A Proposal to Introduce Market-Based
Principles into Domain Name Governance
We
believe that the current Domain Name System (DNS) – the system that enables persons
to use easy-to-remember, common language names instead of numerical Internet Protocol
(IP) addresses to locate other computers and users on the Internet – suffers from
an artificial scarcity of names that is detrimental to Internet users worldwide.
The current system, administered by the Internet Corporation for Assigned Names and
Numbers (ICANN), is one that hampers the release of top level domains (TLDs) and
is the product of a legacy, consensus-based system of governance that inherently
cannot serve the diverse and large groups that have varying and even diametrically
opposed stakes in how today’s Internet is operated. As a possible solution,
we propose a hybrid consensus/market-based system in which the technical aspects
of the DNS are run by consensus through a central organization such as ICANN, but
the political and economic aspects of the DNS – those involved in choosing which
TLDs to use and who will operate them – are best served by companies competing in
an open marketplace. (For those who are new to issues relating to the DNS,
we present as background a brief history of the DNS in the attached appendix.)
In
understanding the DNS, one must keep in mind the difference between naming and addressing.
As Dr. Jon Postel, who had coordinated different Internet protocols including the
assignation of names and numbers in the DNS, stated with admirable clarity, “A name
indicates what we seek. An address indicates where it is.” In other
words, naming allows one to find a given computer more easily; addressing refers
to the way in which a computer is identified.
This distinction is not academic:
there is a clear difference between (1) the decisions regarding addresses – which
relate to the way in which machines function on the Internet including the assignment
of IP addresses, the establishment of Internet protocols and the manner in which
names are mapped to addresses; and (2) decisions regarding names – which relate to
exactly what names should be used by humans to locate the machines within the system
as well as who should administer those names. One set of decisions – addressing
and its attendant issues – is technical; the other – naming – is political and economic.
For example, why must a TLD be “.COM” or “.INFO”? The answer is simple: it
need not. There is no technical reason for that choice. As Paul
Vixie, the author of the DNS server software BIND, has stated, "A million names under
“.” isn't fundamentally harder to write code or operate computers for than are a
million names under "COM"."
Currently, however, one organization administers both
the addressing and the naming space of the DNS, and as such it must try to balance
all three areas – technology, politics and economics – in its work. In so doing,
ICANN has attempted to preserve the consensus-based decision-making model out of
which the technical parameters of the Internet grew. Along the way, ICANN has
faced criticisms and questions regarding its administration, decision-making procedures,
rules, and even its legitimacy. It is no surprise that any such administrative
body would encounter these issues when trying to address political and economic matters,
and especially so when trying to apply a consensus-based decision process to such
matters.
Indeed, those who originally laid out the parameters for moving
the control of the DNS out of the U.S. Department of Commerce were aware of the benefits
of a competitive marketplace. Specifically, in setting forth the “Principles
for a New System” (a key section of the Statement of Policy entitled “Management
of Internet Domain Names and Addresses” -- commonly known as the “White Paper”),
the Department of Commerce stated:
Competition. The Internet succeeds in
great measure because it is a decentralized system that encourages innovation and
maximizes individual freedom. Where possible, market mechanisms that support competition
and consumer choice should drive the management of the Internet because they will
lower costs, promote innovation, encourage diversity, and enhance user choice and
satisfaction.
In keeping with the White Paper’s principle of competition, we propose
that a market-based approach in conjunction with a consensus approach will allow
the DNS to achieve high efficiency and broad representation. This result is
possible because such a combination allows the technical aspects of DNS to be separated
from the political and economic questions concerning the creation of new TLDs.
Under such a system, technical matters would be decided using a consensus-based decision-making
process, and political and economic matters would be determined by market forces.
Accordingly, under this proposal, consensus- and market-driven decision-making processes
are used where they are best suited rather than forcing one into the other’s realm.
This hybrid approach will allow the DNS to serve best the group with the most at
stake: Internet users.
A History of Consensus-Based Decision-making
Since
the early days of the ARPANET, most questions of Internet architecture have been
resolved using a consensus-based system. Indeed, the very nature of the Internet
allows a consensus-based system to work extremely effectively in situations where
technical issues can be carefully considered by knowledgeable parties across the
world. The evolution of the RFC (request for comments) process – which allows
individuals or groups to publish technical proposals for the rest of the Internet
community to comment and build upon – has been and continues to be an essential component
in resolving technical issues quickly and efficiently.
