.org Reassignment: General Counsel's Report on Proposals for Reassignment of the .Org Registry

Posted: 22 September 2002

.org Reassignment: General Counsel's Report on Proposals for Reassignment of the .Org Registry

Prepared by: Louis Touton
ICANN General Counsel
22 September 2002

On 19 August 2002, I submitted a document entitled "Evaluation of Proposals Under Procedural Criteria (#2, #3, and #10)", as a contribution to ICANN's evaluation of the eleven proposals that it received from entities seeking to become the successor operator of the .org registry when VeriSign gives up that role at the end of 2002. My evaluation was posted in conjunction with the Preliminary Staff Report on Evaluation of the Proposals for Reassignment of the .Org Registry. That preliminary evaluation and supporting documents were posted for comment by the applicants and the public; the resulting comments of applicants and the public have also been posted.

This General Counsel's Report is a revision and supplementation of my earlier report, which it is intended to replace. The comments received have indicated several elaborations and adjustments that should be made in my earlier report. They have also pointed out some additional subjects on which my analysis and comment is appropriate.

The first three parts of this Report set forth my revised analysis of criteria 2 (Ability to Comply with ICANN-Developed Policies), 3 (Enhancement of Competition) and 10 (VeriSign Endowment). The fourth part discusses criterion 11 (Completeness of Proposals/Realistic Plans). The fifth part whether Board members have a conflict of interest concerning the reassignment of .org arising from their involvement with the Internet Society.

A. Evaluation Under Criterion 2 (Ability to Comply with ICANN-Developed Policies)

Criterion 2 of the published Criteria document states:

2. Ability to comply with ICANN-developed policies.

As a globally open TLD, the operation of the .org registry must comply with policies defined through ICANN processes, such as policies regarding registrar accreditation, shared registry access, the uniform dispute resolution policy, and access to registration contact data via Whois. Consideration will be given to the adequacy of mechanisms proposed for ensuring compliance with those policies.

Questions relevant to this criterion include C17, C17.1, C17.2, C17.3, C17.7, C17.8, C17.11, C17.12, C17.13, C17.16, C19, C21, C22, and C28.

This criterion is largely a verification step that all bidders are expected to meet. Because of this, I have analyzed criterion 2 with respect to only the proposals likely to be selected based on the overall evaluation, including the evaluations conducted by the Gartner and Usage Evaluation Teams. Consistent with the recommendation in the Final Staff Evaluation Report that only proposals ranked by the Gartner evaluation in its highest tier should be considered "Leading Proposals", I focus my analysis on the ability of ISOC, NeuStar, GNR, DotOrg, and Register.org to comply with ICANN-developed policies. Should an issue later arise concerning another applicant’s ability to comply with ICANN-developed policies, that analysis can be performed at that time.

The principal vehicles to implement policies in the operation of a TLD registry are entry and enforcement of Registry Agreements between ICANN and the registry operator. As part of the Request for Proposals, ICANN published a Model .org Registry Agreement, which details the registry operator's obligations to implement ICANN-developed policies. Each applicant was advised in the instructions:

If your application is selected, you will be required to enter into a registry agreement with ICANN concerning your operation of the .org registry. The model form of the agreement is posted on the ICANN web site; please note that several appendices must be prepared and agreed between you and ICANN based on the details of your proposal. If agreement is not reached on the registry agreement and its appendices, ICANN reserves the right to make another selection of the organization that will become the successor operator.

Thus, the model agreement (already drafted) and its appendices (some already drafted and others to be prepared to reflect the terms of the selected proposal) provide the main mechanism by which the selected registry operator will be obligated to implement ICANN-developed policies.

