Cache Classification and Practical Application Cases for High‑Performance Systems

Cache is a primary tool for improving system performance in internet applications, and this article focuses on classifying caches and sharing practical scenarios encountered in real business.

Cache Classification

Common cache types include client cache, browser cache, CDN cache, NGINX cache, application cache, and unified cache (e.g., Redis). The first four operate at the network transmission layer, while the latter two are used within applications.

Practical Cases

1. Hotspot Key

During a large promotion, HTML fragments for activity pages were cached. The data volume was small but request peaks reached 4 million per minute. Initially Redis was used, but a single hot key caused performance degradation. The solution added a local JVM cache with a one‑minute TTL, falling back to Redis and the database, and later introduced NGINX gzip and proxy caching, reducing response size from 73 KB to 13 KB and stabilizing performance.

2. Category Center Design

The category center serves as a shared data source for many systems. A three‑layer cache architecture was adopted: client cache, JVM cache in the category system, and a unified Redis cache. Updates follow a strict sequence—database update, Redis refresh, JVM cache invalidation across all instances, and version bump—to ensure consistency.

Metrics show that 95 % of client requests hit the client cache, achieving 3.7 million reads per minute with a 99.9th‑percentile latency of 1 ms, while server‑side calls remain well within capacity.

Conclusion

Cache can also serve as a data store to avoid cache‑penetration.

Data updates involve both active and passive cache strategies.

Ensuring data consistency and validity is essential.

Choosing what to cache, how to combine caches, and aligning with business scenarios requires continuous observation and summarization.