It is also clear that consensus
works well in certain situations but is unwieldy -- if not impossible -- as the numbers
of persons engaging in the consensus process grow large. Indeed, a consensus
process loses its effectiveness as it tries to function beyond a fairly homogenous
group. In the case of domain names today where a heterogeneous group
seeks to govern technical, political, and economic matters, consensus administration
becomes unworkable and often produces undesirable results. Accordingly,
it may be that the natural limits of consensus-based decision-making prevent it from
achieving an efficient and broadly representative result.
Appreciating the limits
of consensus-based governance may be difficult for many involved in the continuing
administration of the Internet. It involves reassessing, with an open mind,
the state of the success or otherwise of a consensus-based approach to Internet naming
as a whole. We assert that the evolution of the Internet has resulted in such
a wide diversity of parties having an interest in the naming space that the current
approach, applying a consensus-based process to all aspects of naming, is no longer
the best way to ensure maximum efficiency and consumer benefit.
From early 1982,
decisions as to how the domain name system would work (and the attendant modifications
to the technical aspects of the system) used the RFC process, as did many other aspects
of the Internet’s operation. This process issues standards, informational
pieces and commentary. The standards do not create a law per se in that someone
can choose to operate outside the standards, but deviating from the standards obviously
makes it harder to work with those who have adopted them. Insofar as
changes to these technical matters are required, the tested method of the consensus-generating
RFC makes sense and is desired. This process brings the benefits of the appropriate
group, the technical community – which is focused on and passionate about making
the Internet operate well from a technical standpoint – vetting protocols and giving
each other input. Technical innovation is positively encouraged rather than
stifled.
Limits of Consensus
In contrast, we and many other groups feel
that the current naming process falls short of such interaction and constructive
procedures. Regardless of the various actions that brought about ICANN as the
current body governing the name space, suffice it to say that no single body could
use the process described above for TLD naming issues today. The TLD name space
is not comprised of a small independent group as was the case with Internet Assigned
Numbers Authority (IANA) with its rules and procedures regarding adding extensions
that meet a minimum specification threshold, or the Internet Engineering Task
Force (IETF) consisting of engineers developing protocols via occasional meetings
and email discussion groups to discuss and evaluate those protocols.
Rather, the TLD name space consists of numerous, disparate interests, and thus any
governing body must always attempt to serve many masters at once. In so doing,
it attempts to resolve political and economic issues relating to which TLDs should
be created, how they should be run, who should reap the economic benefits of running
them, and so on.
Any organization addressing political and economic matters
runs into questions of legitimacy and related questions of representation and due
process. As such, it is easy to understand why ICANN, whose regulatory or commercial
nature has been debated, faces numerous questions regarding legitimacy, fairness,
undue influence and accountability, to name a few. Indeed, it is no surprise
that ICANN, in an attempt to get anything done at all in the non-technical policy
arena, has, according to its critics and even neutral observers, chosen to marginalize
many of its constituencies and heed the counsel of a relatively like-minded, more
homogeneous subset of affected groups. Whether well founded or not, all of
these issues and questions surrounding ICANN further hamper the process of creating
new generic TLDs. In fact, many of these questions need not have arisen and
will actually be ameliorated by opening up the naming space to competition and moving
away from a position of artificial scarcity of names.
The Market/Consensus Approach
Something
more than consensus-based decision-making is required when addressing the non-technical
aspects of the name space and trying to expand it. We believe that the
market can serve that purpose. In a market-based name space, while ICANN keeps
guard over the manner in which the Internet evolves on a purely technical level,
the market invites individuals and companies to innovate regarding how those technical
gifts are used. To be clear, anyone trying to innovate would by necessity have
to comply with the technical parameters for the portion of the Internet in which
they operate in order to achieve commercial acceptance.
There are numerous
examples of innovation in the name space that have occurred without ICANN’s official
sanction, but which have been widely embraced by Internet users and provide significant
consumer benefit. These include: (1) New.net’s introduction of domain names
with more descriptive and useful TLDs that are accessible by users that choose to
support New.net, (2) VeriSign’s sale of multilingual domain names that require use
of a client application to enable resolution, (3) AOL’s use of “keywords” that are
accessible only by AOL users, and (4) RealNames’ offering of key words that can be
resolved by many, but not all Internet users.
