Internet Architecture Board


Comment on ICANN ASO membership

Subject: Re: Letter from ETSI to ICANN Board re ASO
Date:    Fri, 13 Aug 1999
From:    "John C Klensin, for the Internet Architecture Board"
CC:      Mike Roberts <>, Mike Roberts <>,
         Esther Dyson <>,
         Geraldine Capdeboscq <>,
         George Conrades <>, Greg Crew <>,
         Frank Fitzsimmons <>, Professor Jun Marai <>,
         Eugenio Triana <>, Linda Wilson <>,, Hans Kraaijenbrink <>,,, ETSIBOARD@LIST.ETSI.FR,, OCG@LIST.ETSI.FR,,,,,,,,
         Scott Bradner <>,,

13 August 1999

Interim Board of Directors
ICANN – Internet Corporation For Assigned Names and Numbers
4676 Admiralty Way, Suite 330
Marina del Rey, CA 90292 USA


An Open Letter to the ICANN Board of Directors

The IAB would like to respond to the European Telecommunications Standards Institute’s (ETSI’s) letter commenting on the “Proposal for an MoU-based Address Supporting Organization (ASO)” submitted to ICANN by the Regional Internet Registries (RIRs).

From its beginnings, one of the key challenges for ICANN has been establishing a balance of perspectives that encompasses both (a) the perspective of technical and engineering principles of the Internet and (b) the perspectives of industry, the consumer base, and elements of public policy. The Supporting Organization structure is intended to provide that balance. As you know, the IETF has generally stood back from ICANN issues and Supporting Organizations other than those related to the PSO. We believe that our doing so is consistent with the ICANN Supporting Organization framework, in which each of the bodies that make up those organizations is assumed to bring its distinct perspective to ICANN. If every organization is represented everywhere, then the Supporting Organization model becomes largely useless.

While we have strongly and consistently supported ICANN and see strong IETF consensus for that support, there are still significant concerns in the IETF community about the risks associated with decisions which have technical or operational impacts being made without sufficient technical review and control. The original model for the PSO and Protocol Council assumed that the IETF would be the principal source of broad technical input to ICANN, input that would be based on the perspective of protocol developers and on high-level Internet operational experience. This has evolved into a model which includes a broader range of organizations who develop widely-deployed standards used on the Internet. The change is good and useful from the standpoint of openness and balance, but increases the risk of underrepresentation of the Internet’s technical core competencies in ICANN circles and hence the risk of destabilization.

Against this backdrop, ETSI’s letter requesting membership in the ASO is cause for some concern. The question of appropriate representation is best addressed within the community that allocates and uses these addresses: the IAB has no special competence to form an opinion in that area. However, we believe it necessary for us to comment on the broader issues, as follows:

  1. ETSI’s letter demonstrates a serious lack of understanding of Internet protocols and operations that calls into doubt not only the appropriateness of their membership in the ASO but even, to the extent that the PSO is seen as a technical resource for ICANN, ETSI’s role in the PSO and its Council.

    Specifically, we do not agree with ETSI’s hypothesis that the convergence of telecommunications networks and services with the Internet and consequent use of IPv6 and IPv6 addresses implies a need for “harmonization of numbering, naming, and addressing systems”. One of the key strengths of the Internet architecture is that it is designed to accommodate a variety of services with different naming systems, such as: the web with its URLs; email with its email addresses; FTP with with its domain names and ability to deal with different host file systems; and a variety of network numbering schemes, such as IEEE 802 LAN addresses, AppleTalk addresses, ATM NSAP addresses. The architecture accomodates these and others WITHOUT requiring the core IP addressing system to be “harmonized” with any of these service naming systems or network numbering schemes. This independence is accomplished through layers of abstraction and, in particular, identifier mapping mechanisms, such as the Domain Name System and the Address Resolution Protocol. These mapping mechanisms then allow IP addressing to be optimized around the needs of efficient IP packet routing and delivery, without unnecessary constraints based on geographic or political boundaries, and without limiting its future support of new kinds of service naming systems and new kinds of network numbering. This efficiency and adaptability is just as important for IPv6 as it has been for IPv4.

    This is, of course, not uniquely an IETF view. The advantages of the layering abstractions used in both the OSI Reference Model and its Internet parallels are well-established. We are unaware of anyone technically qualified and responsible who advocates the principle of abandoning them for an integrated system that requires application “addressing” to be sensitive to the structure of transport “addressing” and so on. Our impression is that the general directions at ITU, which has been ultimately responsible for the OSI layered model and E.164 addressing, still favor future work that maintains and expands upon clear layering.

    The numbering systems of the traditional telecommunications services and networks that fall within the areas of competence ETSI believes need to be incorporated in the ASO, can and should be accommodated in IPv6 just like the naming and numbering systems of any other services that run over IP and any other networks that carry IP. By intentional design of the Internet architecture, those naming and numbering systems have nothing to do with the IP addressing system that is the concern of the ASO.

    We expect IPv6 addresses to be organized on topological significance, and probably to change as topology changes, without explicit consideration of geographical significance. This is in marked contrast to E.164 addresses, where ETSI indicates their special competence lies. The technical expertise for both IPv4 and IPv6 address allocation lies with the existing RIRs and their customers who operate the world’s IP networks.

    While there is clear consensus on the importance of layering abstractions, the broader topic of relationships between the telephone system and the Internet is a matter of considerable debate, even within the telecommunications industry itself. Groups within a number of major industry players believe that the traditional telecommunications infrastructure will ultimately ride on top of an IP infrastructure, but will not interact with it directly. That model would be consistent with the long-standing design principle that the basic structure of the Internet is not altered to accommodate particular applications.

    In our opinion, it would be quite risky to the future of the Internet for ICANN to accidentally prejudge the answer to the complex technical questions involved in this issue. Those questions should be resolved in technical discussion among the subject-matter experts participating in the relevant standards development organizations. They should not be replaced by an administrative procedure derived from assumptions about the answer by one organization that is a relative newcomer to the Internet.

  2. ETSI’s expressed desire to be represented everywhere confounds the Supporting Organization model; granting ETSI’s desire is likely to undermine that model, weakening ICANN’s ability to obtain a balanced collection of perspectives on issues.

  3. ETSI’s argument is based in part on the not-yet-launched “3GPP Phase One”. If ETSI should be part of the ASO based on that argument, all other bodies working in this same area should also be admitted. Of course, this is not the only significant deployment future for Internet-based communications-based services. Should every body within every industry sector that is contemplating or developing a large scale deployment of Internet-based consumer electronics also exert influence directly in ICANN through comprehensive membership in every Supporting Organization? ICANN’s structure will not readily accommodate such a model of competition in the exercise of influence.

We ask that ICANN consider the above points when the creation of the ASO is considered, as well as in reviews of the structure and functioning of other Supporting Organizations.

Thank you for your consideration,

John C Klensin
For the Internet Architecture Board (IAB)