Why Redis Leads Modern Key‑Value Stores: Features, Use Cases, and Performance

Overview

Redis is an open‑source, ANSI‑C, in‑memory key‑value database with optional disk persistence. It supports richer data structures than Memcached: strings, lists, sets, sorted sets, and hashes, all with atomic operations.

Key‑Value Store Fundamentals

Key‑value stores provide high‑performance, scalable access to massive data sets, suitable for large‑scale internet services. They avoid schemas, transactions, and full SQL, enabling distributed performance.

Characteristics

Basic operations : SET key value, GET key.

Distributed : multiple nodes share data and maintain consistency.

Consistency & redundancy : synchronous updates and data replication.

Fault tolerance : node failures do not affect overall service.

High reliability : redundancy and fault tolerance together.

Redis Data Types

Strings : simple values, support atomic increment.

Lists : ordered collections, push/pop, useful for queues.

Sets : unordered unique elements, support set operations.

Sorted sets : elements ordered by numeric score, ideal for leaderboards.

Hashes : field‑value maps, similar to objects.

Persistence

Two persistence options:

RDB (snapshot) : periodically writes the entire dataset to disk; fast but may lose recent writes.

AOF (append‑only file) : logs every write operation; more durable but slower.

Master‑Slave Replication

Redis can replicate data to multiple slave instances, improving read scalability and providing redundancy.

Performance

On a Xeon X3320 2.5 GHz Linux server, Redis achieves ~110 k SET ops/sec and ~81 k GET ops/sec. StackOverflow uses Redis as a cache layer.

Typical Use Cases

1. Retrieve latest N items

Use LPUSH to add IDs to a list and LTRIM to keep only the most recent N entries.

2. Leaderboards (Top N)

Store scores in a sorted set with ZADD; retrieve top elements with ZREVRANGE.

3. Precise expiration

Encode expiration timestamps as sorted‑set scores and periodically remove expired members.

4. Counters

Atomic INCR and DECR commands provide high‑throughput counting.

5. De‑duplication

Insert values into a set; duplicates are automatically eliminated.

6. Real‑time messaging (Pub/Sub)

Redis Pub/Sub enables live chat and other real‑time communication patterns.

7. Queue systems

Lists implement FIFO queues; sorted sets can implement priority queues.

8. Caching

Redis offers superior performance to Memcached with richer data structures.

References

Official site: https://redis.io/ Chinese site: http://www.redis.cn/