TiDB
TiDB is a distributed Structured Query Language (SQL) database that combines the relational semantics of a MySQL-compatible (relational database) with horizontal scalability and strong consistency (distributed database) for transactional and analytical workloads.
- Horizontally scalable, distributed SQL database for OLTP and OLAP workloads (database platform)
- MySQL-compatible syntax and ecosystem integration for applications and tooling (relational database)
- Row-store transactional layer with ACID transactions and strong consistency (transaction processing)
- Columnar analytics engine for real-time and batch analytical queries (analytics engine)
- Automatic sharding, fault tolerance, and online scaling across commodity hardware or cloud environments (infrastructure data layer)
More About TiDB
TiDB is an open-source distributed SQL database developed by PingCAP and designed for scenarios where traditional single-node relational databases encounter scaling and availability constraints. It targets online transaction processing (OLTP), hybrid transactional/analytical processing (HTAP), and mixed workloads that require MySQL compatibility with horizontal scalability and strong consistency across nodes.
Architecturally, TiDB separates compute and storage into distinct layers (distributed database architecture). The TiDB server layer acts as a stateless SQL engine that is MySQL protocol compatible, allowing existing applications, drivers, and tools in the MySQL ecosystem to connect with minimal change. The underlying distributed transactional key-value store, TiKV, manages data distribution, replication, and storage, and works with the Placement Driver (PD) component, which handles metadata management, scheduling, and cluster resource coordination.
TiDB provides a row-based transactional engine (transaction processing) with ACID guarantees and strong consistency through a distributed transaction model. It supports standard SQL, secondary indexes, and other relational database capabilities for OLTP workloads. For analytical processing, TiDB integrates TiFlash, a distributed columnar storage engine (analytics engine) that maintains replicated columnar copies of row-based data. This enables HTAP use cases by allowing analytical queries and reports to run against columnar replicas while transactional workloads continue on the row store.
For operations teams, TiDB offers online scaling and elasticity (infrastructure operations), allowing the cluster to add or remove nodes without application downtime. Data is automatically sharded and balanced through regions and Raft-based replication groups, with automatic failover and data redundancy for high availability (resilience). Management and monitoring are supported through tooling such as TiUP and dashboards for deployment, configuration, and observability, as documented by PingCAP.
Enterprises use TiDB in environments where they need a relational database with MySQL compatibility but must run on distributed infrastructure across data centers or cloud platforms. Common usage patterns include mission-critical OLTP systems, real-time analytics on fresh transactional data, and consolidation of multiple database instances into a single distributed cluster. TiDB fits into enterprise data platforms as a primary System of Record (SOR) database, an HTAP engine, or a scalable backend for Software-as-a-Service (SaaS) and internet applications that require consistent transactions and large-scale concurrency.
From a directory and taxonomy perspective, TiDB aligns with categories such as distributed SQL database, NewSQL database, HTAP engine, and MySQL-compatible relational database platform. It intersects infrastructure, data management, and analytics domains by providing a single system that addresses both transactional and analytical processing requirements on distributed infrastructure.