BEGINNER · DOMAINS · 6 min read · Updated 2026-06-02

A brief history of domain names

How we went from a single file passed around by hand to a billion-name, self-governing directory.

By the Check.Systems team

TL;DR

Domain names exist because a single shared file of computer names stopped scaling. DNS replaced it in 1983, .com arrived in 1985, the system was commercialized in the 1990s, and ICANN took over coordination in 1998 — leading to today's thousand-plus TLDs.

The file that didn't scale

Picture the early internet as a small town where everyone shared one address book — a single sheet of paper, photocopied for the whole town. In the 1970s and early '80s, that's roughly how it worked. Every computer on the ARPANET kept a file called HOSTS.TXT that mapped names to addresses, and the Network Information Center at SRI International — run by Elizabeth "Jake" Feinler — maintained the master copy by hand.

Want to add your machine, or change its address? You emailed the maintainers, and everyone downloaded the updated file. It worked when "the internet" was a few hundred machines. But as the network grew, the cracks showed: the file got huge, downloads strained the network, edits collided, and two people could claim the same name. A town address book doesn't scale to a planet.

1983: DNS is invented

The fix, designed by Paul Mockapetris in 1983, was the Domain Name System — and its key idea was delegation. Instead of one central file, the namespace was split into a tree. A central root would only need to know "who's in charge of .com," and the .com operators would know "who's in charge of example.com," and so on down.

That hierarchy is the breakthrough that still runs today: no single computer holds every name, so the system scales almost without limit. (For how a modern lookup walks that tree, see What is DNS? — or watch one happen in the DNS lookup tool.)

The first domains

The original generic TLDs were defined in 1984–85: .com, .org, .net, .edu, .gov, and .mil. Then came the famous firsts:

March 1985 — symbolics.com, registered by a computer company, became the first ever .com domain. (It still exists, now preserved as a piece of history.)
Universities and research bodies quickly filled .edu and .org.
For years registration was free — handled by government-funded operators as a public service.

For most of the 1980s, almost nobody outside research and government had a domain. There was simply no commercial internet to need one yet.

Money, congestion, and InterNIC

The web arrived in the early 1990s, and demand for names exploded. In 1993 the U.S. government consolidated registration under a project called InterNIC, with a company named Network Solutions handling .com, .org, and .net.

Then, in 1995, the pivotal change: registration started costing money — initially $100 for two years. That's the moment a domain went from a free public allotment to a rentable asset, and it set off the land-rush (and the cybersquatting) that defined the dot-com era. The "you rent, you don't own" model we live with today dates from here.

ICANN and going global

A single company controlling the most valuable TLDs, under one government, wasn't a stable arrangement for a global network. In 1998, a new non-profit — the Internet Corporation for Assigned Names and Numbers (ICANN) — was created to coordinate the namespace internationally.

ICANN's most important move was opening the registrar market to competition. Instead of one operator selling every .com, hundreds of accredited registrars could now compete — which is why a domain costs a few dollars today instead of fifty. The clean split we still use — one registry per TLD, many registrars selling it — comes from this era. (See What is a domain registrar?)

The new-gTLD explosion

For a long time the generic options stayed small — a couple of dozen TLDs. A first wave of additions arrived in the 2000s (.info, .biz). Then in 2012, ICANN threw the doors open with the new gTLD program, inviting applications for almost any word: .app, .shop, .dev, .xyz, even brand TLDs like .google.

The result is today's landscape: well over a thousand TLDs, a few hundred million registered domains, and a system that — remarkably — still rests on the 1983 idea of a delegated tree.

The throughline

Every leap in this story solved a scaling problem: HOSTS.TXT → DNS (one file couldn't hold everyone), free → paid (demand outran a public service), one operator → ICANN + competition (no single party should hold the keys). The same pressures that shaped domains are still shaping the internet.

So the next time you register a name in thirty seconds for the price of a coffee, it's worth remembering it sits on forty years of solving the same question over and over: how do billions of machines agree on who's called what — without anyone being in charge of all of it?

References: RFC 882 (original DNS, 1983) · RFC 1034 (DNS concepts) · ICANN — History