In 1992, the National Science Foundation (NSF), through a competitive
bid, chose NSI to manage and enhance the domain name registration
system. Just two years ago, this small effort required registering only
200 domain names per month onto the Internet. Today the volume has
reached more than 120,000 names per month, with a total of approximately
1.6 million registrations to date.
The current domain name system is relatively easy to use,
inexpensive, robust in the face of growth and financially
self-sufficient. However, many of its critical functions rely on systems
run by volunteers with inadequate legal guidance or protection, its
directories are not centralized, registration policies vary and it is
too vulnerable to technical attack and "name speculation."
International growth in domain names requires greater coordination and
global solutions to these problems.
The growth of the Internet, the development of electronic commerce,
and the pressures on the existing domain name system are factors
signaling the need to transition to the next generation of domain name
services. To plan and implement a successful transition, the Internet
community and the government must work together to address issues that
affect the stability and reliability of the domain name system,
specifically in the areas of operation, management and coordination.
Certain basic administrative functions, such as IP address allocation
and distributed root server management that are part of the Internet’s
historical legacy, are under increasing stress and need to be changed.
A successful transition must also promote continued private
investment. Developing a stable and commercially viable domain
registration system required an investment of millions of dollars by
NSI. The conversion to user fees was critical because it allowed NSI to
keep pace with the annual tripling of demand in domain registrations
over the last two years, while at the same time improving services and
reducing turnaround time for processing applications. Through automated
and improved processes, NSI has reduced processing time from
approximately four weeks to 24 hours. This progress demonstrates how
private investment and commercial incentives can produce improved
services.
Added competition can become the motivator to increased investment,
lower prices and enhanced domain name services. But competition must be
increased in a way that strengthens, rather than threatens, the
stability and reliability of the underlying system. There should be
criteria governing how competitors around the world provide domain name
services -- whether they do so by providing new top level domains or by
acting as new registrars within existing domains.
However, certain essential functions – such as the provision of IP
addresses and management of the distributed root servers – should be
managed in a neutral, centralized fashion that guarantees access to all
Internet users. Similarly, while we should promote competition among
domain name registrars, a designated party should manage a
central registry. The transition to such a central registry and
its proper management structure must occur in an orderly way to protect
the system from confusion and conflicts that would undermine the current
domain name system.
Successful competition requires the proper legal framework to oversee
functions critical to the operation of the Internet and the domain name
system. This legal framework should encompass both domestic and
international issues related to providing competitive domain name and
associated Internet services. In addition, there is a need to clarify
the relationship between the domain name system and existing legal
frameworks -- such as trademark law – and to establish clear and
consistent working principles to govern conflicts and liabilities.
The growth of the Internet and the development of electronic commerce
depend upon a stable and reliable system for domain name services. The
transition to a domain name system as robust as the Internet itself
requires private investment and a global legal framework that promotes
competition and protects the critical functions. NSI supports these
goals to provide a domain name system that will continue the growth and
success of the Internet.
We welcome the opportunity to comment on these important issues and
respectfully offer the following response to the Department of Commerce
Notice of Inquiry.
Competition in and expansion of the
domain name registration system should be encouraged. Conflicting
domains, systems, and registries should not be permitted to jeopardize
the interoperation of the Internet, however. The addressing scheme
should not prevent any user from connecting to any other site.
NSI response to a:
NSI welcomes additional competitors to the business of providing
domain name registration services. NSI believes additional competitors
will lead to new and creative ancillary services, and wider consumer
choice. Competition will induce greater worldwide investment, which is
critical to the growth of the emerging registration services market.
Consumers will benefit through more aggressive pricing and improved
services. As the registration services leader, NSI believes it will
benefit as greater investment will enhance the worldwide market.
NSI supports the introduction of new competition in a manner that
safeguards the stability of the Internet. Whatever form competition
takes, its introduction must be phased-in in a way that protects the
current Internet users. It must be accompanied by policies, procedures
and (technical and legal) safeguards that do not jeopardize the
security, reliability and needs of users on the Internet. To do
otherwise will generate a crisis of confidence in the Internet.
The form of competition that is chosen must preserve the physical
connectivity of the Internet by avoiding balkanization of the
root system. The power of the Internet lies in its total connectivity
and in the use of TCP/IP, a single open protocol.
NSI and other Internet service companies are making the Internet
easier for consumers to understand and to use. Ease of use is a
significant issue for the majority of new users, who need the Internet
to achieve their communication objectives. Proposed changes must
encourage this trend and must avoid increased consumer confusion.
The private sector, with input from governments,
should develop stable, consensus-based, self-governing mechanisms for
domain name registration and management that adequately define
responsibilities and maintain accountability.
NSI Response to b:
Government must play the stabilizing role now, providing a legal
anchor for critical elements of the Internet through a period of
transition to a more mature form of governance. The "rough
consensus" decision mechanism that served the Internet well in the
past is inadequate to deal with the complexity and conflict that
commerce has brought. Internet stakeholders with differing interests
need a better mechanism with legal force to resolve their conflicts.
Whatever the governance mechanism adopted, NSI believes that the
marketplace and commercial interests should ultimately drive domain name
registration. NSI’s experience has shown that the domain registration
services business is investment intensive. Structuring registration
services as a competitive for-profit industry is the only way to justify
the needed investment. In this approach, customer demand creates the
built-in incentive to generate improvements in service, the development
of additional services and the most creative solutions to evolving
problems.
NSI believes that domain name registration is not a "public
trust." The notion that domain names should be managed as a public
trust is based on the misconception that second-level domain names are a
limited commodity. This claim is unfounded; the number of second-level
domain names (or domain names at any level) is virtually limitless, and
the parallels made to radio spectrum are incorrect. The spectrum
comparison is correctly drawn to the 32-bit IP numerical addresses on
the Internet. They are very limited and are in increasingly short
supply.
These self-governance mechanisms
should recognize the inherently global nature of the Internet and be
able to evolve as necessary over time.
NSI Response to c:
Eventually, the administrative functions of the Internet should be
anchored in some international legal authority to ensure representation
from the global community. NSI believes that an appropriate sponsoring
organization has not yet been identified. To allow the time to either
select a suitable organization, or to establish the charter for creating
a new sponsoring organization, we propose that a U.S. Government agency
act as an interim sponsoring organization. Although the Internet has
become a global communication medium, the U.S. Government has been
underwriting its development both technically and financially for more
than 20 years, and the administrative infrastructure of the Internet is
still largely U.S. based. [See Chart One]. Therefore, leadership from
the U.S. Government in transitioning the Internet to the next phase of
its existence is appropriate, and in fact may be necessary. The reality
is that no other entity may have the political strength to generate
consensus in this diverse community.
The overall framework for accommodating
competition should be open, robust, efficient, and fair.
NSI Response to d:
NSI strongly supports a competitive model that is open, robust,
efficient and fair. This implies that it should encompass all top-level
domains (TLDs) and all registrars. A few historical observations will
indicate why NSI supports an open framework.
After two years of government supported registration activity, NSI’s
mandate from the NSF was to develop a process that would allow
registrations to become market supported. The most critical step toward
that goal-and a test of the market-was to initiate user fees for DNS
registration, moving away from government funding. Although NSI suffered
two years of negative profitability due to expansion requirements needed
to meet the huge demand, this effort has proved to be successful. NSI’s
and NSF’s joint action has removed the need to publicly subsidize the
registration of private entities on the Internet, and equally
importantly has generated a source of funds for reinvestment in improved
systems and expanded service. It is unlikely that the DNS registration
activity would have survived the growth demands it experienced had it
remained supported by capped government funding. Open and robust
competition will only further this progress made by the NSF and NSI.
