About Caching Strategies

Where to cache, patterns, invalidation, common bugs. Part of the DevTools Surf developer suite. Browse more tools in the Info / Guides collection.

Use Cases

• Backend developers implementing Redis caching for database queries
• Frontend engineers configuring HTTP cache headers for static assets
• Platform teams designing multi-layer caching for high-traffic APIs
• SREs diagnosing cache hit ratios and stale data in production systems

Tips

• Start with the 'where to cache' hierarchy: browser, CDN, app, database
• Check the invalidation patterns to avoid serving stale data
• Review the common bugs section for cache stampede and thundering herd issues

Fun Facts

• Phil Karlton's famous quote — 'There are only two hard things in Computer Science: cache invalidation and naming things' — dates to the early 1990s at Netscape.
• CDN caching can reduce origin server load by 60-90%. Cloudflare reported in 2023 that it serves over 57 million HTTP requests per second from cache globally.
• The write-through vs write-behind (write-back) distinction originated in CPU cache design in the 1960s and was later adopted for application-level caching patterns.

FAQ

What are the common patterns?

Cache-aside, write-through, write-behind, refresh-ahead. Each has tradeoffs in consistency, complexity, and latency.

Where should I cache?

Browser, CDN, reverse proxy, application memory, or database. The page has a decision flowchart based on update frequency and consistency needs.

How do I invalidate?

The hard problem. TTL (time-based), event-driven (write triggers invalidation), or versioned cache keys (URL-level bust). Each trades off consistency vs complexity.

Thundering herd — what is it?

When cache expires and 1000 requests hit the origin simultaneously. Solutions: request coalescing, stale-while-revalidate, jittered TTLs.