In addition to the obligation to implement policies, however, it is important that proposals demonstrate the capability and disposition to implement them. The following proposals, which were placed in tier "A" by the Gartner team, were reviewed to determine whether they propose adequate mechanisms to promote voluntary compliance with ICANN policies, and otherwise indicate the bidder’s capability and disposition to comply:

• DotOrg Foundation
• GNR
• ISOC
• NeuStar
• RegisterORG

Based on the proposals (particularly items C19, C21, C22, and C28), each of the bidders propose mechanisms (such as dedicated compliance personnel) that should afford them the capability to comply with ICANN policies and each reaffirm their commitment to abide by ICANN policies.

Three of the five proposals (DotOrg, ISOC, and RegisterORG) listed above involve a back-end provider that has an ownership relationship with one or more registrars. This circumstance requires particular attention to compliance with ICANN requirements that all registrars receive equivalent registry access. The back-end providers involved, however, have demonstrated their ability to address these circumstances in connection with TLDs they already operate.

In its comments to my earlier report, NeuStar noted that the registry operator proposed in the ISOC bid will be a yet-to-be-formed not-for-profit company known as the "Public Interest Registry" or simply "PIR." NeuStar argues that the fact that PIR has not yet been formed makes the above analysis under Criterion 2 inadequate:

[T]he Report simply notes that it is sufficient that the new registry operator operate under a TLD agreement with ICANN and that the back-end registry provider has proven its ability to provide equal access in the past. For those applicants with existing registry operators, this analysis, NeuStar agrees, is likely sufficient. It is insufficient, however, with respect to ISOC.

. . . . Unfortunately, PIR does not yet exist and ISOC maintains that it will be run entirely separately from ISOC. Although ISOC will not have any significant role in the registry, its proposal relied heavily on its own credentials and qualifications for its response to critical questions including the adequacy of its mechanisms for compliance with policy. Since ICANN states in its proposal that PIR, and not ISOC, will be responsible for Policy, there is no way for ICANN to accurately assess whether the entity will have mechanisms in place to ensure compliance. It is unclear even who the bidder is given that ISOC will have only a limited role in .org following formation of PIR. Therefore, the ISOC application cannot be measured against this criterion and ICANN Staff must reconsider their evaluation.

This argument does not withstand careful analysis. The ISOC proposal describes with significant detail the mechanisms that will be in place to ensure PIR's compliance with ICANN-developed policies. As ISOC notes, PIR "will enter into the registry management agreement with ICANN, and PIR will contract with Afilias for all back-end registry services." Thus, it is clear that the back-end provider will be Afilias, an organization that is quite familiar (due to its operation of .info) with the steps required to comply with ICANN policies. As the ISOC proposal notes in Item C19, "PIR's agreement with Afilias will call for Afilias to implement without fail all ICANN policies and requirements of the .ORG Registry Agreement." As explained in the ISOC proposal, Afilias will devote the efforts of its full-time compliance officer, as well as its periodic reporting infrastructure, to ensure that PIR will be able to comply with ICANN-developed policies.

Indeed, the particular policies highlighted in criterion 2 ("registrar accreditation, shared registry access, the uniform dispute resolution policy, and access to registration contact data via Whois") are all ones with which Afilias currently complies. The technical plan in the ISOC proposal describes at great length the technical mechanisms, currently used in .info and proposed for .org, to support compliance with ICANN-developed policies. Item C17.8 describes a thick Whois system fully capable of meeting the requirements of ICANN-developed policies. Item C21 of the ISOC proposal provides a sound plan for ensuring equivalent access for all ICANN-accredited registrars. Afilias has gained considerable experience in compliance with the uniform dispute resolution policy. In view of the detailed compliance mechanisms described in the ISOC proposal and already tested by use with .info, there is no reason to doubt that the PIR/Afilias combination will not be fully capable of complying as well.

In sum, I conclude that all five of the “Leading Proposals” demonstrate compliance with criterion 2.

