Update Regarding RIR Submissions
16 September 2002

Committee on ICANN Evolution and Reform
Update Regarding RIR Submissions

On September 13, the three existing regional Internet registries (RIRs) – APNIC, ARIN, and RIPE NCC – posted a statement to the ERC comment list stating that they cannot support the ERC's proposed changes to the composition of the ASO, the proposed changes to the support mechanisms of the ASO, and the proposed changes to the role of the ASO within ICANN because the ERC has not been responsive to their earlier postings. The ERC is concerned that the RIRs have gained that impression because the ERC, while considering comments from all sources, has not explicitly noted in its Interim Implementation Reports each point of agreement or disagreement with each individual comment. We have paid careful attention to RIR submissions and statements (and the numerous private conversations with members of the RIR staffs and the addressing community) over the past several months – and indeed have incorporated some of the suggestions in our recommendations and other postings – even though we have not responded point by point to the RIR submissions any more than we have to those from other sources.

Our failure to respond individually to every comment is not due to inattention or disrespect; the full bandwidth of the committee (and then some) has been dedicated to reviewing and incorporating the large number of comments that have been submitted. We regret, but cannot honestly apologize for, the fact that we have not been able to also prepare individual responses to each comment.

Nevertheless, it is clear that a continuing dialog with the RIRs is important to ensure that there are no communication misunderstandings as we proceed with the complex process of reform. We believe that both informal and formal exchanges are valuable, and are always willing to listen to concerns expressed by any members of the community.

Background

The 13 September posting from the RIRs followed presentations at the APNIC and RIPE meetings over the preceding two weeks at which the same concerns were expressed. In particular, those presentations expressed disappointment that the ERC had not directly and explicitly responded, in its 2 September Interim Implementation Report, to the views offered by the RIRs in their "Regional Internet Registries' Submission to the Committee on ICANN Evolution and Reform" (20 June) and to an earlier Statement of Principles posted by the RIRs on their websites on 8 May. We recognize that the RIRs' concerns are both procedural – due to the lack of a formal, written response from the ERC to comments submitted by the RIRs – and also substantive – the RIRs' dissatisfaction with the decision by the ERC not to include all of their recommendations in its Implementation Reports. The procedural concern is noted above; the remainder of this commentary deals with the substantive issues on which the RIRs and the ERC disagree.

Comments on RIR Proposals

As has been the case with contributions from all ICANN stakeholders, we have found some of the arguments advanced by the RIR community to be persuasive, and have incorporated their recommendations into the Interim Implementation Report; others we have found to be less persuasive.

Some of the changes that have occurred as a result of listening to the RIRs and the addressing community are detailed in Annex 1. The original Lynn proposal advocated replacing the ASO by an Addressing and Numbering Policy Council. It would have applied many changes uniformly to all supporting organizations. It also advocated a Technical Advisory Committee (TAC) that would include the RIRs and that would act as a committee in giving technical advice. Both informal feedback and formal submissions – many of them from the RIRs and the addressing community – have persuaded the ERC and the ICANN Board that, since the ASO has functioned well as conceived, it would be a mistake to replace it by a different supporting organization, or to treat all supporting organizations uniformly. The ERC and the Board have also substantially changed their views on the TAC based on this feedback. It should also be clear at this point that the ERC firmly agrees with the RIRs that the ASO is the ICANN body responsible for developing (as distinct from approving) policies with respect to addressing.

In three areas, however, the RIRs have proposed changes to the current structure and practice with which the ERC does not agree and therefore has recommended no changes. In essence, the existing RIRs: (1) advocate transferring the responsibility for formally approving global policies relating to addressing from the ICANN Board to the Address Council of the ASO, limiting the ICANN Board to the role of ensuring that the RIRs and ASO adhere to their stated procedures in the formal adoption of policies; (2) suggest that the responsibility for recognizing new RIRs be transferred from the ICANN Board to the ASO; and (3) recommend that the IANA's role of allocating blocks of IP addresses to the RIRs according to their demonstrated needs – while reserving unallocated blocks for use of future users and operators of the Internet and for common purposes – be transferred to the RIRs acting collectively.

