Who TiKV is for#
Infrastructure teams building distributed databases
Use TiKV when you need a consistent key-value storage layer and have the expertise to operate distributed systems.
Skip if:
Skip if you need a turnkey SQL database for application developers.
Teams evaluating TiDB storage internals
TiKV helps platform engineers understand and tune the storage layer behind TiDB workloads.
Skip if:
Skip if your database problem can be solved with a managed relational service.
The problem it solves#
Single-node key-value stores are simple, but they become risky when data must survive node failures, scale horizontally, and keep strong consistency. Teams can shard manually or depend on a managed cloud database, but both choices add operational or vendor constraints.
Distributed transactional storage is hard because replication, consensus, rebalancing, and failure recovery must work together. Teams need a storage engine that already solves those distributed systems problems.
How it solves it#
Distributed key-value storage
Splits data across nodes and regions so storage can scale horizontally instead of depending on one server.
Raft-based replication
Uses Raft consensus for replicated data and strong consistency across failures.
Transactional API foundation
Provides a transactional key-value layer used by TiDB and useful for systems that need consistent distributed storage.
Strengths and trade-offs#
Strengths
- Battle-tested inside TiDBTiKV is not only a research project; it serves as the storage layer for TiDB deployments.
- Apache-2.0 Rust codebaseThe permissive license and Rust implementation make TiKV attractive for infrastructure teams evaluating distributed storage internals.
Trade-offs
- -Not an application database by itselfMost product teams should use TiDB or another higher-level database. TiKV is an infrastructure component that requires distributed systems skill to operate directly.
TiKV vs alternatives#
TiKV vs DynamoDB
TiKV and DynamoDB both address scalable key-value storage, but they target different operators. TiKV is an open source distributed storage engine; DynamoDB is a fully managed AWS database product.
TiKV is better for infrastructure teams that need control over a strongly consistent storage layer. DynamoDB is still better when the team wants AWS to own operations, scaling, and availability.
What it's built on#
- Languages
- PerlPythonRust
- Infrastructure
- AWS
FAQ#
Is TiKV open source?
Yes. TiKV is open source under the Apache-2.0 license.
What is TiKV used for?
TiKV is used as a distributed transactional key-value store and as the storage layer for TiDB.
How does TiKV compare to DynamoDB?
DynamoDB is a managed AWS key-value and document database. TiKV is an open source distributed storage engine that teams operate themselves or use through TiDB.
Similar open-source tools#
Apache CouchDB
NoSQL database with multi-primary sync and HTTP API
MongoDB
Popular open source NoSQL document database for modern apps
OrientDB
Multi-model NoSQL database: graphs, docs, and key-value
JanusGraph
Distributed open source graph database for massive scale
ArangoDB
Multi-model database: documents, graphs, and key-value in one
RethinkDB
Open source distributed document database with real-time feeds