Version classiqueVersion mobile

Normative Experience in Internet Politics

Françoise Massit-Folléa
Cécile Méadel
Laurence Monnoyer-Smith

Chapter 3. Some comments on the Institutionalization of the Net: ITU or ICANN, is there an ideal solution?

Herbert Burkert

Texte intégral


  • 1 American Society of International Law: Multipartite International Instruments of Legislative Effec (...)

1As the development of the European railway networks has taught us, infrastructural networks have the inherent tendency to branch out until they finally reach the borders of their authorized territory. They then pose the border question: under which conditions and with which consequences are they to be linked with the infrastructural network on the other side? European landscapes still show traces of unfinished railway tracks, of bridges ending in the void, all witnessing unresolved border problems and failed connectivity efforts. Border problems reach another dimension when we move from two neighboring territories to the topology of three neighboring territories, a move that introduces the problem of transit. Transits need more complex solutions than simple border crossings. Crossings may be handled bilaterally. Transits invite multilateral regimes. This rule not only applies to railway infrastructures, it also applies to communication infrastructures, and it affects them with greater speed, since messaging infrastructure require less resources. It comes as no surprise then that the transit issue had been one of the main driving forces to install the first international organization ever, and that this organization dealt with messaging: the International Telegraph Union established in 1865, more than twenty years before a similar undertaking for the international railway infrastructure proved to be successful (May 15 1886 – Convention Concerning the Technical Unity of Railways 18861)

  • 2 For more details about the history of the ITU: [Tagged, 1994].

2Looking at the history of the International Telegraph Union2 one is struck by similarities with the Internet Corporation for Assigned Names and Numbers (ICANN): Both are hybrid constructs combining public sector and private sector actors. The ITU from its very beginnings admitted private sector parties, telegraph companies, although not as voting members but still as part of the institutional construct.

3Like ICANN the ITU from its beginnings had representation problems: The United States pointed out – some sixty years after the foundation of the ITU – that some European member states had been using their voting power in that organization unfairly by giving separate membership and voting rights to their own colonies. The United States, in the 1930s, seeing themselves proudly as a non-colonial power, made it a condition for their accession that this practice was to be changed. And, finally, the ITU – as it is occasionally suspected of ICANN – was effectively run by one national government: It was the Swiss federal government department responsible for telegraphy and later also telephony that was acting as ITU’s secretariat.


  • 3 See for example [von Arx, 2003] and [Palfrey, 2004].
  • 4 See for example [Huston, 2005].

4Taking into account these historical experiences and looking at the increasingly successful past of ITU, it is surprising that an organization like ICANN does exist at all. It had seemed almost natural that the ITU would absorb all technical questions arising from a specific addressing system of a particular, even if peculiar network such as the emerging Internet. Now a status quo exists between both organizations that no one wants to upset. The reasons for ITU’s «failure» have been discussed in detail elsewhere.3 There has also been a reluctance of the US Government to let – so to speak – its own invention, the Internet, wander off into a regulatory setting that was seen as lacking the necessary adaptive speed and flexibility, and that relied on a policy setting and on rule making mechanisms that were seen to be too strongly influenced by political considerations outside telecommunications proper.4 From the ITU’s perspective, conceptual issues have to be considered as well: The numbering system of the Internet had switched very quickly to a semantic system, dealing with names on its top layer. The semantic connotations of these names were organizations, business names, trademarks, all of them evoking associations quite different from message transport. And the ITU – while dealing indeed with infrastructural questions of networks – had regarded the Internet as a sort of private network issue, a sort of experimental in-house network that could be left alone.

  • 5 See [Rutkowski, 2008].

5However, many important decisions that also affect the semantic and pragmatic level of the Internet still happen and will continue to happen in the context of the ITU. Among them, the most important current issue is in my view the installation or rather the recommendation of installing IP traceback mechanisms, i.e. technical systems that will make it possible to trace any message, any package on the Internet to the IP number of its origin, regardless of whether proxy servers or other means of anonymization are being used in the transport.5 Work on this proposal, which has jointly been entered into the appropriate committee process of the ITU by representatives of the United States and of China, has not yet been concluded.


6Now we are facing a situation where, at last, ITU and ICANN may compete over public policy issues, the ITU using the keyword “convergence” as its legitimate reason to look into public policy issues. ICANN, of course, using the catch phrase of “the stability of the network” sees itself equally entitled to look into any related public policy field.

7But there are now also other institutional competitors dealing with public policy issues of the Net, such as, for example, the UN fora which have been developing around and with the World Summit on the Information Society altogether creating a complex structure of multi-player, multi-interest, multi-process, multi-layer interactions.

8This situation contrasts with a sort of ideal-type single international institution addressing all issues relating to the Internet and including all relevant stakeholders, applying consensual processes of discussion and decision-making which would then serve as guidance to national and regional policy makers. The lack of such an institution seems to be one of the reasons for the dissatisfaction with the current Internet policy-making process.

9The concept of such an ideal-type international institution, that idea to design the institution to end all fragmentation in Internet policies, had played an important role for many stakeholders in participating in the WSIS process with the hope that in the end one such UN-like institution would emerge. This concept of the one due-process organization has, I think, developed with general political change over the past thirty years: the end of the Cold War, the hopes associated with that end, the increasing role of Non-Governmental Organizations, the increasing importance of international institutions as a medium for national policies, and the ascent of “democracy” as a sort of functional benchmark requirement for international institutions.