We believe that these proposals are inconsistent with the basic premise that led to ICANN's creation and motivates its current reform efforts, as expressed in the mission statement and core values that have been developed. This is why we have not included these particular RIR proposals in our Implementation Reports and recommendations to the ICANN Board. The specific reasons for our disagreement on each specific issue are as follows:

1. Global Address Policy Responsibility

The issue here is global address policies, not regional policies. Regional address policies should of course remain the responsibility of the individual RIRs; but global address policies (along with comparable global policies related to names) are at the heart of the rationale for creating ICANN in the first place. The requirement for a fully open and transparent global consensus-development body to coordinate global policy, with stakeholder participation reflecting the geographic and functional diversity of the Internet, is as valid today as it was then. The RIRs are not, individually or collectively, suitably structured to receive, evaluate, and develop consensus positions on those address policies that truly take into account all interests affected by global concerns, although they are in a very strong position to propose and recommend such policies. The RIRs were not created with a broadly based policy-development process in mind. Most importantly, they are membership organizations with the duty to serve their (primarily ISP) member interests; they are not structured to encourage open participation by all the affected stakeholders, ranging from individual and business users to national governments. To change this would essentially require replication of ICANN at the RIR level, which seems duplicative and wasteful.

Our experience over the last four years, and the current reform efforts, demonstrate clearly that there is a global community that has legitimate interests in, and should be heard on, those policy issues that affect the functionally and geographically diverse stakeholders in the global Internet. This is not a criticism of the RIRs individually or collectively; they are in fact very efficient at what they were established to do (to provide numbering resources to their members and to establish regional policies for doing that). However, some additional structure must exist to develop and establish global policies that reflects the aggregate of community interests more broadly and inclusively than the collective RIRs.

Today, the ASO is a critical element in that structure, and it performs very well. But it is not the whole story. We do not believe that the ASO can operate effectively as the ultimate and final decision-maker on global address policies. The broader ICANN, not just the ASO, is the appropriate structure for final determinations on global addressing policy, just as it is for global naming policies. Ideally – and indeed as has always been the case to date – this amounts in practice to what the RIR statements refer to as "ratification": the straightforward administrative approval of a policy developed and recommended by the ASO. In the unlikely event of substantive disagreement with an ASO-developed global policy recommendation (possibly involving disagreement among the RIRs), having the necessary discussion and developing a broad consensus is what ICANN is designed to do. Absent recreating ICANN within the ASO, it is hard to see any plausible approach that ensures that all interested stakeholders have an adequate opportunity to participate in the process.

From our perspective, this debate involves both semantics and substance. It is semantic to the extent that we expect the ASO to continue to develop and propose global policies that in fact have or will receive broad community support, in which case the ICANN Board will simply accept (ratify) the proposal. But it is substantive to the extent that if, for whatever reason, the ASO process does not successfully resolve serious and widely-held community concerns, the ICANN Board – selected as it will be by a broad cross-section of all ICANN stakeholders – must be the place where all those views, including but not limited to those expressed by the RIRs or the ASO, are reconciled.

2. Recognition of New RIRs

One of the responsibilities historically performed by the IANA and assigned to ICANN by the ASO MOU (section 1(c)) is the recognition of new RIRs. This process is performed according to standards that have been jointly agreed to by the existing RIRs, the ASO, and the ICANN Board. ICANN provides a neutral balance in ensuring that new RIRs are recognized based on technical merit and need, and that the decision concerning the recognition of a new RIR is not compromised by the unavoidable and natural self-interest of the existing RIRs.

