Databases & Storage
Relational Database OLTP Benchmark: PostgreSQL 16 vs MySQL 8.4
Head-to-head OLTP benchmarking of PostgreSQL 16 versus MySQL 8.4 under extreme read/write concurrency.
Test Environment
CPUAWS r6i.4xlarge
RAM128GB
OSUbuntu 22.04 LTS
Toolsysbench
Metrics Comparison
Write-Heavy TPS
Unit: Transactions/secPostgreSQL 16
18,500★
MySQL 8.4
15,200
Detailed Analysis
MySQL handles pure read-heavy point lookups slightly faster, while PostgreSQL dominates complex queries and heavy concurrent writes.
PostgreSQLs MVCC (Multi-Version Concurrency Control) handles high write concurrency better, avoiding lock contention seen in InnoDB under stress.
Row Locking
Acquiring locks for updates
Bottleneck: InnoDB page-level contention on hot rows.
Architectural Recommendations
- PostgreSQL for complex schemas and heavy writes
- MySQL for simple, read-heavy workloads
Frequently Asked Questions
Is JSON support better in Postgres?
Postgres JSONB is generally more mature and offers better indexing (GIN).