What happens when you type a URL
You type pangaea.id, press Enter, and the page shows up in a blink. That blink hides a small
global relay race — a name gets turned into an address, your request finds the nearest of
hundreds of data centers, a lock clicks shut, and a page that was built days ago is handed back.
Here's every leg, in plain English.
1. The name isn't an address
pangaea.id is a name, made for humans. The machines that move traffic need a number — an
IP address. Turning one into the other is the job of DNS, the internet's phone book.
Your browser asks a resolver (your ISP's, or one like 1.1.1.1): "what's the address for
pangaea.id?" If nobody nearby remembers, the resolver walks the chain — the root, then the .id
directory, then the authoritative nameserver that actually holds the answer. For us, that
authority is Cloudflare — where we pointed the domain, a story of its
own. It hands back an IP, and the resolver remembers it for next time.
2. One address, many cities
Here's the trick: the IP Cloudflare returns doesn't belong to a single machine in one place. The
same address is announced from hundreds of data centers worldwide — a technique called
anycast. The internet quietly routes your request to the nearest one. Someone in Jakarta
and someone in Berlin type the same pangaea.id and each lands on a copy a few milliseconds away.
Distance is the enemy of speed; anycast deletes most of it.
3. A quick hello, then a lock
Your browser and that nearest edge do a two-part greeting:
- TCP — a short "hello / hello back / got it" so both sides agree they're connected.
- TLS — they swap keys and switch to encryption. That's the padlock in the address bar. From here on nobody in between — café Wi-Fi, your ISP — can read or tamper with the page.
It sounds like a lot, but it's a couple of round-trips, and the browser reuses that one connection for everything else the page needs.
4. "Give me the homepage"
Now the actual request: the browser sends HTTP — literally GET /. The edge usually already
has the answer in its cache, because our pages are built ahead of time into plain HTML and
copied out to every edge. So it replies in milliseconds, with no server "thinking" required. (Only
on a rare first miss does the edge fetch the page once from the origin — Cloudflare Pages — then
cache it for everyone after you.)
5. The browser paints
The HTML comes back and the browser reads it top to bottom. It sees it also needs some CSS, a little JavaScript, and the fonts — and asks for those too, again from the nearby edge (and those files are fingerprinted, so they're cached "forever"). Then it paints: header, headline, the rest. The blink is over.
We chose this on purpose
None of it is luck — it's the setup we picked: the domain points at Cloudflare's edge, and the site is static files on Cloudflare Pages. The how-and-why is in our build diary: Point the domain at Cloudflare → and Ship it: Git → CI/CD → Pages →.
Sources