This process has worked well so far, with the ASO formulating (and the ICANN Board accepting) a detailed statement of "Criteria for Establishment of New Regional Internet Registries" (ICP-2). At the March 2002 ICANN meeting in Accra, the ICANN Board provisionally recognized LACNIC under these criteria. It is expected that the LACNIC recognition process will be completed by the October 2002 ICANN meeting in Shanghai. The recognition process for a fifth RIR, AfriNIC, is also underway, using the same ICP-2 criteria.

The existing RIRs propose that the responsibility for recognizing new RIRs be transferred from the ICANN Board to the Address Council, which is formed from voting members designated by the existing RIRs. Although under the RIR proposal, ICANN would have the role of oversight through review and reconsideration of the decisions in the event of a dispute, placing the decision-making role in a body dominated by the existing RIRs is not likely to inspire confidence in the neutrality of the process. Under the present process, of course, existing RIRs can provide recommendations to the ICANN Board, and their comments have been helpful to the recognition process. But the role of ICANN in facilitating the process – which after all involves the existing RIRs allowing a new entity to assume some of the responsibilities they now carry – has proven to be useful. Diminishing the apparent neutrality by which recognition decisions are made would frustrate the need for a clearly unbiased process and may unnecessarily raise concerns of competition authorities.

The existing process, agreed to by the addressing community and ICANN, worked well with respect to the LACNIC recognition process, and we see no reason for changing it, especially given the potential negative effects noted above.

3. The IANA Address-Allocation Operational Responsibilities

This is one of the historic core functions of the IANA, comparable to the determination of which domains are included in the authoritative root. It involves receiving RIR requests for allocations of numbering resources (blocks of IP addresses and Autonomous System numbers) and verifying whether additional allocations to a particular RIR are appropriate. To ensure that numbering resources are optimally available on a global basis, for both present uses and into the future, it involves a review of RIR-provided usage statistics to ensure that resources are allocated to a particular RIR's region based on demonstrated present and near-term need. A process for independently verifying the RIRs' demonstrations of need is essential to ensure that top-level allocations are made only as required for present uses, and that the remaining pool of IP addresses is conserved so that it is available to support future users and uses as the Internet evolves. It is particularly important in light of the growing possibility of legal challenges to the fairness of allocations.

Ordinarily, this process is largely mechanical: an RIR makes a request, it is justified, and additional allocations are made. The issue here, however, is similar to that with respect to global policies: where does the ultimate responsibility (and accountability) lie? This question has particular relevance in the unusual circumstance in which an allocation request raises substantive questions that cannot be resolved mechanically. The RIR position, as we understand it, is that they should collectively determine whether and how such allocations will be made, but to do so they would have to engage in some form of collective action. Their position is that they would prefer the necessary collective action to be limited to the RIRs rather than accountable to the entire global community of interests. We believe that ICANN was established to ensure that the interests of the entire global community, including but not limited to the RIRs and their present members, are represented in collective actions with global implications.

The key point is that the IANA allocates address space based on demonstrated need by the region, just as the RIRs do in allocating address space to their members. This is a linchpin of global and regional policy. Removing ICANN's role in protecting this policy at the global level would open the door to the policy becoming vulnerable at all levels. In addition, the RIR argument that the IANA's involvement creates a cumbersome two-stage process is hard to understand. Whatever process might be created by the RIRs themselves to accomplish allocations among regions (replacing the IANA's role, but with accountability to a narrower set of interests) would, of necessity, involve a global stage of allocation that would be followed by an intra-regional allocation stage.

Thus, the issue is whether the long-standing procedures for initial allocation of number blocks to RIRs should continue, where it is done through an objective test establishing the justification for the allocation, or whether that function should be eliminated, leaving to the RIRs themselves the discretion to make number allocations as they deem appropriate, without any review by the broader Internet community. We see no value to the proposed change, and real value to continuation of the current arrangement.

