Requiem for Jon Postel

Humanity lost a defender today. Against the squalid backdrop of human nature, one man tested his mettle against the forces of chaos and entropy, without regard for his own personal safety, and gave our civilization some extra time to get itself organized before advancing along the track of history.

The man was Jonathan B. Postel, who passed away 16 October 1998 from complications following cardiac surgery. His family has asked that we all respect their privacy so they can deal with their loss. I'd like to share some of my memories here, to give those who didn't know him a glimpse of what they missed.

My earliest memory of Jon was in reading his network protocol documents. He was a pillar from the ARPAnet's earliest days, and it is no exaggeration to say that the Internet's landscape -- from deep to broad and from high to low -- would have come about very differently without his long work. As an author and as an editor, his documents were clear, readable, and unambiguous. When the IETF was later created, its philosophic charter was inherited in large part from Jon's work as a protocol designer and as the "RFC Editor."

But then in 1987 I had my first run-in with him, over the proper direction of the top-level US domain's development. Jon had a unique perspective on how US-DOM should be organized, and he was surprisingly gracious to me (an unknown newcomer), considering my lack of agreement, my lack of understanding, and my lack of civility.

A few years later, one of Jon's students contributed a moderate amount of work toward some software I maintain, and Jon pretty much let the kid fend for himself as I ripped both the kid and his work to shreds. I ended up feeling that I'd been used by Jon to prove a point, but used to a good purpose.

Finally, in 1993, I met the man himself, at my first IETF meeting. I was in the process of extending the DNS protocol to add a feature, and I needed an opcode number for my new feature. Someone told me, "Go ask the IANA." I'd heard about the IANA and figured that I was in for a long and complicated negotiation with some large and shadowy organization. No, I was told, "He's around here somewhere, I saw him a few minutes ago." The man was found, my need was explained, and off we went to the terminal room where Jon logged in remotely to his computer back at ISI, and edited a file to allocate me my number. The Internet Assigned Numbers Authority had acted decisively and without either haste or delay.

When InterNIC was first given approval to charge money for domain name registration, I was one of about a dozen people who called Jon to rant and rave about how "there oughta be a law" and how the IANA should step in and make things better. Jon listened to everyone, humored them, and then made his own plans -- just like always.

As one of the many people who had been pushing Jon to assert control of the DNS top level domain name space (the so-called "root servers"), I was quite surprised when he called one day and said, "OK, let's test it."

The political environment of that week was such that I had to question the move: "Is this the best time?" He said it was, without elaboration, and so I helped perform the various behind-the-scenes machinations that made his somewhat controversial tests possible. Watching the events of that week was like watching a sailboat stare down a battleship, but ultimately the battleship turned. Jon always did what he had to do, and usually it ended up being the right thing for all of us.

When persuasion was called for, Jon cajoled. When threats were needed, he issued them. If reason could work he tried it. If anger was called for, he felt it but rarely showed it publicaly. As the glue that held the Internet together for its first 30 tumultuous years, he will be sorely missed.

In discussions of the future of the IANA, his own death sometimes came up as a topic of discussion, since many people saw no distinction between IANA (the man) and IANA (the office).

He cared very much about how things would go for the rest of us in the event of his death and, based on past health problems, he had a personal motivation to "outplace" his cherished IANA. That outplacement was to be the crowning jewel of his career, and he planned to retire after seeing IANA take on a new and separate life from his own. He lived long enough to see the paperwork signed, but the rest is up to us.

Paul Vixie is a respected network administrator and consultant. He is the author of the Berkeley Internet Name Domain (BIND), one of the Net's underlying domain name server utilities.

Related Wired Links:

Net Mourns Passing of Giant
18.Oct.98

I Remember IANA
18.Oct.98

A Life too Brief
18.Oct.98

Meet the ICANN Board
8.Oct.98