In addition, the fee-for-service approach pioneered by NSI and NSF
has been widely imitated in many of the other registries, and has even
spawned a rich cottage industry consisting of dozens of businesses
"reselling" domain names along with other competitive
services. The competitive framework should not impede the development of
enhanced registration services by existing or potential registrars, and
should not discourage the entrepreneurial spirit demonstrated by this
cottage industry.
In an effort to promote enhanced competition, NSI formally proposed
to the NSF on June 10, 1997, that additional TLDs and new registrars be
allowed on the Internet root server system under the terms of the
NSF/NSI cooperative agreement. NSF declined this request pending a more
comprehensive review of the system.
The overall policy framework as name allocation and
management mechanisms should promote prompt, fair, and efficient
resolution of conflicts, including conflicts over proprietary
rights.
NSI Response to e:
NSI supports the development of an overall policy framework to
resolve DNS conflicts that reflects the current realities. Any
management mechanisms that are implemented must not burden the
registration process with unnecessary costs or delays. The number of
disputes is a small fraction of the number of registrations. Adopting
procedures that delay or add costs to the vast majority of trouble-free
registrations to accommodate the few exceptional disputes is
unreasonable. The NSI dispute policy, which provides for first-come,
first-served registration of domain names, also recognizes that
trademark owners have rights found in federal law. The policy remains
the best mechanism available in today’s environment.
There is no legal precedent that deals with the international nature
of DNS disputes. The global nature of the Internet frustrates federal
trademarks because of the vacuum in international law. The resolution of
disputes requires a body of international law or, at least, civil
procedure, that is specifically applicable to domain names and
trademarks. Barring that, issues will be dealt with through negotiation
between disputing parties resulting in highly public, but relatively
rare, court suits. Because of its registration volume, NSI has been the
first registrar to experience these conflicts. Growth in the Internet
will eventually bring this problem to all registrars in the hundreds of
other Internet TLDs.
Increasing competition by adding new TLDs and new registrars will
increase the challenges associated with dispute resolutions. If policies
and procedures are not established first, the chances are high that
registration services will not be prompt, fair or efficient.
Care must be taken to avoid instituting a dispute mechanism that
retards the speed of registration. Consumers will not tolerate
significant delays or additional cost to prescreen for trademark
disputes.
A framework should be adopted as quickly as
prudent consideration of these
issues permits.
NSI Response to f:
Although the Internet is in transition, it is not in crisis. The
vacuum in governance and the legal vulnerability felt by the major
players has generated a climate of confusion. The real problem is that
no current mechanism exists through which the Internet can change and
evolve in the way that is required. This is the key issue that the
Department of Commerce should address.
As a corollary, stable procedures are in place to maintain the daily
operational requirements of the Internet. The Internet domain name
system is more resilient than its critics acknowledge. Some elements of
the DNS lack accountability, and it is suffering under some stress, but
its continuing exponential growth and the emergence of new services is
evidence of its overall health. As an example, NSI has registered over
1.6 million domain names to date. In two years registration has risen
from less than 200 names per month to 120,000 per month. NSI is
averaging 92 percent of applications registered and on the Internet
within 24 hours. Since 1988, NSI has faithfully updated the Internet
global root servers nightly, reflecting the world’s changes for the
preceding twenty-four hours. Daily we carefully allocate scarce IP
addresses using long established guidelines to the thousands of ISPs in
the Americas.
The facts outlined above are a clear indication that the DNS is
capable of meeting the daily operational requirements of the Internet.
While NSI recognizes that the DNS needs to come under a more mature form
of governance, there is no need to implement hasty solutions. Time
should be spent to ensure a good outcome for the Internet.
What are the advantages and disadvantages
of current domain name registration systems?
NSI Response to 1:
The current system has managed to sustain the recent growth imposed
on it. The introduction of widespread commercial interests and demands
placed on the DNS, however, have revealed unforeseen deficiencies.
Network Solutions, based on its experience, cautions that any changes
need to be carefully tested before they are implemented. Below NSI has
identified some advantages and disadvantages of the current system that
should be considered:
Advantages of the DNS:
1. Ease of use. The DNS has simple user conventions, within a
simple structure, that are easy to understand. In the current DNS, the
existence of a single registration entity within each of the relatively
small number of TLDs (approximately 200) prevents confusion even for new
Internet users. For those consumers who desire a one-stop registration,
there are dozens of ISPs who will resell domain names in any TLDs for a
premium fee.
2. Large number of available names. Additional top level
domains may be useful to create new brands and encourage new services,
but mathematically will do little to increase the number of second-level
domain names available. There are virtually an infinite number of
second-level domain names available in each existing top-level domain.
In .com, for example, there are approximately 3722
possible combinations for second-level domain names.
To illustrate:
- 22 spaces available to the of the left of the dot
- 37 characters: English alphabet, plus 10 digits and the hyphen
- possibilities (3722) 31,700,000,000,000,000,000,000,000,000,000,000 *
*The exception to this is
that a valid domain name may not have a dash '-' as the first or last
character.
3. Financially self-sufficient. The domain name system has
nearly completed its evolution from government subsidy to financial
self-sufficiency. Based on recent OECD data, the number of registries
being supported by user funding is growing rapidly. There is direct
correlation between the volume of registrations and the likelihood that
user fees will be sufficient to support a registry.
[See Chart Two]
4. Resilient to growth. NSI has operated a registration system
which has demonstrated over the past three years that the DNS can
withstand tremendous growth. Simultaneously, registration
acknowledgments are now generally received in less than an hour, down
from the four weeks it sometimes took in 1994. DNS is a tested system
that has provided stability for the Internet through unparalleled
turbulence.
Disadvantages of the DNS:
As previously mentioned, commercial interests and the challenges that
come with commercialization such as name speculators, hackers, hostile
attacks, legal challenges and explosive traffic volume are adding stress
to the system. These stresses have helped to reveal the disadvantages
described below.
1. Vulnerability to spoofing. DNS as a whole is not a secure
system. It was designed to be a peer-to-peer trusted system and hence is
vulnerable to "spoofing," attacks whereby a machine misrepresents itself
to the system as being "authoritative" when in fact it is not.
As a result, various simple mechanisms have been designed and employed
by programmers to "hijack" an organization's DNS, which has the effect
of making its web-site, and possibly its e-mail, disappear.
2. Directory Inadequacy. No central DNS directory exists
today, and this is a tremendous problem. TLDs therefore have evolved
into de facto directories that allow users only a sense of where
to begin looking for people or things on the Internet. A comprehensive
directory would help alleviate the concern that the Internet is
unorganized and overwhelming for many users looking for information.
3. Software vulnerabilities. The primary server software that
is used to implement DNS is called BIND (Berkeley Internet Name Daemon).
BIND is currently developed and distributed by the Internet Software
Consortium, a 501(c) (3) non-profit organization. There is currently one
programmer working on this software, funding comes from corporate
contributions and software releases are not always timely enough to fix
known "security bugs." In the chain of organizations, machines
and software that together comprise the global DNS system, the BIND
software, along with the volunteer root operators, are the weakest links
in the chain.
4. Uncontrolled critical parameters. There are many technical
parameters, that can affect the smooth operation of DNS. These include
port numbers, protocol definitions and reference implementations of key
software. Changes to any of these can affect most, if not all, users of
the Internet. These changes are currently detailed and approved by
voluntary organizations and clubs which have no legal standing and
reflect differing levels of corporate sponsorship and participation.
Although this consensual mechanism has worked well in the past, it is
too easily subject to manipulation by parochial funding or
overrepresentation by a vocal splinter group. For the first time in the
history of the Internet, there is significant investment together with
corporate and national interests that are impacted by each and every
administrative and technical decision.
5. Multiple, separate whois databases. There is no
standardized mechanism for unified lookups between whois databases among
the registrars. For instance, a whois query against a domain name in the
.us TLD must be run against a whois server at ISI, and not the
most widely used one at NSI which resolves .com, .net.,
.org, .edu and .gov domains. Military domain names
must be looked up at the Defense Data Network Network Information Center
(DDN-NIC) run by the Defense Information System Agency (DISA). This
complication should be resolved in a manner transparent to the consumer.
