In-memory Database

An in-memory database keeps its primary dataset in RAM rather than on disk, trading capacity and durability tradeoffs for orders-of-magnitude lower read and write latency. In-memory engines back almost every cache, session store, leaderboard, rate limiter, and many real-time analytics systems.

Common engines

• Redis. The dominant in-memory key-value store; rich data structures, optional persistence (RDB, AOF), replication, cluster sharding.
• Memcached. Pure cache; no persistence; multi-threaded; very simple.
• Valkey. Open-source fork of Redis after the 2024 license change; AWS and Linux Foundation backed.
• DragonflyDB. Redis-compatible multi-threaded engine optimised for modern hardware.
• KeyDB. Redis fork with multi-threading.
• SAP HANA, MemSQL/SingleStore. In-memory relational and HTAP systems.
• Hazelcast, Apache Ignite. In-memory data grids for distributed compute.

Durability strategies

Most in-memory engines support some level of disk persistence to survive restarts:

• Snapshots. Periodic point-in-time dumps to disk.
• Append-only log. Every write recorded sequentially; replayed on startup.
• Replication. Multiple replicas reduce the chance that all copies fail at once.