B. Evaluation Under Criterion 3 (Enhancement of Competition)

Criterion 3 of the published Criteria document states:

3. Enhancement of competition for registration services.

One of ICANN's core principles is the encouragement of competition in the provision of registration services at both the registry and registrar levels. Promotion of that principle will be a criterion. As one illustration of this criterion, a major purpose of the reassignment of the .org registry is to diversify the provision of registry services by placing the .org registry under different operation than the .com and .net registries. Consideration will be given to the extent to which proposed arrangements are consistent with this purpose. As another illustration, applicants are encouraged to refrain from prohibiting non-affiliated providers of backend services from offering their services in connection with other applications.

Questions relevant to this criterion include C2, C8, C12, C13, C13.1, C13.2, C13.3, C13.4, C13.5, C13.6, C14, C21, C22, C25, C26, C27, C31, C32, and C33.

As is apparent from the text supporting criterion 3 quoted above, the reassignment of .org is in general a pro-competitive event, though the pro-competitive benefits might be mitigated by various features of particular proposals. (Note: Various competitive effects may also result from factors such as the introduction of innovative services and reductions in pricing. Those factors are addressed in criterion 7, which was evaluated by the technical teams.)

The text above gives two illustrations of proposal features that tend to diminish the pro-competitive benefits of reassignment:

• Where the proposal does not diversify the provision of registry services by placing the .org registry under different operation than the .com and .net registries.
• Where a proposal is shown to have prohibited a back-end provider from providing support for other proposals.

There appears to be no need to analyze the second bullet in detail in this report. With the possible exception of a restrictive arrangement between UIA and VeriSign (that restriction does not affect the outcome of this analysis of criterion 3), there appears to be no evidence of lock-ups of the type stated in the second bullet.

With regard to the first bullet, only three of the eleven proposals received appear to involve potentially relevant relationships with the operator of the .com and .net TLDs (VeriSign, Inc.): Global Name Registry; ISOC; and UIA. The following analyzes competitive considerations arising from these relationships:

Global Name Registry has a two-fold relationship with VeriSign. First, VeriSign Capital Management, Inc. is a minority (less than 10%) investor in GNR’s parent holding company. See GNR Proposal, item C33(b). VeriSign does not occupy a board seat nor is there any indication that its minority ownership gives it control over GNR’s operations. In these circumstances, it does not appear that this investment relationship undercuts the competitive benefits of reassignment of .org.

In addition to its minority ownership by VeriSign, GNR has entered an arrangement under which VeriSign provides principal registry services for .name, the TLD that GNR currently operates. This arrangement was implemented at the time that GNR went from batch-mode to real-time shared registry operations. In its .org proposal, GNR’s proposal pledges as follows:

Assuming a winning bid, aside from its role in transitioning the .org registry to Global Name Registry, Global Name Registry will not contract its backend services to VeriSign in connection with the continued operation and maintenance of the .org registry. Clearly, this type of arrangement would contravene the intent and spirit of the entire .org divestiture.

GNR Proposal, item C33(b). This pledge appears fully satisfactory to maintain the pro-competitive benefits of reassignment of .org and, if GNR is selected, it should be included in the Additional Covenants (Appendix W) of the Registry Agreement.

It should be noted, however, that an additional consequence of this pledge is that GNR does not propose to use the technology (that introduced through GNR’s collaboration with VeriSign’s) which has been used to provide real-time shared registration services for the .name TLD. Instead, GNR proposes to use technology it has developed and will develop without VeriSign’s collaboration.

This concern was noted in my earlier report, and subsequently Supplemental Question No. 6 was posed to GNR to obtain more information about its use of VeriSign technology. GNR’s response to that supplemental question, together with the timeline it provided in response to Supplemental Question No. 10, indicates the following:

• GNR originally intended to use technology it would develop to implement the .name registry.
• GNR began developing all systems needed to operate the .name registry.
• GNR entered its registry agreement for .name with ICANN on 1 August 2001.
• In September 2001, GNR entered an outsourcing agreement with VeriSign Global Registry Services.
• GNR began accepting .name registration requests in batch mode on 12 December 2001 and began providing nameservice and Whois service on 15 January 2002.
• GNR continued taking .name registrations for over six months using a batch process, under which each registrar would periodically (approximately every two weeks) submit a batch file of registration requests.
• After submitting its .org proposal (by the 18 June 2002 deadline), GNR began real-time registration services for .org on 26 June 2002 using a combination of VeriSign-outsourced services and internally developed services. In particular, VeriSign provided key software for, and operates, the shared-registry systems that govern registrar interactions. These include the following real-time registration functions:
  • Allowing for EPP connections by the registrars
  • Serving EPP requests from registrars
  • Modifying data for the registrars in the registrar-authoritative database
  • Replicating the registrar-authoritative database to the GNR site London

Thus, it appears that, when implementing real-time registration services in .name, GNR did not employ its own technology, but used outsourced services obtained from VeriSign. While the appropriate use of outsourced services is not necessarily a technical concern, GNR has made it clear that it does not plan to use a similar arrangement with VeriSign for .org, so that its .name experience is not directly applicable for .org with regard to the real-time registration services function.

ISOC’s back-end provider, Afilias, also has VeriSign as an investor. Afilias is organized as a consortium of eighteen gTLD registrars. VeriSign is a minority (5.6%) shareholder of Afilias as one of these registrars. Because the other Afilias shareholders are VeriSign’s competitors, however, VeriSign’s ability to exercise control over Afilias is effectively minimized and, indeed, no VeriSign employee has ever been elected to Afilias’ Board of Trustees/Directors. In these circumstances, it does not appear that this investment relationship undercuts the competitive benefits of reassignment of .org, particularly in view of the fact that the .org registry would be assigned to ISOC, not Afilias.

UIA proposes to subcontract with VeriSign Global Registry Services (the current operator of the .com, .net, and .org registries) to provide back-end registration services for the .org TLD. UIA has entered into a teaming agreement with VGRS under which VGRS will provide to Diversitas (UIA’s non-profit operating subsidiary) back-end registration services for the first three years of the Registry Agreement. At the end of the second year, UIA will recompete the services initially to be provided by VGRS, with a target implementation date for the recompetition of 1 January 2006. Although this recompetition will ultimately result in an increase in diversity over the present circumstances, that pro-competitive increase will be delayed by three years.

In view of the review described above, it appears that criterion 3 renders the UIA bid less attractive, but does not bear significantly on any of the other bids.

In their comments on my earlier report, several applicants argued that the criterion 3 analysis should also consider alleged other (i.e. non-VeriSign-related) competitive drawbacks of selecting various other applicants. For example:

• RegisterOrg commented that "Several of the bidders already manage significantly large global registries" (apparently referring to NeuStar, ISOC/Afilias, and perhaps GNR) and advocated selecting a .org operator "whose selection will further diversify registry management and promote competition."
• GNR stated "The financial problems of both ISOC and NeuStar have been documented and need to be evaluated in consideration of their ability to operate .org." It also argued that, since GNR currently operates (with VeriSign’s assistance) only one TLD registry (.name), it would promote competition to award GNR a second TLD (.org) (to operate without VeriSign's assistance), rather than awarding .org to NeuStar (which operates .biz and .us) or ISOC/Afilias (which operates .info and .vc).
• NeuStar faulted my earlier competition analysis as "too narrow", contending that "a more comprehensive review of each applicant's financial health should be considered" because a "registry operator cannot be successful nor enhance competition if it does not have the requisite financial resources."

Although many of these arguments make for interesting reading, they are highly speculative given competitive realities. VeriSign's registry now supports over 31,400,000 domains, and will support approximately 29,000,000 domains even after the reassignment of .org. NeuStar and Afilias are quite small in comparison, supporting only about 1,000,000 domains each. Contrary to RegisterOrg's claims, these are simply not "significantly large global registries" in competitive terms. Indeed, selecting an operator with an existing registry operation would result in economies of scale that would likely provide that operator an enhanced capability to compete with VeriSign.

