CockroachDB
By Cockroach Labs
CockroachDB is an open-source, distributed SQL database designed to survive machine, rack, and datacenter failures while providing strong ACID consistency and horizontal scalability across many nodes.
Definition
CockroachDB is an open-source, distributed SQL database designed to survive machine, rack, and datacenter failures while providing strong ACID consistency and horizontal scalability across many nodes.
Overview
Traditional relational databases like PostgreSQL typically run on a single primary node, which makes strong consistency straightforward but creates a scaling and availability ceiling. CockroachDB was designed from the ground up to distribute data automatically across many nodes — potentially spread across regions or cloud providers — while still offering full ACID transactions and a familiar SQL interface, largely wire-compatible with PostgreSQL. Under the hood, CockroachDB splits tables into small ranges of data that are replicated across multiple nodes using the Raft consensus protocol, so if any single node or even an entire datacenter goes offline, the cluster keeps serving reads and writes without data loss. This architecture draws heavily on ideas from Google's internal Spanner database, adapted into an open-source system that any team can run on their own infrastructure or consume as a managed cloud service. Because it behaves like a normal SQL database to application code while handling replication, failover, and rebalancing automatically, CockroachDB is often chosen for applications that need both strong consistency and geographic distribution — a niche that sits between traditional single-node relational databases and eventually-consistent NoSQL systems like MongoDB. It's frequently discussed alongside other distributed SQL systems such as YugabyteDB and TiDB.
Key Features
- Distributed SQL with full ACID transaction guarantees
- Automatic data replication and rebalancing across nodes using Raft consensus
- Survives node, rack, or datacenter failures without data loss
- PostgreSQL wire-protocol compatibility for easier adoption
- Horizontal scalability by simply adding more nodes
- Multi-region deployment support for geo-distributed applications
- Automatic range splitting and load balancing across the cluster