Contents

Click a page name

Cost-based Optimization of Queries Across Diverse Data Sources
Motivation and Goals
Assumptions
The Problem
The Key Ideas
Overview of Optimization
Building Plans using Properties
Costing in a Traditional Optimizer
Costing in a Heterogeneous Environment
An Update Statistics Facility
The Framework
Advantages of Our Approach
Experiments
Results
Wrapper Input Matters
It's Not Difficult
Default Model, Simple Source
Default Model, Source w/Preds
Modeling QBIC
Good Decisions Are Made
Related Work
Conclusion
A Default Cost Model for Wrappers
Default Cost Equations for Wrappers
Statistics Supported