Similarly, GNR's "one TLD we win, two TLDs they lose" argument is not persuasive. There was nothing in the RFP that suggested that any current operator of two TLDs should be "marked down" over one that only operates one TLD. (Moreover, it should be noted that the .vc TLD operated by Afilias is quite small and just started up in July 2002.) By the same token, the fact that RegisterOrg’s selected back-end, Registry Advantage, provides service to eight ccTLDs does not suggest that a selection of it should be ruled out on the basis of criterion 3.

Nor would it be justified, as NeuStar urges, to rely on its claimed superior financial resources (which are disputed by GNR and public comments) as a ground of finding its selection to be more pro-competitive. Although the RFP made the size of firm commitments of resources that would be devoted to the .org project a relevant consideration in terms of assessing each proposal, it intentionally did not make the size of an applicant's general financial resources that it might devote to .org a factor. It is extremely hard to reliably assess a particular applicant’s financial stability directly. The spectacular failures of the past few years of large enterprises indicate the difficulties in predicting the financial future of any institution.

In sum, except for the negative competitive consequences of selecting UIA, analysis competitive considerations under criterion 3 does not meaningfully favor any applicant over any other.

C. Evaluation Under Criterion 10 (VeriSign Endowment)

Criterion 10 of the published Criteria document states:

10. Ability to meet and commitment to comply with the qualification and use requirements of the VeriSign endowment and proposed use of the endowment.

To the extent that a proposal contemplates the availability of the VeriSign US$5 million endowment (see subsection 5.1.4 of the current .org Registry Agreement), the proposal should demonstrate that it meets the qualification and use requirements of that endowment. Proposals that employ the endowment should also include detailed commitments about the uses to which the endowment would be put, and consideration will be given to the extent to which those uses are consistent with the smooth, stable transition and operation of the .org TLD for the benefit of current and future .org registrants.

In view of the above description, the following issues are relevant to evaluation of this criterion:

• Does the proposal seek the VeriSign endowment?
• If so, does the proposal appear qualified to receive it?
• If the proposal seeks the VeriSign endowment, what effects would failure to obtain the endowment have on the bidder’s ability to meet the requirements of its proposal?

Questions relevant to this inquiry include C2, C3, C4, C13, C13.1, C13.2, C13.3, C14, C41, C41.1, C41.2, C50.1, C50.2, and C50.4.

The following table summarizes my review of the above issues for each proposal:

Key to codes in “Appears Qualified?” column:

A=Appears likely to be qualified, with perhaps minor modifications, to receive the endowment.
B=Some issues regarding entitlement to the endowment; non-fundamental modifications to proposal may be necessary.
C=Raises significant concerns regarding entitlement to the endowment, which may require substantial revision to address.

As can be seen from the above table, only two proposals (GNR and SWITCH) raise significant issues under the VeriSign endowment. They are considered in more detail below.

• GNR does not itself claim to be qualified for the endowment, but instead proposes that its two non-profit "partners," .orgcentre and the Causeway Community Foundation, receive the endowment. This proposal raises serious questions because the VeriSign endowment is limited by contract "to be used to fund future operating costs of the non-profit entity designated by ICANN as successor operator of the .org registry." GNR does not propose that either of its two non-profit "partners" be designated as "successor operator of the .org registry" and GNR itself is not a "non-profit entity." Thus, it seems unlikely that the VeriSign endowment could be obtained under GNR's proposal. Because it is difficult to see how the proposal could be revised to qualify, GNR is assigned a "C" on the above table.
  
  GNR proposes that endowment funds be paid to the Causeway Community Foundation, as equally matching funds to those paid by GNR to Causeway. Thus, loss of the endowment would diminish Causeway’s funding by up to half. Causeway is envisioned by the GNR proposal also to provide a portion of the funding for .orgcentre, a community outreach initiative that GNR proposes. Thus, .orgcentre’s funding may also be diminished by the failure to qualify for the endowment.
  
