Most SQL databases make you choose: either stick with a single machine and hit scaling limits, or shard manually and lose transaction guarantees. TiDB eliminates this choice by automatically distributing your data across multiple nodes while preserving full ACID compliance and MySQL compatibility. Your existing applications can connect without code changes, but suddenly handle petabytes of data.
What sets TiDB apart is its hybrid HTAP architecture - it runs both transactional (row-based) and analytical (columnar) workloads on the same data simultaneously. The built-in Raft consensus ensures your cluster survives node failures automatically, while the separated compute and storage layers let you scale each independently. At 39K+ stars, it’s proven in production at companies processing billions of transactions daily.
If you’re hitting database scaling walls or tired of managing sharded MySQL clusters, TiDB offers a compelling path forward. The project provides comprehensive documentation, active community support, and you can spin up a local cluster in minutes to test with your existing schema.
⭐ Stars: 39711
💻 Language: Go
🔗 Repository: pingcap/tidb