Skip to main content

tikv

TiKV is a distributed transactional key-value database (distributed data store) designed for horizontal scalability, high availability, and ACID-compliant workloads in cloud-native environments.

  • Distributed transactional key-value store with ACID semantics (database / data store)
  • Horizontally scalable storage with automatic sharding and rebalancing (distributed systems)
  • Replication-based fault tolerance using Raft consensus (data resiliency / consensus)
  • Support for both raw key-value and transactional APIs (application data services)
  • Integration as the storage layer for TiDB and other Structured Query Language (SQL) layers (database ecosystem)

More About tikv

TiKV is an open-source distributed transactional key-value database (database / data store) originally developed to provide a storage engine for TiDB and now used as a general-purpose distributed storage component. It targets workloads that require ACID transactions, linearizability, and horizontal scaling on commodity hardware, positioning it in the category of distributed operational data stores for cloud-native and containerized environments.

The project implements a distributed architecture where data is automatically partitioned into regions and distributed across nodes, enabling horizontal scalability (distributed systems). It uses the Raft consensus algorithm (consensus / data resiliency) for replication and leader election, providing tolerance to node failures while maintaining strong consistency. TiKV exposes both a transactional Application Programming Interface (API) that supports distributed transactions with snapshot isolation and a raw key-value API (application data services), giving platform teams flexibility in how they model data and build higher-level services or integrate SQL engines.

Internally, TiKV leverages a log-structured key-value engine such as RocksDB for local storage (storage engine) and layers distributed transaction protocols, multi-version concurrency control (MVCC), and snapshot mechanisms on top (database internals). These capabilities enable transactional semantics across multiple keys and regions. The system is designed for deployment on commodity servers or cloud infrastructure, often orchestrated by Kubernetes (cloud-native infrastructure) when used in containerized environments.

In enterprise environments, TiKV is commonly deployed as the storage layer under TiDB (SQL database), where TiDB provides a MySQL-compatible SQL interface and TiKV stores table data and indexes as key-value pairs. This separation allows independent scaling of stateless SQL processing nodes and stateful storage nodes (scale-out architecture). Organizations can also access TiKV directly through its client libraries and APIs to build services that require low-level transactional key-value access without a SQL layer.

TiKV is a graduated project under the Cloud Native Computing Foundation (CNCF) (open-source foundation), which places it within the cloud-native ecosystem alongside other CNCF projects. It integrates with observability tooling through metrics and logging interfaces (observability) and can be managed using standard container orchestration practices. For an enterprise directory, TiKV fits under categories such as distributed transactional key-value database, cloud-native storage platform, and Raft-based strongly consistent data store.