Common Redis Use Cases in Real-World Projects

1. Caching Hot Data

High‑frequency items such as flash‑sale products, trending news, or hot‑search lists are pre‑loaded into Redis to boost cache hit rates and shield the database from sudden spikes. Pre‑warming methods include offline statistics (analyzing logs to identify hot keys) and scheduled scripts that load data during low‑traffic periods. The article warns of cache‑related issues like cache breakdown, avalanche, and penetration.

2. Distributed Lock

In distributed architectures, traditional JVM locks (synchronized, ReentrantLock) are ineffective. Redis provides a distributed lock, often wrapped by the Redisson library, which includes a watchdog that automatically extends the lock’s TTL during long business processes, preventing premature expiration.

3. Bloom Filter

The article explains the Bloom filter principle: a key is hashed by multiple functions to set bits in a bitmap; presence is checked by testing those bits. Because the filter stores only bits, it consumes minimal memory while blocking malicious requests that query non‑existent data, thus mitigating cache penetration.

4. Delayed Queue

Small‑to‑medium projects can implement delayed queues with Redis ZSet combined with a scheduled task. Elements’ scores store expiration timestamps; a periodic scan moves due items to the processing queue.

5. Rate Limiting

Redis hash structures implement a token‑bucket algorithm, smoothing burst traffic and rejecting malicious requests to maintain system stability.

6. Boolean Data Statistics

For massive boolean datasets, Redis Bitmap (a bit‑operation extension of the String type) stores each value as a single bit, drastically reducing memory usage while enabling fast bitwise statistics. Typical scenarios include user sign‑in tracking, online status, and permission flags.

7. UV Deduplication

To count unique visitors, three approaches are described: (1) a Set collection where each visitor ID is added and the set size yields UV; (2) HyperLogLog, a probabilistic cardinality estimator using PFADD, PFCOUNT, and PFMERGE; (3) Bitmap, where SETBIT marks visits and BITCOUNT estimates unique counts.

8. Geospatial Calculations

Redis GEO, built on ZSet and GeoHash encoding, stores latitude/longitude as a 52‑bit integer score. It leverages skip‑list structures and the Haversine formula for efficient radius searches, supporting features like nearby‑place search, ride‑hailing matching, and “people nearby” social functions.

9. Lightweight Message Queue

Redis Stream improves on the earlier List (LPUSH + BRPOP) and Pub/Sub by adding message acknowledgment and persistence, making it suitable for asynchronous decoupling in small‑to‑medium projects. For high‑throughput, strong‑consistency needs, the article notes that Kafka or RocketMQ are more appropriate.