Lecture
This lecture covers the optimization and execution of database queries, comparing static optimization with dynamic re-optimization. It delves into the granularity of optimization, discussing single query versus multiple queries, and the benefits of scan sharing. The instructor explains how running multiple OLAP queries concurrently can saturate I/O bandwidth and proposes solutions for sharing CPU and I/O. Examples of disk-based scan sharing are provided, showcasing how queries can share cursors to reduce I/O. The lecture concludes with the importance of query optimization in DBMS performance, scalability, and the challenges faced in cost estimation and space exploration for efficient query plans.