The Referral whois (Rwhois) protocol, which NSI engineers in conjunction
with an Internet Engineering Task Force (IETF) working group have been
developing over the last four years, is able to accomplish this
delegation of lookups, but it is not yet widely used.
6. No uniform registration policies. All of the registration
authorities have different policies, procedures and regulations. As a
result, service levels and sophistication vary tremendously and may
deter user interest in commerce. Also, users with common problems cannot
find a common forum and jurisdiction for resolution.
7. Legal dispute vacuum. There is no uniform, global legal
system to deal with trademark-domain name disputes. The difficulty in
developing policies or a legal framework is that it involves managing
the complexities of trademark law as it intersects with a global
communication medium.
8. Registries are legally vulnerable. None of the registries
are immunized from legal suit. They are all vulnerable to legal
challenges and financial failure, which may have very serious
operational implications for the system and its users.
9. Critical facilities managed by volunteers. Roots servers
are administered through a volunteer network. Under the current system,
root server administrators have no performance contracts, legal
requirements or compensation provisions. One direct result is that these
critical Internet assets are not manned or monitored around the clock.
Generally there is no structured accountability in place.
10. Unscrupulous activities. The DNS is vulnerable to
speculators who clog registries with applications for names to
"resell." These name speculators create tremendous workloads
for the registry, increase costs to the clients and extort money from
business in exchange for the rights to a particular domain name.
How might current domain name systems be
improved?
NSI Response to 2:
NSI’s suggested improvements reflect the disadvantages discussed in
the last question:
Anchor DNS policy. To improve current systems, certain
operational and policy initiatives should be implemented first. The DNS
policy functions of the IANA should be transferred to a legal authority.
The responsibility of this role has greatly expanded and has become
inappropriate for an organization that does not have clear authority,
accountability or immunity from legal challenge.
Establish DNS management. The entire high-level management
structure of the DNS system needs to be anchored in a legally immunized,
user-funded, knowledgeable organization. The organization’s
responsibilities would include the root server system; maintenance of
BIND and other critical softwar; encryption technology which will be
needed for secure DNS; and any other technical decisions that can
potentially impact the successful global operation of DNS.
Secure root servers. The root server system volunteers should
be replaced by secure, contracted services operating under specified
service and performance terms, procedures and policies.
Encrypt DNS software. Encryption must be added to DNS to
improve its security, and widespread industry adoption should be
strongly encouraged. In the interim, attempts to misrepresent, misdirect
or deny DNS identification of an Internet organization or user should be
treated as a crime and prosecuted by the Justice Department.
Standardize TLD access. A standard, such as Rwhois, should be
agreed upon to ensure uniform access to all TLD databases. A basic
precept should be that a consumer should never be required to know the
physical details of domain or host administration to perform basic
navigation and identification.
Immunize registries. Recognize that registries are passive
suppliers of services; they should be immunized from domain name dispute
lawsuits that do not involve negligence. NSI’s vulnerability to lawsuits
has necessitated unpopular and costly policies and substantial legal
fees to defend and protect the registry. These costs and inconveniences
will eventually be passed on to consumers. Ultimately, what is needed is
a rule of law to deal with global trademark vs. domain name issues.
Regulate DNS speculators. Establish Internet commerce specific
regulations that prevent speculation in registering domain names.
By what entity, entities, or types of entities
should current domain name systems be administered? What should the
makeup of such an entity be?
NSI Response to 3:
The growing number of legal actions surrounding the issue of
legitimate, legal authority on the Internet is troubling. Who has the
authority, and from where does the authority derive? These questions are
being asked more frequently in knowledgeable Internet circles.
The administrative functions of the Internet need to be managed by a
body anchored in a legal authority that can assure the stability of the
Internet, oversee policy regulations and reflect the concerns of a
global community of users. Eventually both the anchoring legal authority
and the managing body should be international organizations reflecting
the global nature of the Internet. Representation within the managing
body should be proportional to Internet use, setting policy and handling
any disputes between entities. It is critically important that this
managing body represent all of those involved in Internet administration
in an open process.
NSI believes that an appropriate international anchoring organization
does not currently exist. NSI proposes that a U.S. Government agency
assume the legal responsibility during a transition period. Eventually,
NSI proposes that the U.S. Government sponsor an International Public
Advisory Group (IPAG) as a managing body for the Internet as it moves
towards a more fully competitive, global environment.
Although NSI is not advocating any particular agency, what is needed
is an agency with broad experience in the administration of complex
communications issues and working relationships in these matters with
other countries. This agency should have in-house technical expertise
and extensive experience in balancing regulatory and commercial matters.
The agency also needs mechanisms in place to accept business and
consumer input, through a series of industry driven committees.
Leadership by the U.S. government would be an enormous help in
developing the charter for the organization that can carry these
responsibilities forward on a global basis.
The following administrative functions would be the first charge of the IPAG:
1. Management of IP Registries
The IPAG should set policy and coordinate the activities of the IP
registries worldwide. Three new Regional IP Registries are planned to
supplement Reseaux IP Europeens Network Coordination Center (RIPE) and
Asia-Pacific Network Information Center (APNIC). These are non-profit
entities that work in concert to safeguard the increasingly scarce IP
address space in IPV4. The Regional IP Registries should continue to
manage IP address space and allocate it within their respective regions.
The current proposal to establish three new IP registries-the American
Registry for Internet Numbers (ARIN), the -African Network Information
Center (AfriNIC), and the Latin American and Carribean Network
Information Center (ALyCNIC) -should be accelerated. (As of this
writing, the NSF has approved the establishment of ARIN a non-profit
organization to allocate IP numbers in the U.S.) The remaining three new
registries are required to provide complete global coverage. IP
allocation fees from ISP members and users will support IP registries.
Each regional IP registry should perform the following functions:
1. Allocate IP address space to ISPs and others in its region;
2. Maintain all data relative to its regional IP allocations;
3. Manage and assign inverse addressing (IN-ADDR) and Autonomous
System Numbers (ASNs) within its region.
2. Management of the Global Domain Name System
The importance of the root server system to the functioning of the
DNS requires neutral, vigilant and stable management. An entity selected
competitively to manage the root server system would provide
accountability and stability. NSI proposes that the IPAG select a DNS
Manager (not a registrar) to manage and to maintain the root
server system. The responsibilities of the DNS Manager would
include:
1. Establish and update a set of root servers worldwide, which may be managed individually by other contractors to limit vulnerability;
2. Operate services pertaining only to data related to the root;
3. Maintain the data, updates and top-level directories; and
4. Implement any new TLDs that are authorized into the global DNS.
The DNS Manager would manage the activities of the TLD registrars in
accordance with the policy set by IPAG. The IPAG should address in the
interim transition period critical issues such as the form of additional
competition and mechanisms to resolve trademark vs. domain name
disputes. The TLD registries should support the IPAG and its activities
relative to the maintenance of the root server system.
[See Chart Three]
Are there decision-making processes that can
serve as models for deciding on domain name registration systems (e.g.,
network numbering plan, standard-setting processes, spectrum
allocation)?
Are there private/public sector administered models or regimes that can
be used for domain name registration (e.g., network numbering plan,
standard setting processes, or spectrum allocation processes)?
What is the proper role of national or international
governmental/non-governmental organizations, if any, in national and
international domain name registration systems?
NSI Response to 4:
The North American Numbering Council (NANC) in the telephone industry
is a reasonable example of a functioning International Public Advisory
Group (IPAG) comprising industry and government that provides a model
for an effective legal framework for Internet administrative functions.
Under this model, the group’s membership should be global with
representation based on Internet use.
