Reengineering an object-oriented framework for extensible query optimization

In this thesis, we describe a third-generation extensible query optimization framework that has evolved from the OPT++ framework of Navin Kabra for relational databases. Our framework does not change the infrastructure of the OPT++ architecture, which consists of three components: a Search Strategy component, a Search Space component and an Algebra component. However, we address the problems encountered while building two query optimizers in the framework: a simple bottom-up optimizer and an instance of the PostgreSQL query optimizer and enhance the modularity and the collaboration of the three components of the framework at detailed level, which in turn leads to a more flexible, easier to extend and cleaner implementation. The framework has been validated by implementing the two query optimizers. While both these cases were for the relational data model, we believe the design does cover the optimization process for other data models