10If – in a thought experiment – we were to carry on to speculate about such an organization, what would be its requirements? I want to introduce into the experiment seven – rather formal – requirements usually mentioned in discussions about such ideal-type settings, listed here not necessarily according to their importance:

  1. An international institution (for the Internet) should be the result of a proper intentional process: The institution should not just happen to be there, or evolve from an existing organization by slowly extending its areas of competence. Ideally, the institution should be created by an international treaty.
  2. This kind of institution should have a clear focus and properly delimited subject areas it is dealing with.
  3. The institution should be representative, i.e. the design of the organization and of its decision-making processes should be representative of its constituency.
  4. Such an institution should be inclusive. Inclusiveness means that while the institution may indeed be limited in its representation to those who have created it and who form its membership, it should be aware that it will address issues which involve other stakeholders not represented by this institution. To address this “over-spill” there must be a process in place that provides a mechanism for inclusion.
  5. The fifth requirement is independence. The institution should provide guarantees that it will not fall under the hegemony of one or several of its members. Nor should it make itself dependent on one single source of revenue.
  6. Transparency and information: The institution has to have enforceable and effective mechanisms in place that ensure passive transparency (accessibility) and active transparency (dissemination), for documents and meetings, as well as for structural and procedural information.
  7. The institution should provide for mechanisms of accountability, responsibility, review and redress.

11If we apply these criteria to the ITU or to ICANN, we realize a series of shortcomings. With the regard to the ITU, I just restrict my critique to one element which is interesting because one is not quite sure if on this occasion the ITU may have intended to copy ICANN. The ITU, from its beginning, had two types of membership: one is the member country membership, the national states represented by their governments; and the other is the telecommunications companies membership. This division still exists: there are the members States, and there are what is called the sectorial members. Since the 1990s, ITU has introduced a new type of membership, the “associate membership”: to become an associate member, all you have to do is pay for your membership. This membership has become a new supplementary form of financing for the ITU. This new type of membership (or rather sponsorship in this case), however, raises a problem: what is it that the ITU has to offer to such members? The ITU can offer information. Since its committees are the forum for technical recommendations and the initiator of some standard policies of the future, it is able to provide trend information at the earliest stage to assess into which direction certain technical fields and developments will go. But this information must have an exclusive value to be attractive. So, in order to make the associate membership more attractive, the ITU has gone through its information resources and has declared whole sets of information to be confidential for the sole and only reason that now it can sell them exclusively to the associate members. Such a policy seems to be in contrast to the criteria of an appropriate information policy and transparency, and would shed some doubt on the true interpretation of financial independence.

12With regard to ICANN, five criteria from our ideal-type list immediately come to my mind as being problematic. How did ICANN come about? – clearly not as the result of a careful international agreement. What is the actual focus of ICANN? That is simply one of its problems. What about representation, inclusiveness and independence? Here we can simply point to the repetitive criticism that occurs with each debate on the further prolongation of its mandate.


13Our brief analysis has started with an observation of fluidity, which prompted us to think about criteria for a clearer institutional framework as a new vessel for Internet policy, making it possible to overcome this problem of fluidity. Using the criteria with respect to already existing candidates for such an institutional revival, ITU and ICANN, did not help to alleviate our doubts about the feasibility of building the international organization for the Internet from existing institutions.

14In the current situation of a transition of hegemonies, we might also choose to take a different approach and – again in a kind of thought experiment – simply accept fluidity as a given thing, adapt to its “multi-factors-games”, and look for proper regulatory structures. After all, fluidity does have a number of advantages in comparison to a somewhat rigid, ideal-type monolithic international institution. Fluidity and multi-level-structures for conflict resolution may help to avoid stagnation and even to unblock – apparently unsolvable – confrontations, by the ease with which foray and participants may be exchanged in order to be able to continue the debate. I also tend to think that this fluid structure is more attentive and sensitive to signs of change, and it generally increases choice. But it still creates a heavy burden for the management of these kind of games.


15At this point I think I can end my short comments with one main conclusion and a strategic outlook.

16Ideal-type institution building is perhaps too traditional an approach to solve the governance problems of the Internet. We may use, and do use, the criteria for ideal institutions as a tool kit for criticizing current processes. A constructive approach, however, cannot solely rely on such a tool kit, and may have to accept fluid-border and borderless concepts and settle for efforts to at least optimize their management. Such an optimization could build on four mutually supporting strategies:

  1. Make arrangements for your own multiple presences and the multiple presences of others. It is extremely risky to leave out any of the potential players.
  2. Multiple presences must be constructed in a way that they facilitate cross-institutional coalition building. Coalition building helps to compensate for the costs of multi-presence. There is already a lot of practice and experience in the governing processes of the European Union, where member states build coalitions in forum A to be able to use this coalition in forum B and again in forum C.
  3. You have to be creative in institution-building yourself. That is, creating or suggesting the creation of international institutions has to become an integral part of a national policy. There are indeed some countries which already excel in this policy, although they are not always successful with their suggestions in the long run. Institution invention can always be used strategically to enlarge options or tactically to switch foray.
  4. In the game of multi-presence and institution building, the criteria we mentioned above can always be used for delegitimizing inappropriate uses of multi-presence and institution building.

17Ultimately then those ideal-type criteria may prove the only long-term guidance for a sustainable institution building in future unstable environments of the Net.


1 American Society of International Law: Multipartite International Instruments of Legislative Effect. In: The American Journal of International Law, Vol. 22, No. 2, Supplement: Official Documents (Apr., 1928), pp. 90-108, 92.

2 For more details about the history of the ITU: [Tagged, 1994].

3 See for example [von Arx, 2003] and [Palfrey, 2004].

4 See for example [Huston, 2005].

5 See [Rutkowski, 2008].


Public Law Professor, Research Center for Information Law, University of Saint Gallen, Switzerland. International Fellow, Information Society Project, Yale Law School (USA). Senior Research Fellow, Fraunhofer Institute (Germany).

© Presses des Mines, 2012

Licence OpenEdition Books

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search