Redis (Remote Dictionary Server) is an open-source, in-memory data structure store used as a database, cache, message broker, and streaming engine. Unlike traditional disk-based databases, Redis holds all data in RAM, delivering sub-millisecond read and write latency at scale.
Redis is a key-value store at its core, but it supports a rich set of data structures that make it far more versatile than a simple cache layer. Every value in Redis is associated with a string key, and the value itself can be one of many types: strings, lists, sets, sorted sets, hashes, bitmaps, hyperloglogs, or streams.
Redis is single-threaded for command execution, which eliminates the need for locks and makes all operations inherently atomic. This architecture is what allows Redis to guarantee predictable, race-condition-free behavior under high concurrency without complex locking strategies.
In-Memory Architecture
All data is stored in system RAM. Reads and writes bypass disk I/O entirely, resulting in throughput measured in millions of operations per second on standard hardware. Redis can optionally persist data to disk (via RDB snapshots or AOF logs), but the working dataset always lives in memory.
Key Expiration (TTL)
Every key in Redis can be assigned a Time-To-Live in seconds or milliseconds. Once the TTL expires, Redis automatically evicts the key. This is the foundation of cache invalidation — you set a TTL and Redis handles cleanup without any application-side polling or cron jobs.
Atomic Operations
Commands like INCR, GETSET, SETNX, and MULTI/EXEC (transactions) execute atomically. No other command can interleave between the steps of an atomic operation. This property is critical when Redis is used as a rate limiter or a distributed lock provider.
Pub/Sub and Streams
Redis supports a publish/subscribe messaging model and a persistent stream data type (XADD, XREAD), making it suitable as a lightweight message broker for event-driven architectures.
Cluster and Replication
Redis supports primary-replica replication for read scaling and high availability. Redis Cluster shards data across multiple nodes using hash slots, enabling horizontal scaling beyond the memory of a single machine.
| Use Case | How Redis Helps |
|---|---|
| Response Caching | Store serialized API responses with a TTL; serve from memory instead of hitting the database |
| Rate Limiting | Use atomic INCR and EXPIRE or SETNX with sliding window logic to enforce request quotas |
| Session Storage | Store session tokens as hash maps with TTL-based expiry |
| Job Queues | Use LPUSH / BRPOP (or Redis Streams) to implement FIFO task queues |
| Leaderboards | Use sorted sets (ZADD, ZRANK) to maintain real-time ranked lists |
| Distributed Locks | Use SET key value NX PX ttl for atomic lock acquisition across processes |
Redis Cloud is the managed Redis service maintained by Redis Inc. It offers a permanently free tier suitable for development, staging environments, and low-traffic production workloads.