P. 1
Launching the DNS War: Dot-Com Privatization and the Rise of Global Internet Governance

Launching the DNS War: Dot-Com Privatization and the Rise of Global Internet Governance

5.0

|Views: 2,086|Likes:
Published by Craig Simon
This dissertation investigates the Internet governance debates of the mid 1990s, narrating events that led to the signing of the Generic Top Level Domains Memorandum of Understanding (gTLD-MoU) in May 1997. During that period, an unlikely alliance formed to create a new institutional structure that would administer the Internet’s Domain Name System (DNS). The collaborators included members of the Internet technical community’s “old guard,” leading officials of the International Telecommunications Union, representatives of organized trademark interests, and others. Their ambitious project aimed at constituting a formal procedural apparatus capable of operating at a world-wide level, independent of the sovereign state system. Institutional membership in the new structure was intended to confer participation rights and normative obligations, thereby establishing status relationships that resonated with the kinship, ingroup, and citizenship relationships of legacy social orders.
The example serves as a particularly valid and germane case study that can be used to model power relations among responsible agents in an expressly global system of rule. This postulated case allows for a more useful comparison of power relations within past, present, and future epochs.
This dissertation investigates the Internet governance debates of the mid 1990s, narrating events that led to the signing of the Generic Top Level Domains Memorandum of Understanding (gTLD-MoU) in May 1997. During that period, an unlikely alliance formed to create a new institutional structure that would administer the Internet’s Domain Name System (DNS). The collaborators included members of the Internet technical community’s “old guard,” leading officials of the International Telecommunications Union, representatives of organized trademark interests, and others. Their ambitious project aimed at constituting a formal procedural apparatus capable of operating at a world-wide level, independent of the sovereign state system. Institutional membership in the new structure was intended to confer participation rights and normative obligations, thereby establishing status relationships that resonated with the kinship, ingroup, and citizenship relationships of legacy social orders.
The example serves as a particularly valid and germane case study that can be used to model power relations among responsible agents in an expressly global system of rule. This postulated case allows for a more useful comparison of power relations within past, present, and future epochs.

More info:

Published by: Craig Simon on Jun 27, 2011
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

11/12/2013

pdf

text

original

The November 20 “Internet Names, Numbers, and Beyond” conference was chaired

th

Brian Kahin, a Harvard-based academic who had been advising the government on Internet

commercialization since the inception of the NSFNET. The event was organized as a series

328

231

of panels, featuring Mockapetris, Rutkowski, St. Johns, Postel, Mike Roberts, Landweber,

and Paul Vixie, who was the lead author of the DNS resolution software used by nearly every

ISP in the world. Some other participants were less familiar to the Internet community, but

in most cases it made sense for them to be there, especially Bob Frank, co-chair of the

International Trademark Association’s Internet Task Force, and Mike Corbett, the FCC’s

specialist on the management of the toll free 800 and 888 zones. The audience was full of

outspoken luminaries including Vint Cerf, David Johnson, Carl Malamud, and Elise Gerich,

plus the managers of Asia’s and Europe’s IP registries, David Conrad and Daniel Karrenberg.

During the course of the day, Vixie, St. Johns, Roberts, and Mockapetris advocated

various strategies for closing .com. Rutkowski objected. Dave Graves and the others in NSI’s

contingent must have been relieved to hear him speak up. Rutkowski played up the fact that

the ITU favored reversion to country codes. (This tainted anyone else who supported the

same strategy.) Moreover, businesses which saw themselves pursuing a global strategy in

a global market wanted to avoid the expense and the extra difficulties registering and names

in every country code. Rutkowski’s arguments apparently influenced Mockapetris, who

switched to a position of “Let the market decide.”

Mockapetris and Vixie were like the bookends of DNS software. Mockapetris had

written the first implementation of a DNS server for the TOPS-20 operating system in 1984,

while still at ISI. He called it “JEEVES,” after the smart butler from the P. G. Wodehouse

tale, “Jeeves and Wooster: Ties that Bind.” Managers at the Digital Equipment Corporation

(DEC) were eager to support development of a UNIX version, believing it would enhance

the market for DEC’s TCP/IP-related hardware. In 1985 the company arranged to have an

employee, Kevin Dunlap, work on the project with a team of graduate students who made

up the technical staff at UC Berkeley’s Computer Systems Research Group, the publishers

of BSD UNIX. The group evidently played on the Wodehouse theme by naming the new

software “Berkeley Internet Name Daemon” – BIND.

Vixie was a relatively young programmer working at DEC’s Palo Alto office in 1988

when he was given responsibility for running the company’s corporate Internet gateway and

about 400 subzones of dec.com. To simplify the task, he undertook a rewrite of the UC

232

From an interview with Dave Wreski, “Paul Vixie and David Conrad on BINDv9 and Internet

329

Security,” LinuxSecurity October 3, 2002, http://www.linuxsecurity.com/feature_stories/conrad_vixie-1.html.

Paul Vixie, “External Issues in Scalability,”http://ksgwww.harvard.edu/iip/GIIconf/vixie.html.

330

Berkeley code (which he once described as “sleazy, icky, rotten” and, worst of all,

unstable). He then made his revision publicly available, including source code and an

329

installation kit as part of the package. It was instantly popular. Vixie’s revisions were so

highly acclaimed that the managers BSD UNIX replaced their version of BIND with his.

