PDW AU3 (Chrysalis)
In recent years, massively parallel processors have increasingly been used to manage and query vast amounts of data. Dramatic performance improvements are achieved through distributed execution of queries across many nodes. Query optimization for such systems is a challenging and important problem. One of our major contributions to SQL Server PDW AU3 was the cost-based query optimizer for the SQL Server Parallel Data Warehouse product that was initially developed at the GSL. By properly abstracting metadata we were able to reuse existing logic in SQL Server 2008/2012 for query simplification, plan space exploration and cardinality estimation. Unlike earlier approaches to parallel query optimization that simply parallelized the best serial plan, our optimizer considers a rich space of execution alternatives, and picks one based on a cost model for the distributed execution environment. The result is a high-quality, effective query optimizer for distributed query processing.
- David DeWitt
- Alan Halverson
- Eric Robinson
- Srinath Shankar
- Rimma Nehme
- Srinath Shankar, Rimma V. Nehme, Josep Aguilar-Saborit, Andrew Chung, Mostafa Elhemali, Alan Halverson, Eric Robinson, Mahadevan Sankara Subramanian, David J. DeWitt, César A. Galindo-Legaria: Query Optimization in Microsoft SQL Server PDW. SIGMOD Conference 2012: 767-776.
- Rimma V. Nehme, Nicolas Bruno: Automated partitioning design in parallel database systems. SIGMOD Conference 2011: 1137-1148.
- Query Optimization in Microsoft SQL Server PDW. SIGMOD Conference 2012, Scottsdale, AZ May 2012.