Open Source Key-Value Stores
Key-value stores organize data as pairs: a unique key identifies the value to retrieve, update, or delete. This simple model supports fast lookups and works well for caching, session data, application settings, and other workloads that do not need complex relational queries. Depending on their design, stores may keep data in memory, persist it to disk, or distribute it across multiple machines, with different tradeoffs in speed, durability, and scale.
Open source options include embedded storage engines, standalone in-memory services, and distributed transactional databases. When choosing one, consider data size and access patterns, persistence and consistency guarantees, transaction support, deployment requirements, language integrations, license, and maintenance activity. These tools are useful to developers and operators building applications that need efficient access to structured data by key.
3 repositories · updated January 12, 2026
KeyDB: A Multithreaded, High-Performance Fork of Redis
KeyDB is a high-performance, multithreaded fork of Redis, designed for enhanced memory efficiency and high throughput. It maintains full compatibility with the Redis protocol, modules, and scripts, making it a seamless drop-in replacement. KeyDB also introduces advanced features like Active Replication and a MVCC architecture for non-blocking operations.
FoundationDB: An Open Source Distributed Transactional Key-Value Store
FoundationDB is an open-source, distributed, transactional key-value store developed by Apple. It is designed to manage large volumes of structured data across clusters, offering strong ACID transaction guarantees for all operations. This robust database is well-suited for both read/write and write-intensive workloads, providing excellent performance.
filedb: A Disk-Based Key-Value Store Inspired by Bitcask in Zig
filedb is a Zig-implemented, disk-based key-value store drawing inspiration from the Bitcask paper by Riak. It offers high throughput and efficient O(1) record fetching by storing metadata in a log-structured hashtable and data in append-only files. The project also provides a Redis-compatible client for easy integration and benchmarking.