Back to ICANNWatch.org

The idea of ICANN was the same as the idea of the internet. ICANN was to be open, voluntary, standards-based, decentralized and built on cooperation. These core ideas prefigure the solutions to the organizational issues now being debated in the DNSO working group, the issues to be covered by the At Large Study over the next few months, and questions being raised about ICANN's structure and function in many other contexts.

None of the following is meant as an attack on ICANN or any participant in ICANN. Indeed, many of the ideas advanced in this piece are drawn from ICANN's contracts and its constitutional documents.We believe the current debate about ICANN's structure has confused personal attacks on today's ICANN participants with discussion of possible structural reforms.In our view, this confused discussion is not as constructive as it should be.We hope and expect that many who have contributed to the ICANN experiment will endorse the proposition that ICANN's institutional strength will grow in proportion to the extent ICANN remains true to its core ideals.We have attempted to set forth these principles below.

What it means to be Open

The internet is open in the sense that you do not need permission to join. You just need a connection with a neighbor and willingness to follow the standards that make interconnections work. ICANN should also be open, in the sense that anyone willing to work constructively towards development of consensus-based standards should be allowed to join its activity.[ii] There may be different categories of individuals and groups that play somewhat different roles in ICANN. But because the work product of ICANN is required to be policy standards that registries and registrars agree to follow because they are supported by a broad, demonstrable consensus among impacted parties, the exact composition and admission criteria for particular subgroups within ICANN should not matter very much. No one can claim the right to exclude from ICANN's work processes any voice that offers constructive contributions to the collective work of developing consensus-based policy standards. 

• It follows that the core of the DNSO is the General Assembly, an inherently open body, not the Names Council (which should act only as a steering committee and be selected by some means that fairly canvas the views of the entire GA). 
• It follows that the constituency structure of the DNSO should be more flexible and fluid, and should certainly include a home for anyone who has a domain name registration. 
• It follows that the At Large membership that elects some of the Board should be provided with the means to deliberate continuously regarding policy issues. 
• It follows that the potential for overlapping memberships in these various groups (GA, Individual Domain Name Holder Constituency, At Large Membership) is not troubling, because ICANN should encourage the self-formation of any group to which the consensus question can usefully be asked.
• It follows that ICANN should ensure that its online resources adequately explain, in layman's terms and in many languages in addition to English, what ICANN is and how it does its work.

No one is required to join the internet. Anyone can form their own private network, without penalty, using any protocol they want. ICANN should also be voluntary, in the sense that participation in the policy standards it sets is a matter of contract, not governmental decree. 

• It follows that ICANN should not actively oppose the formation of alternative root servers, and should behave as a good neighbor in relation to any other efforts to provide global addressing and identity management. 
• It follows that ICANN should not be led by governments.
• It follows that ICANN must base its authority and legitimacy on contracts that obligate registries and registrars and registrants to implement its future policies, sight unseen, based on the nature of the process used to develop those policies. The only contracts that make sense, in this context, are contracts based on the willingness of each participant to defer to standards endorsed by most of those who are impacted by them and not vigorously and rationally opposed by any substantial minority. That is the basis for the idea of "consensus" that is built deeply into existing ICANN agreements. That is why ICANN is not a government and can exercise no sovereignty. 
• It follows that the At Large membership structure cannot be based on the theory of representative (or direct) democracy. The principle of one-person-one vote provides a basis for delegating a people's sovereignty to a government. It does not provide legitimacy for a system that seeks voluntary compliance with policies that have the support or acquiescence of all groups particularly impacted by those policies. 
• It also follows that low-percentage participation in ICANN is not as troublesome as is low voting turnout in democratic systems. Lack of participation may represent willingness to acquiesce in ICANN's decisions (following adequate publication of the nature and impact of proposed policies) or the existence of a marketplace for alternative policies that allows individuals who prefer other regimes to seek them out.

The internet is not a single system. It isn't owned by anyone. It can't be shut off from any single point. Participation in it doesn't subject the participants to rules made by a global governing body. While ICANN is technically a California not-for-profit corporation (with bylaws that require an internationally representative Board), it doesn't "own" the authoritative root system, and it isn't "owned by" any group of shareholders or members. The only thing that makes any root "authoritative" is the fact that ISPs and others running name resolution software point at it.

• It follows that ICANN should not create a class of "members" with special rights.
• It follows that the ICANN Board cannot behave as if it should be granted a perpetual property right to decide what goes into the "authoritative" root. 
• It follows that the Board must enlist the respect and active support of a vast majority of internet participants. Otherwise, some other organization that can enlist that support will, sooner or later, take ICANN's place. 
• It follows that ICANN should make clear to all participants that its constituting documents and contracts sharply limit the types of policies it can develop and enforce.
• It follows that ICANN policies should be required to be re-endorsed from time to time to make sure the impacted community still supports them by consensus. 
• And it follows that financial support for ICANN is warranted to the extent that it usefully aids the policy development process -- for policies the relevant communities believe are necessary.

The internet emerges from the combination of vast numbers of decisions taken at its edge. No one has to ask permission to put up a web site. ICANN, in taking decisions regarding policy standards for the domain name system, must continue to recognize that most of the actions that affect the functionality and value of the domain name system are taken by independent parties, acting on their own, without any need for permission. 

