Relational Databases (RDBMS)
Tables, rows, and rigorous schemas. The boring, proven technology that powers 90% of the world's functional software.
The Default Choice
If you are starting a new project and someone suggests anything other than PostgreSQL or MySQL as the primary data store, make them defend that choice rigorously. Relational databases enforce data integrity at the lowest level, preventing application bugs from permanently corrupting your state.
When to use RDBMS:
- Strong ACID guarantees are non-negotiable (financial transactions).
- Data structure is relatively stable and well-understood.
- You need complex querying involving multiple entities (Joins).
The Scaling Myth
The primary argument against RDBMS is "they don't scale horizontally." While true that active-active write scaling is hard, vertical scaling (buying a bigger server) gets you incredibly far. A modern dedicated server with NVMe SSDs and hundreds of GBs of RAM can handle tens of thousands of transactions per second on Postgres.
Major Players
PostgreSQL
The StandardObject-relational, highly extensible, incredible JSON support. It has practically eaten the "NewSQL" market by simply adding features.
Compare with MySQLMySQL / MariaDB
Historically faster for simple reads, easier replication setup (historically). Powers WordPress and most of the old web.