An adaptive approach for learning operator costs in Polypheny-DB (Bachelor Thesis, Finished)
Author
Description
A major limitation of the current query routing system of Polypheny-DB is the comparability of the calculated execution costs between different data stores. The goal of this project is to implement support in Polypheny-DB for learning and adjusting this cost model at runtime.
The project includes
- Introducing an additional metric for operator costs
- Adjusting the query routing to use this new cost metric
- Implementing a system for learning operator costs at runtime
- Evaluating the implementation
Start / End Dates
2022/03/28 - 2022/07/27