By having ICANN focus on its
core technical competency, Internet technology can continue to use consensus to keep
it technically vibrant and stable. The market, in turn, will drive private
companies to address the name space in a manner that is both efficient and responsive
to Internet users’ needs, thus keeping the name space vibrant.
The market
side of the equation allows groups to find new ways to work within the current DNS,
thus providing consumers with more options. As with any innovation, adoption
by consumers will drive growth of the product and related commerce. Insofar
as those options are limiting on how the consumer interacts on the Internet, any
innovator must inform the consumer of that issue and demonstrate its offerings’ value
if there will ever be widespread adoption. Consumers thus can make the choice
as to whether they wish to be part of a certain naming system within the Internet
or not.
Perhaps the simplest analogy is to the cable television industry in the
United States. For some time, a home cable subscriber was beholden to
his or her cable operator. Programming was broader than with broadcast television,
but one could not deviate from the set of channels offered by the cable operator.
As new cable channels launched, the economics of the marketplace, either through
direct incentives or customer demand, allowed the new channels to gain better carriage.
Cable channels often bought advertising educating potential viewers about the benefits
of their programming, thus encouraging viewers to call their cable operator and ask
that the channel be part of the line-up. The most famous of these advertisements
may be MTV’s “I Want My MTV” campaign. In addition, competition introduced
through the emergence of satellite television further accelerated the cable television
industry’s efforts to be more responsive to its user market.
Similarly, companies
with new TLD or “alternative naming” strategies can use incentives to persuade ISPs
to “turn them on” and reach Internet users. In addition, Internet users can
ask that their ISPs enable their domain name servers to recognize New.net’s or others’
TLDs. The ISPs will be able to choose whether to do so based on the incentives
in place and the demands of their customers. Yet unlike the cable situation,
users can switch ISPs fairly easily (admittedly with some potential switching costs
such as changing email addresses) to get access to domain names that the users want
to access. Users also are empowered by the availability of software that
enables them to use alternative domain names if they happen to use an ISP that chooses
not to provide the user with access to such domain names. Accordingly, users
are given a large amount of freedom of choice and control over how they wish to use
the Internet.
ICANN’s current insistence on a constrained set of TLDs is
analogous to a user being locked into a single cable operator that decides that it
alone – absent any pressures from economic forces or consumer demand – should choose
what channels the user can view, claiming that too many choices would be confusing
to consumers or break the delivery system. It’s clear that the facts don’t
support the latter claim, and we believe that the former is overly paternalistic
at the least.
The benefits of a market-based approach are clear. By allowing
companies to develop new ways of working within the DNS technical system, to raise
capital, to market their products, and to do everything in their power to serve their
users, ICANN and Internet users will benefit in two ways. First, ICANN will
be able to conserve its resources and focus on developing better technical standards
to enhance the DNS. Second, if a company is trying to serve customers without
the shield of the virtual monopoly of being an ICANN registry, it will by necessity
be more responsive to serving its users to gain acceptance.
In a market-based/consensus
name space, ICANN would no longer use its test-bed procedure and no longer need to
issue new TLDs. Innovators could introduce their TLDs and develop them to the
best of their abilities. Though inclusion in ICANN’s or another root is not
necessary, once an innovator achieved success with its TLD, it ought to be included
in the root servers controlled by the U.S. Department of Commerce as a matter of
course. A more broadly representative ICANN could assist in setting an objective
standard for inclusion in the root servers, which might include minimal technical
operating standards and a minimum number of domain names being used by disparate
users. Once the standard is established, there would be little room left for
discussion (and thus politics), and the process would be essentially an administrative
one. ICANN could return to spending more of its time and resources on setting
appropriate technical standards, fulfilling its originally intended role.
In other
words, by allowing companies to rely on their own resources to develop new ways of
using the DNS, the root system can essentially deploy an ongoing test-bed with little
political or economic cost to the root system. Internet users would gain by
having innovators strive to find new ways in which to serve Internet users using
the DNS while complying with the technical standards set by ICANN and other Internet
engineering bodies. As innovators succeed, Internet utility is increased, and
users gain from broader ways in which they can use the Internet. After a company
has proven itself, ICANN could bring that company into its system and thus reward
the innovator with lower costs for further adoption on the Internet. In addition,
ICANN would be liberated from the financial and political drains associated with
non-technical policy matters and that raise questions regarding its legitimacy.
