https://github.com/danolivo/pargres
[Prototype] [WIP] Extension on PostgreSQL for parallel query execution
https://github.com/danolivo/pargres
parallel postgresql sql
Last synced: 5 months ago
JSON representation
[Prototype] [WIP] Extension on PostgreSQL for parallel query execution
- Host: GitHub
- URL: https://github.com/danolivo/pargres
- Owner: danolivo
- License: other
- Created: 2018-10-01T09:54:38.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2018-11-09T06:16:45.000Z (over 7 years ago)
- Last Synced: 2025-10-14T19:36:26.530Z (5 months ago)
- Topics: parallel, postgresql, sql
- Language: C
- Homepage:
- Size: 137 KB
- Stars: 2
- Watchers: 1
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# ParGRES - [prototype] PostgreSQL extension for parallel query processing in shared-nothing architectures
## Introduction
This code devoted to demonstration of one approach [1] to parallel query execution in a shared-nothing architecture.
In accordace to the approach, parallel DBMS uses PostgreSQL core as a container for tuples. PDBMS manages metadata about data distribution. Input query is transfered to the nodes contained some part of tuples of distributed relations, involved into the query. Correctness of JOIN, Aggregate and other operations is provided by a parallel plan generator. It inserts custom `exchange` nodes into the plan positions, that needs tuples shuffling between nodes.
This code do not take into consideration such problems as `global snapshot` and `distributed commit`. Thereunder all transactions that need writing to distributed relations must be executed in sequental mode.
## Authors
Andrey Lepikhov a.lepikhov@postgrespro.ru, Postgres Professional, Moscow, Russia
## Installation
## Links
Lepikhov A.V., Sokolinsky L.B. Query Processing in a DBMS for Cluster Systems // Programming and Computer Software. 2010. Vol. 36. No. 4. P. 205-215. DOI: 10.1134/S0361768810040031