The physics are stubborn: a request from Sydney to a server in Virginia pays for the distance in every round trip. A CDN sidesteps the problem by keeping copies of your content in hundreds of cities, so most requests are answered around the corner instead of across the planet.
Hits, misses, and what gets cached
When the edge has a fresh copy, that is a cache hit, served in tens of milliseconds. When it does not, the request travels to your origin, and the response is stored on the way back for the next visitor. By default, CDNs cache static assets, images, fonts, CSS, JavaScript, and treat HTML as dynamic, because HTML is where logged-in states and personalization live.
Getting more out of the cache
- Version your assets. Build tools fingerprint filenames, so assets can be cached for a year and still update instantly on deploy.
- Cache HTML deliberately. Marketing pages and articles that are the same for everyone can often be cached at the edge for minutes at a time, which turns traffic spikes into non-events.
- Purge on publish. A cache you can clear on demand is a cache you can trust with longer lifetimes.
Watch the hit ratio, not the ad copy
The number that tells the truth is your cache hit ratio. If 90% of requests are answered at the edge, your server does a tenth of the work and visitors get the speed regardless of geography. The gains compound with application-level tuning covered in building a web performance culture, and they matter doubly on image-heavy storefronts, where speed is revenue. Content-driven sites see the same effect; a well-cached WordPress install feels like a different product.