Competition Will Enhance the Internet, Not Harm It
It is our belief that the
benefits of innovation in the name space outweigh any perceived costs. As discussed
above, innovators within the name space include those issuing new TLDs as well as
those seeking to use DNS functionality to expand the way in which we find information
through browsers or other means. When suggesting that innovation or competition
in the name space or the deployment of new roots be allowed, the cost most often
offered as the reason not to do so is that those actions will somehow “break the
Internet.” This argument is sometimes called splitting the root.
At bottom, the argument equates universal resolvability with stability, but the two
are not inherently the same. Put simply, one can have stability without universal
resolvability. A system can be stable in that it works for those within that
system. The system may not be universally resolvable until it performs well
enough that it gets accepted as a standard, but it is still stable.
Enabling more
competition is not likely to result in many competing companies releasing numerous
conflicting top level domains. A competitor choosing a TLD that conflicts with a
widely registered TLD would have to commit resources to an economic battle that it
would most likely lose, or win only at a high price, rather than choosing a TLD with
few or no collisions. Though choosing a conflicting TLD is a possibility, assuming
that Company A has achieved some success in establishing a user base with ISPs and
software distribution, Company B would be hard pressed to convince ISPs, users, and
the market in general that its offering of the same TLD as Company A should be recognized
or is worthwhile. Faced with such a decision, rational market players would
choose developing new TLDs. That decision in turn, would have the result of
further opening up the name space and producing further consumer choice, thus enhancing
the Internet rather than harming it.
Another way of looking at the issue of competitive
systems is to consider the current situation regarding operating systems and Internet
browsers. If all users used Microsoft Internet Explorer, then Web design would
be easier and less expensive. Yet, we are all better off with choices.
Even in the case of having only a few options, consumers get some ability to choose
between competitors. In the case of browsers, we have gone from Viola and Midas
to NCSA Mosaic to Netscape Navigator and Internet Explorer. Along the way,
certain browsers supported some features such as the use of frames while not supporting
the Marquee, or scrolling function, and vice versa. Today, however, to be competitive,
browsers tend to support more rather than less functionality. The presence
of competition and consumer choice pushed the browsers to provide functionality they
might not have otherwise supported. Accordingly, despite dominance by one player,
the existence of competitors forces the dominant player to maintain a higher level
of service and utility than if it were the only option. Of course, should one
player or a small group of players get so dominant as to be anti-competitive, antitrust
laws would act to limit such control.
To summarize, competition within the name
space will enhance the Internet and does not threaten the existence of a root system.
At worst, competition may threaten the existence of the current means of governing
the name space in what is currently the dominant root. Yet, even that possibility
is remote – to have the chance to be successful within that name space, a company
must adhere to the technical rules of the name space and avoid conflicting with other
widely used TLDs. At best, by encouraging more players to develop either new
ways to use the current system or encouraging them to develop entirely new functionalities
based on the current system, competition and innovation in the name space only threaten
to put the Internet back into a creative, user-oriented posture. In the end,
by letting the consensus and market approaches each thrive in the arena in which
it is most effective, we open the way to a more vibrant Internet that delivers more
benefits to a ready and willing world.
APPENDIX 1: Background
This section
provides a brief overview of the development of the Domain Name System (DNS).
1971-1994 - IP addressing and the development of the DNS.
In the early days
of the Internet, relatively few computers were part of the network that comprised
the Internet. As such, the administration of how computers were identified
and found on the network was a fairly straightforward process. Computers on
the network were given an identifying number called an Internet protocol (IP) address.
If a user knew the address, he or she could simply use that address to contact that
computer. Early users of the Internet typically would look up the IP address
for a particular computer from a white pages style list that was stored on every
computer. This process worked well in the early days of the Internet,
but became overly cumbersome as the numbers of computers on the network grew significantly.
Just keeping the list up to date, let alone finding the computer within the list,
became quite difficult.
