yugabyte/yugabyte-db

YugabyteDB is a distributed SQL database that combines PostgreSQL compatibility with horizontal scalability and cloud-native deployment capabilities.

YugabyteDB addresses the tension between transactional consistency and scalability by implementing a distributed architecture based on Google Spanner's design. It reuses PostgreSQL's query layer to support standard SQL features while adding distributed transaction support through Raft consensus replication and hybrid logical clocks. The system provides strong consistency by default for both reads and writes, with configurable read consistency to allow queries from followers and read replicas when eventual consistency is acceptable. Automatic failover and repair mechanisms enable the database to tolerate disk, rack, node, zone, region, and cloud failures without manual intervention.

YugabyteDB suits cloud-native OLTP applications that require both data correctness and at least one of scalability, high failure tolerance, or global distribution. Teams building mission-critical applications on public clouds or Kubernetes should consider it when they need relational database semantics without sacrificing the ability to add capacity by adding nodes. The PostgreSQL compatibility means existing applications using PostgreSQL drivers and SQL patterns can migrate with minimal changes. It is particularly valuable for deployments spanning multiple zones, regions, or clouds where traditional single-region databases create operational complexity or data residency challenges.

The project maintains active development across its core distributed transaction engine, consistency guarantees, and multi-region replication features. Work continues on expanding PostgreSQL compatibility and optimizing performance for various workload patterns. The codebase shows ongoing refinement of failure recovery mechanisms and deployment flexibility across different infrastructure targets.