DEC subsidized his continuing work on BIND even after he went out on his own in 1993,

providing him with state of the art hardware as he set out to build a new star in the root

server constellation (Hardware contributions came from NSI as well). In short order, Vixie

also received a significant vote of confidence from Postel, who gave him official

responsibility for Root Server F. In 1994, Vixie and Rick Adams of UUNET organized the

Internet Software Consortium (now called the Internet Systems Consortium) to support

Vixie’s work on major upgrades to BIND. As a result of all this activity, Vixie had become

someone to be reckoned with in both the IETF (though he held no formal offices there) and

the North American Network Operators Group (NANOG)

A running theme in Vixie’s discussion at the Harvard conference was the undesirable

effect of “domain envy,” especially the flattening of the namespace. As he saw it, nearly

everyone who wanted a presence on the Web was trying to register in .com because there was

virtually nowhere else to go. He articulated a theme that resonated for years to come: For

the sake of introducing competition, root zone operations needed to be moved from NSI to

the IANA, “where it should have been all along.” And he also insisted on the Internet’s

exceptionalism. Since the DNS namespace was “universal,” he argued, it was “inappropriate

for any regional government to take any kind of leadership role in designing, maintaining,

or funding it.” He wanted to see IANA in charge, under the aegis of ISOC.330

Anticipating that as many as 250,000,000 domain names would eventually be

registered, Vixie believed it was important to start as soon as possible steering new

registrants into a more diverse array of TLDs. He offered different proposals for adding tens

of thousands of TLDs to the root, either by creating short combinations of all the numerals

233

and all the letters in the English alphabet (excluding known words), or by creating zones such

as .a001, .a002, and soon. One of his metaphors was the concept of “Create license plates,

but not vanity plates.”

Postel opened his panel on a note of bemused exasperation, “Great – more work for

the IANA!” He didn’t hide his disapproval of what had happened at NSI. “The reason we are

here is that something changed and we don’t like it.” As he saw it, the challenge was to

create a competitive environment that would provide good service to consumers. Postel

raised a problem that later became known as the “lock-in” dilemma: An owner of a domain

name would be unlikely to drop a name in one TLD and switch to a name under another

suffix simply because of price differences between the two zones. Switching costs were

simply too high to justify it. His logic echoed Vixie’s concerns about the way “domain envy”

was creating a flat namespace. Because of the fast expansion of the .com registry, it was

important to attract consumers into using new suffixes as soon as reasonably possible. Postel

also adopted elements of Bradner’s and Gilmore’s arguments about the need to create

“multiple registries for each domain” rather than simply create new .com style monopolies.

When Postel announced his .a through .z TLD concept, it struck a chord with Kahin,

who believed that new TLDs should be sold to the highest bidder through an auction. Kahin

thought this would solve a host of funding problems, but most others at the conference

weren’t so sure. If there was any confluence of opinion among them it was that the problem

was not money, but trust. With so much demand for names, one could imagine many

different ways of creating solvent – even lucrative – allocation regimes. Yet there was no

organization that everyone could regard as an impartial but qualified authority. International

participants and quite a few Americans didn’t trust the US Government. Too many others

harbored lingering doubts about ISOC.

Rutkowski was still nominally ISOC’s Executive Director at this time, but he was

backhandedly critical of Landweber’s proposal for an ISOC/IAB joint authority. Rutkowski

took the position that no existing organization was capable of providing the kind of oversight

that should be expected of a global administrative authority. Moreover, the sort of leadership

that could be provided by traditional governments was “vestigial and unnecessary.” Instead,

234

he wanted to see the “creation of some new kind of alliance” to do the job. Rutkowski touted

his own World Internet Alliance as an organization that could speak for a broad swath of

stakeholders, particularly ISPs, who were the rising forces of Internet commerce.

Furthermore, he argued, the new organization should be incorporated in a way that would

provide “significant immunities” from litigation, either through a special legislative

enactment, or by siting itself in a jurisdiction such as Switzerland where the legal system was

tilted toward providing such immunities.

The fireworks had begun. Not surprisingly, St Johns refused to back down from his

position that the US Government owned the number space and the root of the names space.

Karrenberg shot back, complaining about US imperialism. For many of those who resented

US power, it made sense to put ISPs in charge. Whoever owned the root, the ISP’s servers

made the root the root. It wouldn’t be long before the number of ISPs outside US borders

would exceed the number inside. And that trend would never reverse. With a joint effort, the

world’s ISP operators could anoint a new root altogether. There was no law to stop them

from doing so. ISP operators had real power to change things, presuming they could figure

out a way to coordinate their actions.

But there were problems with this line of reasoning. If ISPs could aspire to lead the

name allocation regime, what about the numbers regime? Putting ISPs in charge of that as

well seemed like a dangerous prospect. It wasn’t hard to imagine that one or two leading ISPs

might end up having the strongest hand regulating the allocation of a resource that all ISPs

had to share. This concern was expressed most strongly by Alison Mankin, who argued that

it would be important to “decouple” the issue of IP numbers from domain names so that each

could be dealt with fairly.

The administrative connection between the names and numbers registries were really

just a vestige of history... a result of Postel’s longstanding technical involvement, as well as

the legacy of funding and operational ties that threaded through the DDN-NIC and InterNIC

years. The whole point of inventing hosts.txt and the DNS was to decouple names from

numbers at a network’s application-level. There was no technical reason to keep the two

235

IETF Role in DNS Evolution BOF (DNSEVOLV) December 5, 1995. No minutes were published,

331

but there was evidently a presentation, since lost, by Brian Carpenter.

Randy Bush, Brian Carpenter Jon Postel “Delegation of International Top Level Domains (iTLDs)”

332

Originally at ftp://rg.net/pub/dnsind/relevant/draft-ymbk-itld-admin-00.txt .Forwarded to the Newdom archives
by Matt Marnell as, “ITLD Draft,” Newdom-iiia January 23, 1996.

technologies joined in a bureaucracy. The pragmatic virtue of unraveling them and pursuing

separate solutions was starting to exert an appeal.

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->