The Internet started as a scientific project and was managed
by the people using it the most: scientists and academics. This group used
the Request For Comment (RFC) process to circulate proposals, comment on them and
finally issue a standard as to how a certain portion of the Internet should function.
As a response to the problem of locating computers on the Internet, a group of scientists
at the Institute for Scientific Information (ISI), including Dr. Jon Postel, used
the RFC process to develop the "domain name system" (DNS). The domain
name system was designed to use a hierarchical database structure, which enabled
different people on different computers to manage different parts of the naming hierarchy.
The initial top level domains (TLDs) included seven generic TLDs, designed to identify
the type of host, such as .com for commercial organizations, .net for network providers,
.org for not-for profit organizations and .mil for the military root domain name
servers. In addition, there were two-letter country code top level domains
such as .jp, .us, and .fr to identify geographical locations.
1994-1998 – the DNS
gains in commercial value; Postel and others advocate the introduction of competitive
forces into the running of the addressing and naming space.
From around 1994, the
combination of the introduction of the World Wide Web and a useful Internet browser,
Mosaic, fueled recognition of the commercial value of the Domain Name System as the
potential of the Internet became clear to a community increasingly made up of commercially-minded
players. At this point, the National Science Foundation (NSF) who had taken
over funding the Network Information Center, entered into a cooperative agreement
with Network Solutions, Inc. (NSI). Under the agreement, NSI took over
the registration services previously run by the Stanford Research Institute.
This meant that NSI ended up registering second-level domains in .com, .net, .org,
and .edu and administrating the main root server, or the "A" root server.
Policy authority however, remained with Postel and the Internet Assigned Number Authority
(IANA).
By 1995, many more people around the world outside of the scientific community
were using the Internet and the World Wide Web for an increasing amount of non-research
activities. Domain name registrations were correspondingly increasing.
In response to this change, the NSF decided to no longer pay for registrations and
executed an amendment to the cooperative agreement with NSI that allowed NSI to charge
a $50 annual fee to each domain name registrant. This change was a fundamental
shift in the operation of the Internet addressing system. Rather than only
a group of scientists interested in and running the DNS for researchers and scientists,
an outside corporation with a distinct financial interest became involved and now
served an international community using the Internet for a wide variety of non-research
purposes.
Consequently, there was growing unhappiness in certain sectors
with the new fees being charged by NSI and the structure of the DNS. Some wondered
why they should be stuck with the service provided by and fees charged by NSI when
registering domain names with generic top level domains. NSI also generated
animosity with its domain name dispute policies, under which it asserted the right
to (and did) suspend any domain name upon receiving a complaint from a registered
trademark holder, without regard to whether the registered trademark holder had a
superior legal claim to the domain name.
This same period saw the beginning
of a scarcity of easy-to-remember domain names as domain name sales increased dramatically.
At the same time, Postel and others in the technical community began to agree that
many more TLDs were technically possible and needed to address increasing demand.
Indeed, in 1996, Postel suggested that IANA authorize up to 150 new generic top level
domains to be operated by new registries. Postel’s view was clear: “…positive
market forces dictate that diversity [in the top level domain space], obtained through
free competition, is the best means available to insure quality service to end-users
and customers.”
As Postel’s proposal regarding adding new TLDs was developed,
IANA and the Internet Society created the "Internet Ad Hoc Committee" (IAHC) to consider
the question of adding new top level domains. IAHC members included representatives
from several international organizations representing corporate interests, including
the International Telecommunications Union, the International Trademark Association
(INTA), and the World Intellectual Property Organization, which marked the introduction
of corporate influence into DNS policy-making. In one example of this corporate
influence, the INTA representative, worried about the effect of new TLDs on trademark
interests, argued that the number of new domains be limited rather than broadly expanded
as Postel and others had suggested. The trademark lobby was successful,
and the IAHC proposed that only seven new top level domains be added as an initial
matter.
At around the same time, the U.S. government established a working
group that included representatives from numerous government offices, including the
National Telecommunications and Information Administration (NTIA), the Patent and
Trademark Office, and the NSF, among others, to determine what should be done with
the administration of the burgeoning Internet naming and addressing space.
As a result of that working group, in July 1997, the NTIA issued a request for comments
addressing the best way in which to govern the Internet and the DNS.