Another example can be drawn from the striking parallels between DNS
and 800 number portability administration. In this case, a small amount
of government sponsorship acts as a significant and stabilizing anchor
for private sector administration of the system. One could adopt an
approach similar to the role of the FCC with respect to competitive 800
number allocation in the telephone market. FCC sponsorship has worked
well for the NANC and for other privately administered numbering
organizations and issues.
Should generic top-level domains (gTLDs),
(e.g., .com), be retired from circulation? Should geographic or country
codes (e.g., .us) be required? If so, what should happen to the .com registry?
Are gTLD management issues separable from questions about International
Standards Organization (ISO) country code domains?
NSI Response to 5:
Retiring TLDs from circulation solves no problems and generates new
ones.* Promoting global electronic commerce requires uniting registries
on policy, standards and operational matters and removing geographic and
national barriers. These interests are not served by retiring TLDs or
requiring ISO TLDs. Corporations have expended large sums on their
Internet identities. Additional costs and consumer confusion would
result from the elimination of TLDs. In fact, many businesses and
organizations want to have a global "feel" and identification,
which accounts for the interest in worldwide TLDs. The current worldwide
TLDs have become market brands for more than a million companies. The
cost, business impact and disruptions to the Internet’s simple
conventions would be substantial and do not support the development of
global electronic commerce.
The notion that somehow the so-called "gTLDs" are the
problem is illogical. In fact, the very distinction made between TLDs
and country code TLDs is invalid. Today, there are some 60 domain name
registries and associated TLDs that have no residency requirement for
registration. These registries and their TLDs are physically and
functionally identical to .com on the Internet. These worldwide
TLDs include some of the largest registries in the world, such as
.uk (United Kingdom) and .de (Germany). The most recent
addition is the .to domain associated with the island nation of
Tonga. That commercial registry operates out of San Francisco, charges
fees identical to those charged by Network Solutions’ InterNIC operation
and, like the other open registries, admits registrants from any country
in the world.
[See Chart Four]
The notion of country code is less clear than it first appears. At
this time, the relevance of geographic and country codes is
questionable. As a Network Wizards Survey of domain hosts notes, there
is no correlation between a host name and its location. A host with a
.nl domain name could easily be located in the U.S. or any other
country. In addition, hosts under domains .edu, .org,
.net, .com and .int could be located anywhere.
There is no way to determine where a host is without asking its
administrator. Thus, although the registrant might supply a registration
address in the country he is registering, his host location might be
anywhere in the world. There is currently no way to control this.
One must recognize that the problems that are being encountered in
.com are related to volume, not geography. These problems will
occur in any TLD that reaches substantial mass in terms of the number of
users. Retiring TLDs will only shift the problem to the other TLDs
operations, or to the country codes TLDs as Internet use expands
geographically. Nominet, the .uk name registry, after close
consultation with NSI/InterNIC, has recently issued a domain name
dispute policy. During a visit to the U.S., the company indicated that
.uk was about two years behind .com in experiencing growth
and the inevitable dispute problems. (NSI was informed recently that
Nominet has encountered its first instance of dispute litigation.) Other
registries are probably in a similar position as a result of growth.
*Generic (or global) TLDs (gTLDs) was a term created by the International/Internet Ad Hoc Committee (IHAC) to distinguish certain three-letter TLDs from other TLDs. The term was created on December 20, 1996 by IAHC in their press release "IHAC Makes Its Preliminary Conclusions on Internet Top Level Domains Available for Public Comment," for the purposes of isolating .com, .org and .net TLDs for further discussion. The term gTLD is not used in NSI's responses.
CHART FOUR
Registry franchises where you can register without a local address or company:
| Country | Domain | Country | Domain
| - Antigua and Barbuda
| .com.ag | - Macedonia
| .mk
| - Armenia
| .am | - Malaysia
| .com.my
| - Azerbaijan
| .com.az | - Malta
| .com.mt
| - Bahamas
| .com.bs | - Martinique
| .mq
| - Bolivia
| .com.bo | - Mexico
| .com.mx
| - Burundi
| .bi | - Mongolia
| .mn
| - Cambodia
| .com.kh | - Mozambique
| .mz or .co.mz
| - Channel Islands, Guernsey
| .co.gg | - New Caledonia
| .com.nc
| - Channel Islands, Jersey
| .co.je | - New Zealand
| .co.nz
| - Chile
| .cl | - Nicaragua
| .com.ni
| - Congo
| .cg | - Norfolk Island
| .com.nf
| - Cook Islands
| .co.ck | - Panama
| .co.pa
| - Denmark
| .dk | - Peru
| .com.pe
| - Dominica
| .com.dm | - Poland
| .com.pl
| - Dominican Republic
| .com.do | - Puerto Rico
| .com.pr
| - Ecuador
| .ec | - Romania
| .ro
| - Fiji
| .fj | - Russia
| .com.ru
| - Germany
| .de | - Rwanda
| .rw
| - Gibraltar
| .gi | - Saint Kitts and Nevis
| .com.kn
| - Grenada
| .gd or .com.gd | - Saint Lucia
| .com.lc
| - Guadeloupe
| .gp | - Saint Vincent & the Grenadines
| .com.vc
| - Guatemala
| .com.gt | - South Africa
| .co.za
| - Isle of Man
| .co.im | - Sri Lanka
| .lk
| - Israel
| .co.il | - Suriname
| .sr
| - Jamaica
| .com.jm | - Switzerland
| .ch
| - Kazakhstan
| .kz | - Tonga
| .to
| - Latvia
| .com.lv | - Turkey
| .com.tr
| - Libya
| .com.ly | - United Kingdom
| .co.uk
| - Liechtenstein
| .li | - U.S. Virgin Islands
| .co.vi
| - Luxembourg
| .lu | - Uzbekistan
| .uz
| - Macao
| .com.mo | - Venezuela
| .com.ve
| | | | - Zaire
| .zr
|
|
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
Are there any technological solutions to
current domain name registration issues? Are there any issues concerning
the relationship of registrars and gTLDs with root servers?
NSI Response to 6:
Technical solutions. The DNS performs reasonably well,
especially in the non-hostile environment that has existed until
recently. The biggest difficulties in registration now are not
technical. The recommended technology improvements to today’s DNS
registration system were previously discussed in the first two questions
in this section. They involve standardization, more secure DNS software
and additional directory capability. Changes to the operational model
for registration, such as shared registration, will require technical
development and operational changes.
To clarify our points, consider the following observations on the
morphology of DNS. DNS is a hierarchical system with exclusive registry
points at each level in the tree, not just at the top (i.e., TLDs). At
each authority level in the DNS tree, the substructure is also a tree
with the current top node acting as a TLD. There is nothing magical
about the absolute top level; any distinction made between TLDs and the
lower levels in the DNS is artificial. Every level in the DNS is the
(exclusive) top level and authority for everything below it.
"Shared" registration calls for changing this only at one
level in the DNS.
[See Chart Five]
To implement the proposal for registry "sharing" at only
the TLD level in the DNS will require technical solutions including
software and databases to accommodate multiple concurrent registrars
with time-stamping requirements, standardized online authentication of
registrars and a very sophisticated certificate program to reduce
unauthorized access. The more challenging problems are operational and
are addressed in section D: Policies for Registries.
Root server relationships. Today the root server operators’
relationship with the registrars, TLDs and the larger Internet community
is informal. The problems with this framework are detailed in section
B question 1. A legal framework, as well as performance contracts or
compensation provisions, would shape a relationship with accountability
between the root server operators and registrars.
How can we ensure the scalability of the domain
name system, and address spaces, as well as ensure that root servers
continue to interoperate and coordinate?
NSI Response to 7:
DNS today is highly scalable, as is confirmed by respected technical
experts.
To secure DNS and ensure its scalability the following measures
should be taken:
- Coordinate an industry approach that protects Internet users from
any entity that displays hostile action. Examples of such hostile
actions should include:
- DNS hijacking through software vulnerabilities.
