DNS: what it does and why it exists

You type google.com into your browser. The browser does not know where Google is — it knows names, not numbers. Before it can open a TCP connection or send an HTTP request, it needs the IP address. That lookup is DNS, and it happens silently on every page load.

What DNS does in one sentence

DNS (Domain Name System) translates a human-friendly name like google.com into the numeric IP address (142.250.185.46) the network layer needs to route packets. Without DNS you would need to memorise dozens of IP addresses per service and update them every time a server moved.

The tree of servers

DNS is a global hierarchy. Three tiers handle every query:

• Root servers — 13 named servers (A–M), each replicated to hundreds of locations via anycast. They do not store domain records; they know which TLD server to ask next.
• TLD servers — one set per top-level domain (.com, .org, .uk, .io, etc.). They know which authoritative server owns a given domain.
• Authoritative nameservers — own the zone files for individual domains (google.com, example.co.uk). They return the final answer.

The phonebook metaphor

Imagine you want to call a restaurant you know only by name. You call directory assistance (the recursive resolver). The operator does not know the number but knows how to find it. She calls a central office (root server), gets redirected to a regional directory (TLD server), and from there to the restaurant’s front desk (authoritative server). The front desk gives the number; the operator relays it and writes it down. The next caller asking the same question gets the cached answer instantly.

The resolution dialogue

A user’s browser sends a query to its configured recursive resolver. If the resolver has no cached answer, it walks the tree:

1. Browser asks resolver: “What is the IP for example.com?”
2. Resolver asks root: “Who handles .com?” Root: “Ask the .com TLD server.”
3. Resolver asks .com TLD: “Who handles example.com?” TLD: “Ask this authoritative server.”
4. Resolver asks authoritative: “IP for example.com?” Auth: “203.0.113.10, TTL=300.”
5. Resolver returns IP to browser, caches the answer for 300 seconds.

On the next query within those 300 seconds the resolver skips steps 2–4 and answers from cache.

One scenario end to end

You click a link. Underneath, the browser does DNS first — to learn the IP. Once it has the IP it opens a TCP connection to that address, then adds TLS on top, and finally sends the HTTP request. Every new browser tab repeats the DNS step unless the answer is still cached. DNS is always the first step of an internet connection.