Looking Back at Postgres
Postgres pioneered extensible, object-relational database features that led to the world's most popular independent open-source database.
2019/01/07 by Joseph M. Hellerstein, Hellerstein, Joseph M. · 17 voices · 1 citation
#cs.DB
paper · pdf · doi:10.48550/arxiv.1901.01973
Abstract
This is a recollection of the UC Berkeley Postgres project, which was led by Mike Stonebraker from the mid-1980's to the mid-1990's. The article was solicited for Stonebraker's Turing Award book, as one of many personal/historical recollections. As a result it focuses on Stonebraker's design ideas and leadership. But Stonebraker was never a coder, and he stayed out of the way of his development team. The Postgres codebase was the work of a team of brilliant students and the occasional university "staff programmers" who had little more experience (and only slightly more compensation) than the students. I was lucky to join that team as a student during the latter years of the project. I got helpful input on this writeup from some of the more senior students on the project, but any errors or omissions are mine. If you spot any such, please contact me and I will try to fix them.
Summary
The paper reviews the history and technical architecture of Postgres, highlighting its focus on extensibility and object-relational features. It argues that the system's success stems from a sound architectural core that allowed it to survive as a widely adopted open-source project and a foundation for numerous commercial products.
machine-generated · gemma4:31b
In simple words
A long time ago, a man named Michael Stonebraker made a tool for storing data. He wanted it to make many different kinds of things work well together. This tool became very popular because people could add new parts to it easily. Many companies used his ideas to make money and billions of dollars were spent to buy those companies.
machine-generated · gemma4:31b
Outline
- Opening — Introduces Postgres as a successful 'second system' with massive industry impact.
- Postgres: An Overview — Details the technical goals and features, including ADTs, active databases, and storage recovery.
- Software Impact — Discusses the transition to PostgreSQL open-source and the commercial companies built upon it.
- Lessons — Reflects on extensibility, general-purpose architecture, and the value of open-sourcing research.
machine-generated · gemma4:31b
Argument
- Postgres was designed as an extensible 'object-relational' system to handle complex data types beyond simple rows and columns.
Design papers from 1986 and 1991, and the implementation of ADTs and R-trees. - This architectural focus on extensibility allowed Postgres to incorporate new features (like JSON/XML) without requiring a full re-architecture.
The evolution of PostgreSQL's support for nested data types over time. - The system provided a foundation that was robust enough for both an open-source community to maintain and commercial vendors to fork.
The existence of PostgreSQL as a top-ranked DBMS and the list of acquisitions (Illustra, Netezza, etc.) totaling over 2.6 billion. - Postgres's success proves that a broad, general-purpose architecture can be more sustainable than highly specialized systems.
Comparison between 'Berkeley Stonebraker' (general purpose) and 'MIT Stonebraker' (specialized).
machine-generated · gemma4:31b
Assumptions
- The transaction dollar value of acquisitions is a reliable proxy for the commercial success and market value of Postgres-based technology. [unstated]
- The current popularity of PostgreSQL (via DB-Engines rankings) indicates the technical superiority or utility of its architectural bones. [unstated]
machine-generated · gemma4:31b
Claims
- Postgres is an example of a 'second system' that succeeded despite being overburdened with features. [argument]
- The open-source PostgreSQL system is the most popular independent open-source database system in the world. [citation]
- Companies built from a Postgres base have generated over 2.6 billion in acquisitions. [citation]
- Postgres's approach to extensible access methods (like R-trees) influenced modern GIS systems like PostGIS. [argument]
machine-generated · gemma4:31b
Cited by
Discussions
- Looking Back at Postgres [hn, 290 points, 74 comments]
- Looking back at Postgres (2019) [hn, 32 points, 1 comments]
- Looking back at Postgres [hn, 19 points, 1 comments]
- This is a great read, a retrospective on how #postgresql started and how it became the world's most popular database 👇 Truly a great piece of technology 👏 arxiv.org/pdf/1901.01973 [bsky, 14 points, 0 comments]
- Looking Back at Postgres (2019) [hn, 12 points, 0 comments]
- Looking Back at Postgres [lobsters, 9 points, 1 comments]
- Looking Back at Postgres [hn, 4 points, 0 comments]
- History of Postgres, starting from early days [hn, 2 points, 0 comments]
- Looking Back at Postgres [pdf] [hn, 2 points, 0 comments]
- Looking back at the Berkeley Postgres project [hn, 2 points, 0 comments]
- Looking Back at Postgres [hn, 2 points, 0 comments]
- A view into early years of Postgres [hn, 2 points, 0 comments]
- Looking Back at Postgres [hn, 2 points, 0 comments]
- Looking Back at Postgres (2019) [hn, 1 points, 0 comments]
- Looking Back at Postgres [hn, 1 points, 0 comments]
- arxiv.org/abs/1901.01973 [bsky, 1 points, 0 comments]
- Looking Back at Postgres [hn, 1 points, 0 comments]
Related