- Generation of extremely excessive traffic such as flooding a
machine by pings, DNS queries, SMTP mail bombing, etc.
- ISPs should be held accountable for a level of performance that
includes having fault tolerance, entering correct routing
information, correct maintenance of customers’ DNS records,
etc.
- The industry should be encouraged to adopt IPV6 and secure DNS as
soon as possible.
- Move forward with establishing the American Registry for Internet
Numbers (ARIN) and other IP registries to manage IP number allocation
under IPV4.
- Root zone servers should also be subject to a secure framework as
the number of servers is technically limited to 13 under the current DNS
structure.
How should the transition to any new systems be
accomplished?
NSI Response to 8:
An important element of this transition would involve
institutionalizing the functions performed by the IANA relating to the
"dot" root, IP numbers and network identifiers. These
functions must be managed in a way that transcends the personal interest
of any group. We have proposed earlier in our responses that a legally
anchored International Public Advisory Group (IPAG) be formed under U.
S. sponsorship. This approach would allow a stable period of
transition for public input in the development of new processes and
new structures. It is not unreasonable that the sponsorship of the U. S.
Government continue through these turbulent times.
A realistic structure that includes appropriate, yet limited,
government oversight is needed for the transition of management of all
of these functions and services. This proposed structure will provide a
badly needed stabilizing sponsorship for all of the Internet
administrative functions and services that are required by the growing
number of business (and personal) users of the Internet.
As we have noted before, although it has become a global
communication medium, the Internet was primarily developed and funded by
the U.S. Government and research community. For more than 20 years, the
U. S. Government underwrote the development of the Internet both
technically and financially. Given the tremendous importance of the
Internet and the billions of dollars now being invested in the Internet
by global businesses, it seems unreasonable to divest the U.S. of a
sponsorship role in the administration of essential functions until
stability is assured and a long-term global alternative can be agreed
upon.
NSI therefore proposes that a branch of the U. S. Government, through
the IPAG, assume an interim sponsorship role to guide Internet
administration through the transition period. The IPAG should have
international membership (like the North American Numbering Council)
that is based on Internet use.
Are there any other issues that should be
addressed in this area?
NSI Response to 9:
- Stability should be the preeminent issue as we forge a new framework
for the growth and expansion of the Internet. Stability is an absolutely
necessary for consumers to have confidence in the new technologies and
for businesses to make investments in new products and
services.
- Any attempted disruption of key elements of the Internet
structure should be treated as a criminal act with associated property
damage. The concept that "hacking" is not a crime should not be
tolerated in the future. The industry and government should close ranks
to ensure that proper measures are taken to protect the network and the
million of dollars in investment that it represents.
- The global nature of this technology should be recognized and
should be dealt with within each and every new software generation. This
includes internationalization such as currency translation and a
priori assumptions as to business hours.
There is no technical impediment to the creation of new TLDs, however there are other issues such as the impact on trademark
disputes and the potential for consumer confusion. Adding new
TLDs does not significantly increase the number of second-level
domain names available. The distinctions often made between TLDs
are artificial and are not supported by reality.
Are there technical, practical, and/or policy
considerations that constrain the total number of
different gTLDs that can be created?
NSI Response to 10:
Technical issue. NSI believes that there is no technical
reason why the number of worldwide TLDs needs to be limited. Domains are
equally scalable at all levels of the DNS. There can be as many TLDs as
there are potential second- and third-level domains.
This view is shared by Dr. Paul Mockapetris, the original creator of
the DNS. In an interview that appeared in a May issue of Web
Week*, Dr. Mockapetris stated, "I believe that there are no
problems in expanding the TLD space that we couldn’t solve in time. In
fact, there may be no problems at all. But we don’t know for sure, so I
would favor adding a small number (say 20) to start, then maybe 100,
then 1,000, etc. if you want to get to a very large number. I’m just
being cautious in my old age."
Other factors. Though there is no technical reason to limit
TLDs, an increase in the number of TLDs may generate consumer confusion.
In all likelihood it will increase the number and cost of trademark and
domain name disputes. The addition of new TLDs will also generate a new,
more vigorous round of speculator activity.
*Web Week, volume 3, issue 15, May 19, 1997 © Mecklermedia Corp.
Should additional gTLDs be created?
NSI Response to 11:
Market interest should be the best determiner of the maximum number
of TLDs. The addition of new TLDs to the market as exclusive brands
establishes a greater incentive for registrars to invest in improved
services. It appears that consumers are split on this issue.
There are no current market data to support the notion of consumer
interest in new TLDs. Demand for new TLDs is largely assumed. Moreover,
in some segments of the commercial world -- particularly among
well-known consumer product companies -- there is a desire to have no
new TLDs. These companies are purveyors of many popular brands and spend
significant sums to police these brands. They will feel compelled to
register in whatever new TLD emerges to offset the possibility of
"warehousing" by domain name speculators. This generates a
particularly difficult dilemma for owners of strong trademarks.
On the other hand, small businesses and those who failed to get the
domain name that they initially wanted see new TLDs as an additional
chance. The reality, however, is that without large numbers of new TLDs,
the system does not scale. Adding seven TLDs only allows seven new
entries for each desirable name that is not a strong trademark. In the
case of strong trademarks, new TLDs offer no relief and, as we have
noted, they likely would create new and additional difficulties.
The demand may increase for new TLDs if they are perceived to add
benefit. Specific industry groups may form their own TLDs to offer
unique services within industry practices. For example, .law
could offer DNS registration and possibly software or digital
identification specific to the legal profession.
Whatever the decision, we must ensure stability of the Internet by
introducing new TLDs to the DNS in phases.
Are there technical, business, and/or
policy issues about guaranteeing the scalability of the
name space associated with increasing the number of gTLDs?
NSI Response to 12:
Without large numbers of new TLDs, the system does not scale. Adding
seven TLDs only allows seven new entries for each desirable name.
(See the first two questions in this section for additional
thoughts.)
Are gTLD management issues separable from questions
about ISO country code domains?
NSI Response to 13:
Uniformity in the management of all TLDs would be both appropriate
and fair to the registrars. There is no rationale for imposing practices
on worldwide TLDs that are not also imposed on the others. The topology
of TLDs, hosts and networks does not conform to our a priori
notions of geography.
In fact, the distinction between worldwide TLDs and ISO country code
TLDs is not all that clear, as virtually all TLDs are managed by private
enterprises, not governments, and over 60 of the ISO TLDs have no
residency requirement for registration. [See Chart Four]. They are
functionally identical to worldwide TLDs.
Note there is not necessarily a correlation between a host’s domain
name and where the host is actually located. A host with a .nl
domain name could easily be located in the U.S. or in any other country.
In addition, hosts under domains .edu/.org/.net/.com/.int could
be located anywhere. There is no way to determine where a host actually
is without asking its administrator. There is not necessarily any
correlation between a network number and a domain name. A single network
number could span many countries, and a single domain may have hosts on
multiple network numbers. (See section B question 5 for additional
related discussion.)
Are there any other issues that should be
addressed in this area?
NSI Response to 14:
The .us TLD remains largely underdeveloped and is unused by
the commercial sector and individuals. The future of the .us TLD
should be addressed in the overall consideration for TLD administration.
A transition to shared registration in TLDs will require research,
design, and the development of software systems. It must be extensively
tested before being introduced to the current DNS. The scope of the
operational challenges significantly exceed the technical development
complexities. The Internet remains a frontier where exploitation of any
weakness in the system is likely.
Should a gTLD registrar have exclusive
control over a particular gTLD? Are there any technical limitations on
using shared registries for some or all gTLDs? Can exclusive and
non-exclusive gTLDs coexist?
