Case Study: Baidu achieves scalable SQL at 40TB and 100k QPS with Cockroach Labs

Baidu needed SQL at massive scale

Baidu, a $90B+ internet company serving nearly a billion users across search, shopping and cloud services, needed a relational database that could support massive scale—40TB of data and 100k QPS—without the complexity of sharded MySQL and middleware. The DBA team sought a SQL-compatible solution that could handle continuous high-concurrency inserts, secondary indexes, and basic real-time analytics while improving operational reliability.

Cockroach Labs’ CockroachDB was selected after a deep evaluation and contributor collaboration; Baidu deployed it on bare-metal servers across 20 clusters (supporting two new applications) with a load balancer in front. CockroachDB’s horizontal scaling, automatic routing/rebalancing/replication, and automated failover reduced operator complexity, improved recovery time objectives, and fully utilized provisioned hardware—enabling Baidu to replace sharded MySQL for these workloads and streamlining developer workflows while continuing to contribute to and promote the product with Cockroach Labs.

Baidu