Skip to content

A prototype implementation of Bao for PostgreSQL

License

AGPL-3.0, AGPL-3.0 licenses found

Licenses found

AGPL-3.0
LICENSE
AGPL-3.0
COPYING
Notifications You must be signed in to change notification settings

frances-h/BaoForPostgreSQL

 
 

Repository files navigation

Bao loves PostgreSQL

This is a prototype implementation of Bao for PostgreSQL. Bao is a learned query optimizer that learns to "steer" the PostgreSQL optimizer by issuing coarse-grained query hints. For more information about Bao, check out the paper.

Documentation, including a tutorial, is available here: https://rmarcus.info/bao_docs/

While this repository contains working prototype implementations of many of the pieces required to build a production-ready learned query optimizer, this code itself should not be used in production in its current form. Notable limitations include:

  • The reward function is currently restricted to being a user-supplied value or the query latency in wall time. Thus, results may be inconsistent with high degrees of parallelism.
  • The Bao server component does not perform any level of authentication or encryption. Do not run it on a machine directly accessible from an untrusted network.
  • The code has not been audited for security issues. Since the PostgreSQL integration is written using the C hooks system, there are almost certainly issues.

This software is available under the AGPLv3 license.

About

A prototype implementation of Bao for PostgreSQL

Resources

License

AGPL-3.0, AGPL-3.0 licenses found

Licenses found

AGPL-3.0
LICENSE
AGPL-3.0
COPYING

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C 59.5%
  • Python 29.6%
  • TSQL 5.6%
  • Jupyter Notebook 4.9%
  • Other 0.4%