NSI Response to 15:
Exclusive control at every level is the basis of the DNS and has been
since its inception. Adding new, exclusive TLDs is certainly the
easiest, quickest and most stable manner in which to increase
competition in DNS registration. This approach has the collateral effect
of expanding the name space. NSI is prepared to support the development
of a stable and secure shared registration system, if that is the
ultimate determination. However, NSI believes that any requirements for
sharing should be applied equally across all TLDs, second-level domains
and ISO country codes. At this time, NSI is unaware of any existing
technical system capable of achieving global "shared"
registration of TLDs. Over the last four years, NSI has built the most
advanced registration system available anywhere. The difficulty NSI
experienced in this effort indicates that the operational complexity and
technical requirements of achieving a global shared registration system
is underestimated by virtually everyone. The answer is more complex than
"sticking a database on a server" somewhere. The peer-to-peer
nature of DNS and the stability of the Internet demands that rigorous
testing of the system must be done and that adequate safeguards must be
in place prior to implementation.
Shared Registration. A "shared" registration system
proposal likely has a two-tiered structure consisting of a
Central Registry Facility (CRF) for all TLDs at the top-level, and a
potentially unlimited number of authorized competing registrars at the
second-level. All registrars must have concurrent online access to the
CRF database and systems in order to register, update and delete
second-level domain names on behalf of their clients.
In some cases there will be a three-tiered structure
consisting of an ISP or name "re-seller," a registrar and the
CRF registry. This assumes that ISPs and name "re-sellers,"
who are not approved registrars, are not authorized to access the CRF
systems directly, and therefore must deal through a registrar.
How customer data (both back-office and front-office)
will be duplicated or distributed among these three parties is an
important operational issue. There will be a significant potential for
confusion, endless communications and costly telephone conferences to
resolve customer difficulties. Without carefully tested policies and
operational procedures, NSI’s experience has shown that unforeseen
problems arise and client frustration levels are very high.
Technical Requirements. Starting with the basic assumption of
a shared registration system, we will examine the technical requirements
of such a system. The most significant technical problems arise in
implementing the CRF systems. These systems will need to at least
functionally duplicate those of the InterNIC registry, which have taken
several years and many millions of dollars to develop and refine.
Designing these systems so that they are accessible to a large number of
registrars, and yet simultaneously provide secure and protected
operations, presents a significant new challenge. To avoid chaos, this
system must be very well designed and will require coordination among
all of the registrars before operations commence. To implement shared
registration on the Internet will require research, design, development
and testing of software to introduce this new technology into the DNS.
The most visible system that is needed is the CRF database. If we
assume that it employs a modern relational database management system
(RDBMS), we estimate that it could be initially operational in 18 months
of design and testing after completion of the operational concept (which
we estimate would require at least six months).
Operational Complexities. There are too many unknowns to
specify the operations of a CRF system. Our approach here is to indicate
a few areas where careful consideration of the operational policies and
procedures is needed, and to illustrate the nature of problems (based on
NSI’s experience) that might be experienced. This is not meant to be a
thorough or exhaustive operational critique.
- System Loading.
The rush to register names on the first day
of operation, before they are taken by competitors, is likely to result
in tremendous initial loading. The InterNIC registry experienced a
gradual buildup of registration volumes over a several year period,
allowing phased operational improvement, a luxury the new CRF is
unlikely to experience.
- Client Responsibility
. Who ultimately is responsible to the
registrant? Is it the registrar, the CRF, the ISP or a name reseller?
This is a critical design decision. This could generate massive customer
confusion.
- Changing Registrars.
Registrants would be able to switch
registrars. If the registrar is also the ISP, then switching registrars
has implications on the choice of ISP, collateral hosting and IP
numbers. Recall that IP addresses are not portable over ISPs. In some
cases, switching registrars may result in a costly and disruptive
situation for customers and users.
- Delinquent Payments.
If customers owe money to a registrar,
would they be allowed to switch to another registrar without paying
their debt to the current registrar? How would customers know who to
call for assistance, and couldn’t customers be switched unscrupulously
such as in the long distance telephone industry? Malicious activities
would be much more difficult to detect, discourage or prevent.
- Template Updates.
Could updates be made through any
registrar? Updates across registrars would require agreed-upon
standards. Authentication (identification) and authorization (legal
representation) would be extremely important operationally. An
authentication system standard would have to be the same
"end-to-end" from registrar to registrar as well as from ISP
to registrar to CRF. Would registration updates be free or would they
require a fee? This has a significant financial impact as indicated by
the fact that at NSI the volume of registration updates exceeds that of
new registrations.
- Online Directory.
The CRF must allow free exchange of data
with a new whois directory interface. This has competitive implications
for registrars: who will maintain the new whois directory? What data
will it have? Will use of the data be restricted?
- Help Desk and Customer Support.
The CRF would need a
significant help desk to assist in disputes and to enforce registrar
policies and procedures. The assumption is that the individual
registrars would have their own customer support facilities.
- Complaints and Disputes.
The CRF would require a new
system for dispute resolution. A universal dispute policy would be
needed for registrar disputes, end-user complaints about registrars,
etc. Who handles complaints about the CRF? What are a customer’s options
if they are dissatisfied with the CRF? How would disputes between
contacts be handled? Will the geographic separation exacerbate dealing
with disputes?
- ISP Host Data
. How will ISPs change host (name server)
information? Will the CRF do it? If not, how will multiple registrars
handle host information?
Coexistence. At the risk of possible consumer confusion,
both shared registration and the current exclusive registration could
coexist. This is fortunate, since it would seem to be prudent to fully
test a newly developed shared system in a significant operational
setting on new TLDs before attempting to use it on existing exclusive
Internet TLDs.
Should there be threshold requirements for
domain name registrars, and what responsibilities should such registrars
have? Who will determine these and how?
NSI Response to 16:
There should be threshold requirements for domain name registrars,
however they should be kept to a minimum so as not to discourage
competition. Those who meet minimum technical and financial criteria
should be allowed to compete in DNS registration. Consumer demand will
quickly improve the quality of the offerings.
We propose the following approach:
- Interested parties meeting the minimum technical and
financial requirements should be allowed to compete in the TLD
registration business.
- Specific, individual, and exclusive TLDs would be proposed by every
registrar and would be issued on a first-come, first-served basis (with
some limitation, e.g. three per registrar).
- Each registrar would deploy and operate a set of servers to support
their own specific TLDs where they deemed appropriate, or contract this
service to others.
- All TLD registrars would be assessed fees, based on business volume,
to support the IPAG and its DNS root server contractor(s).
Registrars would establish their own operating procedures, including
second-level domain name dispute policies, marketing, service standards,
and collateral services consistent with their own brand identity for
their own TLDs within the policy guidelines of the IPAG. Consumers could
then select from among widely differing and competing concepts and
price/service offerings. Competition will drive quality. Registrars will
set their own contractual terms, including pricing. Client information
databases would be proprietary, while DNS databases would be public.
The framework for handling practical and policy issues allows for a
limitless number of TLDs. The procedures NSI is proposing will encourage
investment by participating registrars and guarantee ever-increasing
gains in efficiency and improved customer service.
Are there technical limitations on the
possible number of domain name registrars?
NSI Response to 17:
No technical limitations. There is nothing magical about the
top-level of the DNS. Any subset of the DNS system is structurally
analogous to the whole DNS. Just as there can be a virtually limitless
number of second-level domain names (and therefore
"registrars"), so can there be a limitless number at the
top-level. This is the nature of hierarchical tree structures. As Dr.
Mockapetris stated in an interview, "I believe that there are no
problems in expanding the TLD space that we couldn’t solve in time. In
fact, there may be no problems at all."
Are there technical, business and/or policy
issues about the name space raised by increasing the number of domain
name registrars?
NSI Response to 18:
The Internet is still a frontier and retains somewhat the character
of the "Wild West." NSI’s experience has shown that abuse will
occur, if possible, and that people will exploit any hole in any system,
no matter how obscure.
