PostgreSQL is an open-source object-relational database descended from the POSTGRES project started by Michael Stonebraker at UC Berkeley in 1986. It implements a large subset of the SQL standard, enforces ACID guarantees, and uses multi-version concurrency control (MVCC) so that readers never block writers and writers never block readers.
Its architecture is a process-per-connection model: a postmaster supervisor forks a backend for each client, backends coordinate through shared buffers and WAL (write-ahead log), and background workers handle autovacuum, checkpointing, and logical/streaming replication. The WAL is also the foundation for physical replication, point-in-time recovery, and logical decoding that feeds tools like Debezium. The query planner is a cost-based optimizer with support for parallel sequential scans, parallel hash joins, JIT compilation via LLVM, and partitioned tables.
What sets Postgres apart from most databases is its extensibility: user-defined types, operators, index access methods (B-tree, Hash, GiST, SP-GiST, GIN, BRIN), procedural languages (PL/pgSQL, PL/Python, PL/Perl), and loadable C extensions. That surface enables PostGIS for geospatial, TimescaleDB for time-series, pgvector for embeddings, Citus for sharding, and hundreds more. PostgreSQL backs a large portion of the hosted database market (AWS RDS, Aurora, Google AlloyDB, Azure Database for PostgreSQL, Supabase, Neon, Crunchy Data) and is the default SQL choice for most new application workloads.