  GNR's proposal portrayed the Causeway Community Foundation as a "good works" initiative, and .orgcentre as a component of its community outreach program. Both of these are mentioned in the Usage Evaluation Report as contributing to GNR's responsiveness (criterion 5) score, by way of its input/governance and "good works" sub-scores. The loss of the endowment would likely diminish GNR's score on criterion 5 to some degree, though this is difficult to quantify.
  
  
• SWITCH proposes to use US$ 1,800,000 of the VeriSign endowment for the following purposes:
  
  
  • building-up the Community infrastructure for registrars and registrants
  • differentiation of the ORG TLD

  It states in its proposal that it "will have no ownership or other rights or interests in such funds or in the manner in which they are used or disbursed. The funds will be devoted to the ORG community." As it is crafted, the SWITCH proposal suffers from a similar problem to that described above for GNR. The endowment is "to be used to fund future operating costs of the non-profit entity designated by ICANN as successor operator of the .org registry." Since the proposed successor operator, SWITCH, is proposed to have no ownership in the funds, it would not appear to qualify for the endowment.
  
  Since (unlike GNR) SWITCH is itself a non-profit organization, it might be able to revise its proposal to receive the funds and therefore qualify, yet still apply them to the intended purposes. In view of this, SWITCH is assigned a "B" instead of a "C" on the above table.
  
  A failure to receive the endowment would diminish the infrastructure-build-up and differentiation purposes mentioned above. It appears that this diminution might have lowered SWITCH's differentiation (criterion 4) score in the Usage Evaluation Report, though this is difficult to quantify.

As noted in my earlier report, several issues beyond ICANN’s control may affect the entitlement of a bidder to receive the entitlement as projected. Steps should be taken to ensure that it is clear that bidders assume all risks of not receiving endowment funds. Although in the model .org Registry Agreement ICANN makes no warranties as to the endowment, the selected applicant should be required to expressly assume all risks of non-receipt of the endowment.

D. Evaluation Under Criterion 11 (Completeness of Proposals/Realistic Plans)

Criterion 11 concerns the "completeness of the proposals submitted and the extent to which they demonstrate realistic plans and sound analysis." The Gartner Evaluation Report explicitly evaluated criterion 11 as it relates to the technical aspects of the proposals; each applicant is scored on this criterion by Gartner. It should be noted that Gartner's scoring on criterion 11 closely tracked its cumulated scoring on the other technical criteria, so that additional consideration of criterion 11 would have been cumulative in effect. Criterion 11 was not explicitly scored with respect to the usage aspects, but the Usage Evaluation Team's report shows that completeness, realistic plans, and sound analysis counted significantly in its evaluations. Completeness, realistic plans, and soundness of analysis are also incorporated as significant factors in my analysis of criteria 2, 3, and 10 above. In sum, the elements of criterion 11 were significantly reflected in the overall analysis.

E. Conflicts-of-Interest Issues

Some comments have expressed concern that members of the ICANN Board might have a conflict of interest concerning the Internet Society bid. Ten of the ICANN Directors are members of the Internet Society, but none of them would financially benefit in any way, directly or indirectly, from the selection of the Internet Society as the successor .org operator. ICANN's conflicts-of-interest policy prohibits each ICANN Director from participating in decisions on "any matter in which he or she has a material and direct financial interest that will be affected by the outcome of the vote." It does not prohibit Directors from voting on matters merely because those matters involve a philanthropic society which the Directors may philosophically support. Conflicts-of-interest principles are intended to prevent improper personal financial gain, not to negate the philosophies of ICANN's Directors. Directors are not expected to discard their philosophies and past experiences when they undertake their service at ICANN, but instead are expected to incorporate those philosophies and experiences in formulating well-reasoned decisions. (For similar reasons, the fact that the Usage Evaluation Team may have included some ISOC members does not constitute a conflict of interest.)

Respectfully submitted,

Louis Touton
ICANN General Counsel