Management challenge. Managing a very large number of
registrars could introduce major policy issues and procedural problems.
Unlike today, where registrars are functionally independent, they would
be required to cooperate.
Viability. The financial viability of a large number of
registrars and the inherent instability of marginally successful
registrars is a concern. A failed registrar can have severe impacts on
the business of its clients and effect the overall perception of the
Internet as a viable medium for commerce.
New form of speculation. Speculation could become a much more
significant problem than it is today. The CRF must be able to handle the
expected load from a new class of semi-legitimate speculators.
Registrars could register large numbers of names at minimal cost (the
cost charged by the CRF) to resell later at market prices. As an
example, web-hosting firms could be tempted to become speculators to
better position themselves for gaining clients. One could conceive of a
scenario where a hosting firm registers a large number of names for
themselves that are known or desirable marks, and then offers them to
clients as a condition for contracted services.
Should there be a limit on the
number of different gTLDs a given registrar can administer? Does this
depend on whether the registrar has exclusive or non-exclusive rights to
the gTLD?
NSI Response to 19:
No physical limits. There can be an unlimited number of
registrars at each level within DNS - either shared or exclusive. The
CRF’s and registrar’s bandwidth, memory and CPU capability define the
limit of their respective delegations. There is empirical evidence that
the top level of the DNS can easily handle 1.5 million delegations, and
this empirical limit is increasing daily.
Procedural limits. NSI is concerned that moving to a totally
unregulated marketplace and allowing unlimited TLDs could result in an
undesirably chaotic environment if the appropriate infrastructure
(policies, procedures, technical systems and operational support) is not
in place. This is especially true in the case of non-exclusive
registration models. It would seem prudent in either case to limit the
initial number of new TLDs, adopting a phased program for their
introduction.
Are there any other issues that should
be addressed in this area?
NSI Response to 20:
Unique legal concerns. Adopting non-exclusive TLDs generates
some unique
policy questions with legal impacts. These following issues must be
addressed prior to system design:
- How will enforcement of policies and procedures across
registrars be managed? What will be the legal enforcement mechanism?
Shared TLDs could motivate unscrupulous tactics between registrars. How
will registration information be protected to avoid malicious changes?
What would be done to protect one registrar from the possibility of a
competing registrar using its customer information to steal customers?
How would any of this activity be detected and controlled, particularly
across national boundaries?
- How would delinquent accounts be handled in light of registrar
switching? Would switching to another registrar be allowed without
payment to the first one?
- Would shared TLDs necessitate a universal policy for trademark
disputes, registrar disputes, and end-user complaints about registrars?
How would disputes between different points of contact for the same
domain name be handled?
- If customer information is shared across registrars, will there
need to be controls about how that data is used and/or disseminated by
registrars (e.g., mailing lists)?
- Would the shared TLD model force registrars to operate at the
lowest common denominator with regard to security? What would be done to
avoid this potentially destabilizing requirement?
- Updates of host name server records associated with one domain
usually have impacts on all other domain names associated with that
host; what would happen when the host involved has names registered by
different registrars?
- What systems would be put in place to provide for an audit
trail of domain name record changes to provide the necessary history for
dispute resolution? How would these systems be implemented across
registrars sharing the same TLDs? In cases where multiple registrars
have had involvement throughout a name’s history, would all associated
registrars be liable for damages assessed against one? Would there be
joint liability just because it is a shared process?
- How would the first-come, first-served rule be applied in
relation to inherent disadvantages such as multiple time zones, and
temporary disadvantages such as partial network outages?
There is a vacuum in international law defining the relationship
between domain names, which are global, and trademarks, which are
governed by various national statutes. As a result of a few highly
visible cases, the extent of conflicts between trademark owners and
domain name holders is perceived to be greater than the actual facts
support. Employing its dispute policy since July 1995, NSI has
experienced 1300 disputes in 1.6 million registrations. This is less
than one-tenth of one percent of registrations or, in other words, one
in one thousand registrations result in a dispute. Some of these
disputes have resulted in law suits brought by either the trademark
owner or the domain name holder against the other claimant. To date, 30
such suits also have included NSI, as registrar of the domain names in
issue. All registrars will eventually experience such difficulties as
the worldwide volume of registrations grows to U.S. levels.
Pre-registration review of domain names against an ever-changing
database of trademarks is fatally flawed. Immunization of registrars in
such disputes would allow registrars to utilize simpler dispute
policies.
What trademark rights (e.g., registered
trademarks, common law trademarks, geographic indications, etc.), if
any, should be protected on the Internet vis-a-vis domain names?
NSI Response to 21:
Given the state of the law at present, it is unclear what rights
should be protected with respect to domain names. Owners of valid
trademarks are protected from infringement and dilution of those
trademarks by federal statutes enacted in virtually every country of the
world. If trademark laws apply to domain names, anyone who obtains a
domain name under a registrar’s policies of issuance must do so subject
to whatever liability is provided for by law.
A domain name can be more than a mere Internet address. At the
present time, it also identifies the Internet site to those who reach
it, much like a person’s name identifies a particular person, or more
relevant to trademark disputes, a company’s name identifies a specific
company.
Confusion on the part of the public, as to the origin or source of
goods and/or services, is exactly what the trademark laws were designed
to protect against. Although some trademarks are easily discerned, such
as Pepsi or Ford, other trademarks, such as Disc or Acme, are more
difficult to protect. Federal trademark statutes do not specifically
address the issues presented on the Internet. Case law, however, is
beginning to emerge. Unauthorized use of a domain name by a domain name
holder, which includes a protected trademark, to engage in commercial
activity over the Internet has been held to constitute "use in
commerce" in violation of federal statutes. Such use has been held
to be in direct conflict with federal trademark law. How the trademark
laws of the world will be applied or interpreted remains to be seen. For
all of the instances when a domain name is considered by the trademark
owner to be "close enough" to its protected trademark to be
disputed, it is for the trademark owner, in its discretion, to police
its own trademark rights.
Until legislation or case law defines the relationships between the
various intellectual property rights and domain names, there can be no
answer to "what trademark rights, if any, should be protected on
the Internet vis-à-vis domain names." As with each new
medium (radio, television and, now, the Internet), the law will continue
to be the law and will continue to apply to the actions of the
Internet’s participants until a more suitable framework for addressing
the complexity and global nature of these disputes can be implemented.
Should some process of preliminary
review of an application for registration of a domain name be required,
before allocation, to determine if it conflicts with a trademark, a
trade name, a geographic indication, etc.? If so, what standards should
be used? Who should conduct the preliminary review? If a conflict is
found, what should be done, e.g., domain name applicant and/or trademark
owner notified of the conflict? Automatic referral to dispute
settlement?
NSI Response to 22:
It has been proposed that applications for domain
name registrations be reviewed to determine if the domain name
"conflicts" with a trademark, a tradename, a geographic
indication, etc. While this is an interesting concept, and one that has
certain appeal to the trademark bar and trademark owners, it is a
concept that would be neither reasonable nor practical in practice.
To conduct a pre-registration review of a domain name application
would probably require something like the "notice of
publication" period in the trademark registration process. To
establish such a process in the domain name registration system would
impose an inordinate cost and burden. A representative indicator of the
effect of such a requirement lies in the U.S. Patent and Trademark
Office (U.S.P.T.O). Its trademark registration process includes a
16-month pendency period between the time an application is submitted
and the time an application is granted. Network Solutions, as one of
approximately 200 registries worldwide, is currently registering a
volume of domain name applications that is several times larger than the
volume of U.S.P.T.O. applications each year. Significant additional time
and cost would be involved by applying a U.S.P.T.O.-type process to
domain name registrations.