At the same time as it developed the request for comment, members of the working
group began negotiating with Postel regarding turning IANA into a more structured,
corporate body with greater accountability to the international Internet community.
1998 to the present day – The U.S. Government and the Creation of ICANN.
In
January 1998, the U.S. Government released a paper entitled “A Proposal to Improve
Technical Management of Internet Names and Addresses,” which became known as
the “Green Paper.” By the time the Green Paper was issued, the IAHC process
had stalled, having run into resistance from NSI regarding adding new TLDs to the
“A” root server without express approval from the U.S. Government, which the U.S.
Government had not provided. As such, the Green Paper made no reference to
the IAHC process.
The Green Paper proposed the creation of a new not-for-profit
corporation, "operat[ing] as a private entity for the benefit of the Internet as
a whole," to administer the DNS. To allow this new corporation to have such
control, the Green Paper proposed that the IANA staff would be folded into the new
organization with the U.S. Government handing over existing IANA functions, the root
system, and the appropriate databases to the new corporation and "participat[ing]
in policy oversight to assure stability" for up to two years. The Green Paper made
clear that the new organization and its board "must derive legitimacy from the participation
of key stakeholders," envisioning a continued process of consensus as the best means
to run both the addressing and the naming space. To achieve this goal, the new organization’s
board was to consist of representatives from various membership organizations relating
to the technical aspects of the DNS, such as IP addresses and Internet technical
standards, as well as representatives of "the direct interests of Internet users"—including
individual, corporate, and non-profit interests.
Four months later, after extensive
consensus-based commentary on the Green Paper, the U.S. Government issued its “White
Paper.” In a broad non-specific document, the Department of Commerce acknowledged
the value of introducing competition into the name space: “The pressure of competition
is likely to be the most effective means of discouraging registries from acting monopolistically.”
Yet the White Paper did not give exact details regarding how the new corporation
would be formed. It offered the U.S. Government’s support for a new organization
that was created by "private sector Internet stakeholders" in the form of contracting
with it, advocating for it internationally, and guaranteeing that NSI would give
the necessary access to its databases and software. The actual creation of
the corporation was left an open issue.
Shortly after the U.S. Government’s release
of the White Paper, the International Forum on the White Paper (“IFWP”) formed and
met often to discuss the way in which this new corporation should be run.
Indeed the U.S. Government’s key policy adviser on domain names, Ira Magaziner, attended
two meetings and “ ‘blessed’ the IFWP process.” At the same time as the
IFWP was meeting and discussing how to implement the mandate of the White Paper,
IANA, the Internet Society (which comprises the Internet Engineering Task Force (IETF)
and the Internet Architecture Board (IAB) among other groups), ISOC, Postel, attorney
Joe Sims, IBM, Magaziner, and some foreign governments met and drafted their own
plan for the new corporation. Following those discussions, Postel sent
the Department of Commerce the articles of incorporation for a newly incorporated
company, the Internet Corporation for Assigned Names and Numbers (ICANN), as well
as biographies of the board of directors and bylaws for the company. Postel
described the materials as representing "the consensus judgment of the global Internet
community as to how to form a corporation that will include the IANA function."
Postel died two weeks after delivering the corporate documents to the Department
of Commerce. After reviewing three proposals regarding the new corporation,
the Department of Commerce accepted the ICANN proposal despite controversy regarding
which proposal was the best.
The White Paper called for the new corporation
to have “the functional and geographic diversity of the Internet and its users.”
To meet this mandate, ICANN currently has a 19-member board that operates with the
help and advice of three supporting organizations, the Address Supporting Organization,
Domain Name Supporting Organization and Protocol Supporting Organization and the
At-Large Membership. The organizations are in turn made up of constituencies
ranging from business to intellectual property to addressing groups such as the American
Registry for Internet Numbers to the IETF. Individuals are represented
through the At-Large Membership. All four groups get seats on the ICANN board
with the supporting organizations getting three seats each and the At-Large Membership
getting five seats. The organizations establish working groups to address questions
regarding the DNS by gathering information and then making recommendations to the
board.
Despite ICANN’s efforts to be representative, the structure has met
with some criticism that ICANN’s structure is too complex; its operations are not
truly representative; its decision process does not work; and that ICANN itself is
illegitimate. The paper to which this history is appended presents a
solution that would address many of these criticisms.