• It follows that ICANN should make and attempt to enforce global policies only when there is a clear need for uniformity and well-documented consensus support exists among those who must implement such policies and are impacted by them. In the absence of an ICANN consensus-based policy, the default should be that registries may act as they see fit to enhance the value of the TLD they administer, that registrars are free to seek ways to create value for their customers, and that registrants can use the names they control in any otherwise lawful fashion and in compliance with their registration contracts.
• It follows that selection of the Board should stress diversity of outlook so the Board can serve as a device for detecting and unmasking bogus claims that consensus exists (when relevant impacted groups have not been adequately consulted and would otherwise not be heard). For example, to encourage such diversity on the Board, closed ccTLDs or chartered gTLDs might usefully be allowed to form their own Supporting Organizations. Closed registries have different interests than open registries and should be able to speak effectively and collectively with respect to issues that affect them.

• It follows that there should be many more DNSO constituencies and working groups and task forces discussing various issues. Failure to reach a global consensus may be a success rather than a failure, however, because it leaves undisturbed the power of many diverse and decentralized actors to make their own decisions. These actors may find even better ways to proceed than might have emerged from a compromising committee.
• It follows that ICANN should not take upon itself the burden of attempting to prevent "consumer confusion" by, somehow, restricting the proliferation of names or addressing protocols or top level domains. The internet grew quickly because no one attempted to gate its development. Decentralized standards development necessarily creates some uncertainty and confusion but also generates much greater choice and a market for services designed to help consumers make good decisions. 
• The essence of the internet is not reduction of choice. It is insistence on end-to-end connectivity and avoidance of conflict. This suggests that the guiding principle for ICANN decisions regarding opening up new top level domains should be to encourage lawful innovation and to seek only to avoid destructive collisions between names or undesired fragmentation and disruption of end-to-end connectivity.

The internet arose from the self-interest of various network operators seeking to share communications via standard protocols. It represents a major victory for decentralized collective action -- a form of social ordering we ignore at our peril. Its construction required a lot of hard work.To succeed, ICANN needs to inspire and catalyze the same kind of sharing and much more hard work than has been performed to date. Articulating, exploring, and discussing the details of policies likely to attract consensus support requires patience and precision -- not just open listservs and occasional meetings in places most impacted parties find it hard to get to. Reaching consensus requires compromise, and may require better-structured online meetings. 

• It follows that ICANN's structures must reward the production of persuasive, broadly supported, well-documented policy proposals and nothing else. 

• It follows that ICANN must commit itself to true bottom-up consensus building, and must quickly make available online tools that facilitate this hard work.
• It follows that the Names Council should remember that its role is to recognize consensus, not to act as a legislature in originating and recommending policy to the Board. The Names Council should be far more active in facilitating the consensus-building process.
• It follows that ICANN's structures must not reward seat-claiming or grandstanding. It must seek to avoid conferring "offices" or helping any faction to exclude (or impose its will on) others. It must never provide a central point of "control" that could (and almost certainly would) be hijacked by particular interests -- for purposes of closing down access to the consensus standard-setting process, imposing involuntary rules, or building up centralized power.

• Allow any group that obtains support from at least five percent of the members of the open General Assembly (or that represents registries or registrars with, collectively, at least five percent of all registrations) to form a constituency of the DNSO, without any requirement for approval by the ICANN Board, and to select a member of the steering committee (known as the Names Council).
• If a proliferation of constituencies would create a Names Council that is too large to conduct the day-to-day business of facilitating the development of consensus policies, split the Names Council into subgroups that could play that role for particular subsets of policy development or that are charged with reaching out to particular constituencies.
• Allow any group that wants to form a working group or task force for the purpose of developing reports to the DNSO regarding consensus-based policy development efforts to do so.
• Allow any group that has a clear, substantial, and continuing involvement with domain name policy (or other matters within ICANN's mandate) to elect at least one member to the ICANN Board. Eliminate any other pre-established entitlement of any group to any particular number of seats on the ICANN Board. Allow the Board to make some additional appointments to fill obvious holes of expertise or outreach capability. Require all groups that elect ICANN Board members to pay fees approved by those who collectively pay two-thirds of the amounts levied on any particular group.
• Provide mechanisms for the At Large membership to communicate continuously among themselves regarding policy issues. Allow access to the list of At Large members to any interested group, while allowing individual members to opt out of receiving any email communications sent using this list.
• If a proliferation of Supporting Organizations would create a Board too large to conduct the day-to-day business of ICANN, establish a small group within the Board charged with administration of the organization, and charge the larger Board group with reaching ultimate decisions on questions regarding whether a particular policy proposal has been demonstrated to be supported by consensus and otherwise complies with ICANN's articles, bylaws, and contractual obligations.
• Require all registries reflected in the ICANN root to undertake contractually to follow consensus-based policies. If a registry declines to do so, cooperate with an alternative root and its associated policy-making apparatus to transfer clear and exclusive policy control of such TLDs (and responsibility for operation of the root that includes them) to another party those registries will voluntarily join and support. If a new registry meeting reasonable minimum operational and technical standards, and operating a TLD that does not conflict with the operation of another TLD in the ICANN root, applies for admission to the ICANN root, admit it.
• Reaffirm at the Board level, and in ICANN's constitutional documents, that it may deal only with policies clearly related to the sound operation of the domain name system and that it may not impose its policies other than through voluntary contracts with those who undertake to abide by documented consensus among impacted parties.