Abstract

Abstract: Ever wanted to learn how to build a query optimizer? This talk delivers a primer on the core academic concepts behind Microsoft SQL Server’s query optimizer. This covers topics including top down vs. bottom-up, elements of the CASCADES framework such as properties, rules, and the memo plan fragment storage mechanism, and an overview of the kinds of optimizations that are implemented within the framework as used in SQL Server. This talk is targeted towards people who are interested in how query optimizers work internally and how they should be designed.

Biography: Conor is Principal Software Architect on the SQL Server Query Processor team at Microsoft. He has worked primarily on query processors and optimizers for 10+ years. Conor has worked on a number of areas including the search framework, index matching, distributed query, update processing, language design and semantics, paging and navigation, hierarchical queries, data warehousing, object-relational databases, and a number of other areas. He holds numerous patents in these areas and is the author of several academic papers. Recently, he authored a chapter for the book “SQL Server 2008 Internals” on how the SQL Server Query Optimizer works. Some of his current interest include self-learning/self-tuning database systems, hintable query languages, and parameter-sensitivity in finding optimal query plans.