Determination of trademark infringement involves the application of
up to 16 different factors. In different parts of the world, courts
apply varying degrees of emphasis to those factors. A domain name
registrar should not be expected to have the knowledge, expertise and
experience of an intellectual property attorney or a judge on a global
basis and, without that, determination of infringement cannot be
rendered. Additionally, having to make such evidentiary determinations
would expose a domain name registrar to significant liability in the
event a wrong decision was made.
The entire issue of pre-registration determinations becomes even more
difficult and unreasonable when considering the global nature of the
Internet. Pre-registration determinations of trademark conflict would
require the availability of a database of trademark registrations from
every country in the world. Such a database does not exist.
Aside from a preliminary review process, how
should trademark rights be protected on the Internet vis-a-vis
domain names? What entity (ies), if any, should resolve disputes?
Are national courts the only appropriate forum for such disputes?
Specifically, is there a role for national/international
governmental/nongovernmental organizations?
NSI Response to 23:
Trademark rights should be policed and protected on the Internet in
the same manner as they are in other aspects of commerce. It should be
each trademark owner’s responsibility to monitor domain name
registrations, and to take appropriate action against a domain name
registrant when the trademark owner, in its discretion, believes its
intellectual property rights are being violated. The same forums that
deal with intellectual property rights in international commerce are
also the appropriate forums to resolve disputes when domain name
registrants and trademark owners are residents of different nations.
How can conflicts over trademarks best be
prevented? What information resources (e.g. databases of registered
domain names, registered trademarks, trade names) could help reduce
potential conflicts? If there should be a database(s), who should create
the database(s)? How should such a database(s) be used?
NSI Response to 24:
Conflicts over trademarks cannot be prevented, but they can be
reduced. Nothing will eliminate either the intentional or unintentional
registration and use of a domain name on the Internet by someone who
either seeks to confuse the public as to the source of the product or
service or, alternatively, was unaware that his/her actions caused such
confusion. The question, however, starts with the premise that the
millions upon millions of potential domain names will, in fact, conflict
with trademarks. The premise is incorrect and will, if applied, forever
change the nature and direction of this global communications medium. In
fact, to date Network Solutions has applied its Domain Name Dispute
Policy to about 1,300 domain name registrants, or approximately one
tenth of the one percent of the total number of domain name registrants
in the database.
The conflicts, which will occur on a global basis, can be drastically
shortened and resolved with minimal cost to both the trademark owners
and domain name holders of the world through the passage of a simple
rule of civil procedure to be enacted, jurisdiction by jurisdiction, as
it proves successful. The procedure would have the following goals:
- Allow for the rapid expansion of the Internet;
- Keep any technical portion of the solution simple and universal;
- Keep the "first-come, first-served" assignment of
second-level domain names;
- Allow registries world-wide to follow their own procedures for
disputes;
- Limit government involvement, administration and regulation to
increase global commercial investment in the use of the Internet; the
model for commerce would be contractual rather than
regulatory;
- Allow for the necessary divergence of trademark law and legal
systems worldwide;
- Prevent the formation of an "administration" over
disputes, including personnel and regulations;
- Use only existing global legal systems;
- Protect trademark owners to the maximum extent possible; in
instances where a domain name directly conflicts with an identical
federal registered trademark, allow for an administrative alternative to
judicial proceedings for domain name holders;
- Protect domain name holders’ rights to use their own names or
their assignment of domain names which predate trademarks; and
- Allow jurisdictional prerequisites for personal jurisdiction over
the disputants to be unaffected worldwide.
Should domain name applicants be
required to demonstrate that they have a basis for requesting
a particular domain name? If so, what information should be
supplied? Who should evaluate the information?
On the basis of what criteria?
NSI Response to 25:
No. Inherent in this question is the concept of prescreening and the
accompanying delays and bureaucratic review that will bring this medium
to a standstill. Proposals have even surfaced that would require such
invasions of privacy as the production of the tax returns, social
security numbers and corporate charters. Who is qualified to review
these materials and who would make such decisions? Who would decide what
an appropriate "basis" would be? How would these decisions be
appealed?
How would the number of different gTLDs and the
number of registrars affect the number and cost of resolving trademark
disputes?
NSI Response to 26:
An increase in the number of different TLDs would, in all likelihood,
significantly increase the number and cost of resolving disputes over
the registration and use of domain names. Ignoring the ISO-3166 two
letter country code TLDs, trademark owners presently have to monitor and
"police", if they so choose, three TLDs: .com,
.org, and .net. A domain name in any one of the TLDs could
be the subject of a dispute initiated by a trademark owner that believed
the registration of the domain name violated the trademark owner’s
intellectual property rights. All other things being equal, a trademark
owner could currently be involved in up to three separate disputes, each
with its corresponding cost, if the trademarked word or phrase was
registered as a domain name in each of the three TLDs. It logically
follows that there is a direct relationship between the number of TLDs
and the potential number and costs of disputes related to the protection
of trademark rights.
An increase in the number of worldwide TLDs could also affect costs
to individual companies as they seek to protect their trademarks on the
Internet by registering all their trademarks as domain names in each of
the worldwide TLDs. If, for example, there were a total of 50 TLDs, a
large company with 1,000 trademarks could decide to incur the costs
(time and resources) associated with registering and maintaining 50,000
domain names (50 times 1,000). Such a burden would be unreasonable.
As long as domain name registries or registrars are susceptible to
allegations of contributory infringement of intellectual property rights
of trademark owners, an increase in the number of registries or
registrars would, in all likelihood, significantly increase the number
and cost of resolving disputes over the registration and use of domain
names. More registrars would simply provide more candidates for
trademark owners to include as a named party to litigation. The
potential for cost growth would be compounded if the registries were
located outside the United States. Registries located outside the United
States increase the possibility that trademark-domain name disputes
would take on an international character, with the proportional increase
in cost and complexity.
An increased number of worldwide TLDs also increases the possibility
of a type of dispute not yet seen. That dispute would be between two
registrants, each with the same second-level domain name but in
different TLDs, arguing over confusion and superiority of rights based
upon the date of registration or activation of each second-level domain
name in its respective TLD. The significance of the nature of these
disputes lies in the fact they would not necessarily involve entities
with trademarks. They would involve simply "first use" on the
Internet.
Where there are valid, but conflicting trademark
rights for a single domain name, are there any technological
solutions?
NSI Response to 27:
There may be some technical relief available. Registrars might employ
a shared top-level home page with HTML "links" to each user.
Therefore Delta Airlines, Delta Faucet and other Delta’s would share a
simple top-level home page, Delta.com, with "links" to
each one’s unique home page. This is a service that each registrar could
provide. NSI has attempted to persuade disputing claimants to employ
this tactic. All have declined, preferring to seek exclusive control at
great risk and expense.
Are there any other issues that should be
addressed in this area?
NSI Response to 28:
The legal issues related to the registration and use of domain names
are complex and multifaceted. Answers are not readily at hand because of
the absence of legislation pertaining to the relationships between
domain names and intellectual property rights. Even when existing
legislation may apply, it varies from country to country. Case law is
just starting to evolve, and that too, will vary from country to
country.
Beyond the applicability of existing and emerging intellectual
property laws, legal issues must include consideration of international
laws such are embodied in the Paris and Berne Conventions. These, and
possibly others, must be examined for applicability to the area of DNS
and the Internet.
In addition to intellectual property matters, the United States has
federal statutes which limit or prohibit the use of specific words or
phrases. For example, the U.S. Amateur Sports Act addresses the use of
the words Olympic, Olympiad and Citius Altius Fortius (the
Olympic motto). There are, in all likelihood, similar statutes in
most countries of the world that must be considered when dealing with
legal issues related to global domain name registration and use.
![[See Chart One]](rtserver.jpg){kind=link}
![[See Chart Two]](dname.jpg){kind=link}
![[See Chart Three]](interim.jpg){kind=link}
![[See Chart Five]](dns.jpg){kind=link}