What is a Cache?
7 min read
A cache is a specialized high-speed data storage layer that temporarily stores copies of files, database queries, and rendered web pages so subsequent requests can be served significantly faster than querying the origin server. In web architecture and technical SEO, caching occurs at four distinct tiers: browser cache (local client storage), Content Delivery Network (CDN) edge cache (Cloudflare, Fastly), application cache (Redis, Memcached, Varnish), and database query cache. Proper caching minimizes Time to First Byte (TTFB), improves Core Web Vitals, and conserves search engine crawl budget.
The 4 Core Tiers of Web Caching Explained
Whenever a user browser or a search crawler requests a page on your domain, caching mechanisms determine whether that request travels thousands of miles to your origin database or resolves instantly from memory:
| Caching Tier | Where Data Resides | Typical Response Time | SEO & Infrastructure Role |
|---|---|---|---|
| 1. Browser Cache | Visitor device (RAM / Local Disk) | 0 to 10ms | Caches CSS, JS bundles, and brand logos via Cache-Control: max-age for instant repeat pageviews. |
| 2. Edge CDN Cache | Global PoPs (Cloudflare, Fastly, CloudFront) | 15 to 50ms | Delivers full static HTML and WebP assets from servers physically closest to the visitor or Googlebot. |
| 3. Application Cache | Origin Server Memory (Redis, Memcached) | 20 to 100ms | Stores pre-rendered HTML snippets and session tokens, preventing repeated server-side PHP/Node rendering. |
| 4. Database Cache | Database Engine (MySQL, PostgreSQL) | 50 to 150ms | Retains recurring SQL result sets in RAM so the database engine avoids disk reads. |
How Caching Directly Impacts Core Web Vitals & Crawl Budget
Search engines prioritize fast, reliable user experiences. Uncached origin servers take 800ms to 2,000ms just to begin delivering the first byte of HTML, severely damaging performance metrics:
- Time to First Byte (TTFB): CDN edge caching routinely drops TTFB below 150ms globally, satisfying Google’s "Good" Core Web Vitals threshold (under 800ms).
- Largest Contentful Paint (LCP): Caching hero images, WebP assets, and font files locally on user devices and edge nodes eliminates render-blocking latency.
- Crawl Budget Efficiency: When Googlebot crawls a large website, slow server responses force the crawler to throttle request velocity. A properly cached origin allows search bots to index thousands of pages without degrading site uptime.
Essential HTTP Headers for Cache Control
Web servers instruct browsers and proxy servers how to handle cached copies through standardized HTTP response headers:
Core Caching Headers Cheat Sheet
Cache-Control: public, max-age=31536000, immutable
Used for versioned static assets (hashed CSS, JavaScript, WebP images). Tells browsers and CDNs to store the file for one full year without revalidating.
Cache-Control: public, max-age=3600, must-revalidate
Ideal for dynamic HTML pages. Stores the page at the edge for one hour, then checks origin for updates.
ETag: "686897696a7c876b7e"
A unique content fingerprint. If the file has not changed, the server returns an ultra-lightweight 304 Not Modified response with zero body bytes.
Cache Invalidation Pitfalls & How to Avoid Stale Content
As the classic computer science adage notes, cache invalidation is notoriously tricky. If your caching rules are configured incorrectly, two severe issues emerge:
- Stale Commercial Information: Prospective customers and AI answer engines might see outdated pricing tiers, obsolete feature limits, or superseded compliance data if CDN caches are not purged upon publish.
- Personalization Leakage: Caching dynamic user-specific pages (like logged-in user dashboards or checkout carts) at a public CDN tier can inadvertently expose sensitive account data to other users. Public cache headers must strictly apply only to anonymous, static content.
Frequently Asked Questions
Does caching help Googlebot index my website faster?
Yes. When your pages resolve quickly from memory or an edge CDN, Googlebot consumes fewer origin server resources per request, allowing it to crawl and index more pages within your allocated crawl budget.
What is the difference between client-side and server-side caching?
Client-side caching stores files directly on the user device (browser), whereas server-side caching stores pre-rendered HTML or database query results in server RAM (like Redis or a CDN edge node) to serve multiple different users without repeated computation.
How does an ETag header work with caching?
An ETag (entity tag) is a unique string hash assigned to a file version. When a browser requests the page again, it sends the ETag; if the content has not changed, the server replies with a fast 304 Not Modified code, avoiding full data transfer.