TiDB

Distributed SQL database that runs both transactional and analytical workloads on the same data without ETL. Built around a MySQL-compatible SQL layer, it delivers horizontal scaling through automatic sharding while keeping strong ACID guarantees across a decoupled compute-and-storage architecture.

TiDB is fully open source and widely used for high-concurrency, petabyte-scale systems. It differentiates from sharded MySQL by providing a single logical database with automatic distribution, plus a columnar engine for real-time analytics on live transactional data.

Key capabilities:

• MySQL protocol and syntax compatibility for straightforward migration
• HTAP support with row store for OLTP and column store for OLAP on the same cluster
• Decoupled compute and storage with cloud-native object storage
• Horizontal scaling via automatic sharding and rebalancing
• Kubernetes deployment with TiDB Operator, or self-managed via TiUP
• Migration tooling including TiDB Lightning, Dumpling, and TiCDC for sync

Common uses include fintech and payment systems needing consistency at scale, multi-tenant SaaS with high table and connection counts, operational analytics on transactional data, and RAG-style AI workloads that combine structured data with vectors and full-text search.