Jon Postel

Jon Postel: The Man Who Ran the Internet Alone

In January 1998, Jon Postel sent an email to eight of the twelve organizations running the Internet's root name servers, quietly asking them to redirect their machines away from the government-designated authority and toward a server he controlled himself. They did it — because for nearly thirty years, when Jon Postel asked the Internet's infrastructure to do something, it did. Within days, a presidential science advisor was threatening him personally: "You'll never work on the Internet again." It was possibly the last moment in history when one individual could plausibly redirect the internet's core plumbing by email and be obeyed without a formal vote — and it captures, in miniature, exactly how much informal, personal trust the network had run on for its entire early existence.

A Mediocre Student Among Future Legends

Jonathan Bruce Postel was born August 6, 1943, in Altadena, California. He attended Van Nuys High School alongside Steve Crocker and Vint Cerf, two classmates who would become fellow architects of the early internet — though by his own academic record, Postel was the least distinguished student of the three. He went on to UCLA, earning a bachelor's degree in engineering in 1966, a master's in 1968, and a Ph.D. in computer science in 1974 under advisor Dave Farber. In December 1969, before finishing his doctorate, Postel began working at UCLA as a postgraduate research engineer on ARPANET, the Defense Department network that would become the internet's direct ancestor, collaborating with Cerf and Crocker on the earliest network protocols and pushing for additional protocols to handle communication between separate networks — the conceptual seed of what became the internet protocol suite.

Editing the Rules Everyone Followed

Postel's most enduring institutional role began almost immediately and never really ended: from 1969 until his death, he served as editor of the Request for Comments series, the documents through which the internet's technical standards are proposed, debated, and formally adopted. He personally authored or co-authored more than twenty of them, including RFCs 791, 792, and 793 — the foundational specifications for the Internet Protocol and Transmission Control Protocol that still underlie essentially all internet traffic — and the RFCs that established the architecture of the Domain Name System. From the earliest days of network numbering, Postel also directed the Internet Assigned Numbers Authority, IANA, initially performing the entire task of allocating network addresses and protocol parameters by hand, and administered the .us country-code domain while serving on the Internet Architecture Board and the Internet Society's board of trustees. His day job throughout this period was directing the Computer Networks Division at the University of Southern California's Information Sciences Institute, a post he took up in March 1977.

The Robustness Principle

Postel's single most quoted contribution to computer science is not an institution but a sentence — a design principle first articulated in RFC 760 and later restated in RFC 1122: "an implementation should be conservative in its sending behavior, and liberal in its receiving behavior," commonly summarized as "be liberal in what you accept, and conservative in what you send." Known ever since as Postel's Law, the principle shaped how network software was written for decades, encouraging systems to tolerate malformed or unexpected input from others while holding themselves to strict standards in what they transmitted — a piece of engineering philosophy that outlasted most of the specific protocols it was written to govern.

The 1998 Root Server Test

The defining controversy of Postel's career arrived near its end. On January 28, 1998, with the U.S. government and Network Solutions Inc. locked in an unresolved dispute over control of the domain name system's commercial future, Postel emailed eight of the internet's twelve root nameserver operators and instructed them to redirect their servers from Network Solutions' authoritative root at A.ROOT-SERVERS.NET to an IANA-controlled machine, DNSROOT.IANA.ORG. The operators, many of whom had known and trusted Postel personally for decades, complied without incident, and ordinary internet traffic continued uninterrupted throughout. But the political fallout was immediate and severe: Ira Magaziner, the Clinton administration's internet policy advisor, personally warned Postel that he would "never work on the internet again" and ordered the test reversed. Within a week the National Telecommunications and Information Administration issued proposals that ultimately expanded, rather than reduced, formal U.S. government oversight of DNS management — arguably the opposite of what Postel's action had been intended to demonstrate was unnecessary.

Death and an Outpouring of Tribute

Postel died on October 16, 1998, at age 55, from complications following heart valve replacement surgery in Santa Monica — just months after the root-server confrontation and shortly before the creation of ICANN, the organization that would formally institutionalize the authority Postel had exercised informally for decades. The tributes that followed were unusually personal for a technical field: Vint Cerf wrote RFC 2468, "I Remember IANA," memorializing Postel directly within the internet's own core technical literature, an honor with no real precedent. The Internet Society created the Jonathan B. Postel Service Award in his name in 1999, naming Postel himself its first posthumous recipient; he was inducted into the Internet Hall of Fame as a Pioneer in 2012; and USC's Information Sciences Institute named its Postel Center after him. When a journalist once called him "the god of the internet," Postel deflected the label with characteristic understatement: "Of course, there isn't any 'God of the Internet.' The Internet works because a lot of people cooperate to do things together." The Economist was less willing to let him off the hook, writing that "if the Net does have a god, he is probably Jon Postel."

Why Jon Is Called a Genius

The genius attributed to Postel is unusual among the figures who get the label: it is not a single invention or product but three decades of quiet, largely unpaid stewardship that held together a technical system too complex and too rapidly growing for any formal bureaucracy of the era to manage as well as one trusted individual could. Cerf's tribute in RFC 2468 and the Economist's "god of the internet" framing both point to the same underlying claim — that the internet's early cohesion depended less on any protocol document than on Postel's personal judgment, evenhandedness, and willingness to arbitrate technical disputes that had no obvious formal authority to resolve them. Postel's Law itself is regularly cited in software engineering curricula as a genuinely original and durable insight about system design.

The honest complication is that this model of stewardship — one person, largely unaccountable, making consequential decisions about global infrastructure by personal email — was also precisely what the 1998 root server incident exposed as fragile and, to the U.S. government, alarming. Postel's authority worked because people trusted him individually; it was never a democratically accountable structure, and its replacement by ICANN after his death was widely understood as a necessary formalization of power that could not, in an increasingly commercial and politically contested internet, continue to rest on one engineer's judgment.

Legacy

Postel died before the internet he helped govern became the mass consumer and commercial platform it is today, but the protocols he wrote or edited — TCP/IP and the Domain Name System chief among them — remain the literal substrate of the modern internet, and the institutional structures created in his wake, from IANA's continuation to ICANN itself, still bear the shape of the informal, trust-based system he built almost single-handedly.

Achievements

• Internet Hall of Fame — 2012
• Educated at Van Nuys High School, University of California, Los Angeles and University of California, Los Angeles
• Worked as computer scientist and engineer