LEO - DB2's LEarning Optimizer

By Michael Stillger, G. Lohman, Volker Markl, M. Kandil. Very Large Data Bases Conference, 2001

436 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/24cad9705ccc13fce969021a708ca703db3ae31c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Bao: Making Learned Query Optimization Practical (2020, methodology): “Several works have applied machine learning techniques to these stubborn problems [27, 29, 33, 39, 41, 47, 60, 61, 64].”
Self-Tuning Database Systems: A Decade of Progress (2007, background): “The Learning Optimizer (LEO) project [62], part of IBM’s autonomic computing initiative, identifies incorrect cardinality estimates and saves the execution feedback for future optimization.”
The History of Histograms (abridged) (2003, methodology): “The LEO system [70] generalizes these e(cid:11)orts as it uses result sizes of much more complicated queries to modify its statistics, including join and aggregate queries, queries with user-de(cid:12)ned functions, and others.”
CORDS: automatic discovery of correlations and soft functional dependencies (2004, background): “leo, the DB2 learning optimizer [20], is a typical example of a qfs.”
How Good Are Query Optimizers, Really? (2015, background): “Some commercial systems support the use of sampling for base table estimation, multi-attribute histograms (“column group statistics”), or ex post feedback from previous query runs [38].”
A Survey on Big Data Market: Pricing, Trading and Protection (2018, methodology): “For optimizing such queries, a big data learning scheme was proposed by [108].”
Neo: A Learned Query Optimizer (2019, background): “IBM DB2’s LEO optimizer was the first to introduce the idea of a query optimizer that learns from its mistakes [54].”
Exploiting Dynamic Resource Allocation for Efficient Parallel Data Processing in the Cloud (2011, unknown): “, from collected statistics, similarly to the way database systems try to optimize their execution schedule over time [24].”
Predicting Multiple Metrics for Queries: Better Decisions Enabled by Machine Learning (2009, methodology): “• They attempt to predict only a single performance metric, such as the elapsed time or actual cardinality of the underlying data [9], [10], [11], [1], [12], [13], [14].”