is true both for the relational platform, leveraging the Star Join optimization in SQL Server 2008 database engine, and for Analysis Services which was fundamentally re-architected in 2005 to support true dimensional models. This session covers the key concepts in dimensional modeling, including facts, dimensions, change-tracking with slowly changing dimensions, junk dimensions, mini dimensions, and bridge tables. We look at examples of each concept, and of several dimensional models for different business functions and industries. We’ll also explore the primary differences between 3rd normal form models and dimensional models.
Warren Thornthwaite has been building decision support and DW/BI systems since 1980. He worked at Metaphor for 8 years, managing the consulting group and implementing many major data warehouse systems. After Metaphor, Warren was the DW/BI program director at Stanford University, and he co-founded InfoDynamics LLC, a DW/BI consulting firm. Warren moved to Microsoft/WebTV to help build a world class, multi-terabyte DW/BI system in 1998. He returned to consulting with the Kimball Group in 2002. Warren is co-author of the best-selling Data Warehouse Lifecycle Toolkit and the Microsoft Data Warehouse Toolkit books. He also writes articles for Intelligent Enterprise and SQL Magazine. He holds an MBA in Decision Sciences from the Wharton School, and a BA from the University of Michigan.
Register Now for PASS Summit 2010!