A History of Databases

E. F. Codd of IBM's San Jose Research Lab published 'A Relational Model of Data for Large Shared Data Banks' in CACM. Grounded in set theory and first-order predicate logic, the paper proposed the relational algebra and established 'data independence'—the principle that applications need not depend on the physical layout of stored data. As a fundamental challenge to the then-dominant hierarchical and network DBMS (IMS, IDS), it became the theoretical origin of every RDBMS that followed, from Oracle and DB2 through SQL Server, MySQL, and PostgreSQL. Codd received the ACM Turing Award in 1981.

Read more →

Software Development Laboratories—founded by Larry Ellison, Bob Miner, and Ed Oates in 1977, later renamed Relational Software, Inc. and then Oracle Corporation—shipped Oracle V2 in June 1979. It became the first commercial SQL-based relational DBMS on the market. The company that turned Codd's 1970 relational model paper into a product was not IBM, but this small startup running on a PDP-11. With the US CIA as an early customer, Oracle grew into a software giant: by 2024 its revenue exceeded US$50 billion and Ellison was a perennial member of the global top-ten wealthiest.

Read more →

Michael 'Monty' Widenius, David Axmark, and Allan Larsson founded MySQL AB in Sweden and released the first version of MySQL—a GPL/commercial dual-licensed RDBMS. As the 'M' in the LAMP stack (Linux, Apache, MySQL, PHP/Perl/Python), it powered the Web boom of the 2000s. In 2008 Sun Microsystems acquired MySQL AB for US$1 billion; with Oracle's acquisition of Sun in 2010, MySQL fell under Oracle's roof. Monty forked MariaDB in 2009 to keep an open successor line alive.

Read more →

PostgreSQL 6.0 was released as the successor to the UC Berkeley Postgres project (1986-1994) led by Michael Stonebraker. Having adopted SQL and migrated to an external development community, it relaunched as a serious BSD-licensed RDBMS. With strict ACID compliance, MVCC concurrency, an extensible type system, and the rich ecosystem of add-ons such as PostGIS (geospatial) and pgvector (vector search), it built a deep following—and in Stack Overflow's Developer Survey it overtook MySQL in the 2020s to become the perennial 'most-loved database'.

Read more →

10gen (later MongoDB Inc.), founded by Dwight Merriman and Eliot Horowitz, released MongoDB 1.0—a document-oriented NoSQL database that stored BSON-formatted JSON documents instead of rows and columns, with schemaless modelling and horizontal scalability. It became the public face of the 'NoSQL movement'. MongoDB Inc. went public on NASDAQ in 2017, and by 2024 reported annual recurring revenue exceeding US$2 billion—establishing the document model as a durable choice in a DBMS market that had been dominated by relational systems.

Read more →

AWS announced Amazon Redshift at re:Invent 2012 (general availability followed in February 2013). Built on ParAccel's MPP technology, the column-oriented analytical data warehouse promised petabyte-scale queries at roughly a tenth of the cost of incumbent on-prem DWH like Teradata, Netezza, and Oracle Exadata. By overturning the assumption that a data warehouse meant buying dedicated hardware, Redshift—together with BigQuery (GA 2010)—established the 'cloud DWH' category, opening the path that Snowflake and Databricks would later disassemble and re-architect.

Read more →

Snowflake—founded in 2012 by Benoit Dageville, Thierry Cruanes, and Marcin Żukowski, all veterans of Oracle and Vectorwise—reached general availability in October 2014 after a June 2014 beta. By fully separating storage (object stores such as S3) from compute (virtual warehouses), it built a cloud-native DWH where compute could be spun up only as needed. Its 'multi-cloud, multi-cluster, shared-data' architecture differentiated it from Redshift and BigQuery and powered explosive growth: the September 2020 IPO closed day one at roughly a US$70 billion market cap—the largest software IPO ever at the time. By 2024, annual revenue exceeded US$3 billion.

Read more →

Through 2023 the PostgreSQL extension pgvector advanced from v0.4.x to v0.5.x and exploded in popularity as a vector database for RAG (Retrieval-Augmented Generation) in the LLM era. Embeddings—high-dimensional vector representations of text or images—are stored inside ordinary PostgreSQL tables, queried by cosine similarity, L2 distance, or inner product. As RAG demand surged after ChatGPT's November 2022 release, pgvector's 'add an extension to your existing PostgreSQL' approach went head to head with purpose-built vector databases such as Pinecone, Weaviate, and Chroma. The major managed PostgreSQL services—AWS RDS, Azure Database for PostgreSQL, and Supabase—rolled out pgvector support in quick succession.

Read more →