Domain Name System

The Domain Name System emerged because the early internet was drowning in its own success. By 1983, ARPANET and CSnet had grown to thousands of hosts, each needing to know how to reach the others. The solution was a single text file—the "host tables"—maintained by SRI International in Menlo Park. Every computer on the network had to download this file to know where to send traffic. But the file was growing exponentially, updates took days to propagate, and name collisions became inevitable.

Jon Postel at USC's Information Sciences Institute recognized the problem was architectural, not administrative. He assigned Paul Mockapetris to devise a replacement. Mockapetris proposed something radical: instead of a central phonebook, create a distributed database where each domain maintained its own records. The system would be hierarchical, like a postal address—city, state, country—but inverted, reading right to left: .edu, .mit, .web.

Mockapetris published the DNS architecture in RFC 882 and RFC 883 in November 1983. The first working nameserver, called "Jeeves," ran on DEC Tops-20 machines at USC-ISI and SRI's Network Information Center. The six original top-level domains—.edu, .gov, .com, .mil, .org, and .net—partitioned the namespace by function. DARPA prodded adoption; by 1986, DNS had entirely supplanted host tables.

The adjacent possible that enabled DNS included packet-switched networking, which allowed queries to traverse the network efficiently; the hierarchical namespace concept, borrowed from file systems; and the growth of the network itself, which created the problem DNS solved. Without thousands of hosts, a single file sufficed. Success created the crisis that demanded distributed infrastructure.

The cascade from DNS made the internet human-usable. Users could type "mit.edu" instead of memorizing 18.7.21.69. Websites could change servers without breaking bookmarks. Email addresses became memorable. When Tim Berners-Lee invented the World Wide Web in 1989, URLs depended entirely on DNS to translate domain names to IP addresses. Every web page load, every email delivery, every API call begins with a DNS query.

Path dependence locked in the hierarchy. The original six TLDs expanded to hundreds, including country codes (.uk, .jp, .de) and vanity extensions (.io, .ai, .club). ICANN now governs the root zone, controlling which TLDs exist. The DNS root servers—thirteen logical clusters distributed globally—form one of the internet's most critical and most attacked infrastructure layers.

By 2026, DNS handles trillions of queries daily, invisible to users but essential to every internet interaction. Mockapetris' 1983 architecture remains fundamentally unchanged—a distributed, hierarchical, cached database translating human-readable names to machine-readable addresses. The "too big to update" problem of host tables became the infinitely scalable system that makes the global internet navigable.