How does your browser find a website from just its name?

Your browser asks a global directory called the Domain Name System, DNS. Computers find each other by numerical IP addresses, but people prefer names, so DNS translates human-friendly host names into addresses. It is often called the phone book of the internet, with one important difference: it is not a single book.

It started as one. At the Stanford Research Institute, a text file called HOSTS.TXT mapped host names to the numerical addresses of the computers on the ARPANET. Addresses were assigned manually: to add a machine, you telephoned the Network Information Center, run by Elizabeth Feinler, during business hours. Feinler also suggested domains, such as edu for computers at educational institutions. But by the early 1980s a single, centralized table had become slow and unwieldy. In 1983 Paul Mockapetris created the Domain Name System instead, a design meant to avoid one large central database.

In DNS, no one knows everything and everyone knows who to ask next. Your device usually passes the question to a DNS server run by your internet provider, whose resolver reads the name from right to left. It asks a root server about the last part, such as com, and gets pointed to the servers responsible for that domain, and then to ever more specific servers, until one of them gives the actual answer. Because that information rarely changes within hours, intermediate servers and your own browser cache the answers, so a repeated question can be answered from a cache instead of climbing the chain again.