* * *

In sum, in these substantive points of disagreement, we believe that an appropriately structured ASO principally responsible for developing positions on global address-allocation matters, with the opportunity for informed input by all affected within the ICANN process prior to approval and implementation, will provide significant support to the private-sector-led approach for coordination of Internet addressing allocations. As has been observed in numerous instances already, notably in the case of domain-name policy, the structural balance and breadth of composition of ICANN, operating in the context of an appropriately focused mission, can provide a significant and necessary counterbalance to other approaches.

Committee on ICANN Evolution and Reform
16 September 2002

Annex 1: Significant Adjustments to ICANN Evolution and Reform Proposals Based on Inputs from RIRs and Members of the Address Community

In understanding the important contributions that the RIRs and members of the addressing communities have made to the ICANN evolution and reform process, it is useful to summarize the many significant changes in the course of reform that have resulted from their comments and analyses. The differences between the original February 2002 Lynn proposal for reform and the current proposals of the ERC that affect the RIRs and the ASO illustrate the way in which the ERC's thinking has evolved in response to comments (both formal and informal) from the RIRs and other members of the community.

• Lynn Proposal advocated three supporting organizations, one of which was an Addressing and Numbering Policy Council (ANPC). Based on persuasive arguments from the RIRs and the addressing and technical communities, the Blueprint advocates retaining the ASO as is.
  
  
• Lynn Proposal advocated combining the functions of the PSO and the ASO into a single body. In response to RIR and addressing/technical community input, the Blueprint retains the ASO and replaces the PSO in part by a Technical Advisory Committee with a structure different from that which was advocated by Lynn (see below).
  
  
• Lynn Proposal advocated an ANPC populated by designees of the IETF and those RIRs with whom ICANN had formal agreements, and by the NomCom process. The RIRs and members of the addressing community expressed the view that addressing policy recommendations should be formulated by an organization drawn only from the addressing community. The Blueprint eliminated any NomCom appointees to the ASO Council (and, of course, there are no IETF-appointed members).
  
  
• Lynn Proposal advocated that the Chair of the ANPC Steering Committee be ex-officio an ICANN Board Director. The RIRs and the addressing community responded that this could result in ICANN Board decisions that did not sufficiently reflect their concerns. The Blueprint provides for two ICANN Directors selected, in the same way as now, by the ASO Council.
  
  
• Lynn Proposal advocated that all three supporting organizations (Policy Councils) be treated uniformly as far as structure, funding, staff support, etc. are concerned. Comments from the RIRs and the addressing communities pointed out that the different Councils faced vastly different circumstances, and that these differences should be recognized. The Blueprint advocates that these different circumstances be recognized, and that each Council be considered separately.
  
  
• Lynn Proposal advocated applying the same policy-development process to all three supporting organizations. In line with the RIR and addressing community comment that uniform approaches are not optimal, the Blueprint does not require this. (For example, the Policy Development Process described in the ERC September 2 Report applies only to the GNSO.)
  
  
• Lynn Proposal advocated a Technical Advisory Committee (TAC) that includes the RIRs. Based on RIR comments, the Blueprint does not include the RIRs in the TAC.
  
  
• ERC May 31 posting floated the idea of a TAC that might have a broader role in oversight of certain IANA operational functions. Drawing on input from the technical and addressing communities and the RIRs, the Blueprint discarded this idea, reverting to the original Lynn Proposal of a TAC with a purely advisory (non-operational) role. Based on comments highlighting the importance of this limitation on the TAC's scope, the ERC September 2 Report included explicit charter limitations on the TAC to ensure that it has no operational responsibilities. Under the proposed charter, the TAC does not provide advice as a body, but acts as a conduit to the most expert technical advice available.
  
  
• Lynn Proposal did not address the mission of ICANN, but assumed the current mission. Many in the addressing and technical communities indicated that they believed a clarified and clearly limited ICANN mission was a matter of first priority. The Blueprint and subsequent ERC documents reflect a thorough review and refinement of ICANN's Mission and Core Values that clarifies ICANN's limited mission.

Page Updated 16-Sep-2002
©2002 The Internet Corporation for Assigned Names and Numbers. All rights reserved.