Robust Query Optimization Methods With Respect to Estimation Errors: A Survey (Rapport de recherche IRIT)
Résumé
The quality of a query execution plan chosen by a Cost-Based Optimizer (CBO) depends greatly on the estimation accuracy of input parameter values. Many research results have been produced on improving the estimation accuracy, but they do not work for every situation. Therefore, "robust query optimization" was introduced, in an effort to minimize the sub-optimality risk by accepting the fact that estimates could be inaccurate. In this survey, we aim to provide an overview of robust query optimization methods by classifying them into different categories, explaining the essential ideas, listing their advantages and limitations, and comparing them with multiple criteria.