While last week's IG news was dominated by the ICANN meeting in San Francisco, including the Board's watershed moment approving .xxx, other important IG decisions were occurring elsewhere. In accordance with the ARIN Policy Development Process (PDP), the ARIN Advisory Council (AC) held a meeting on 17 March 2011, deciding on some important policy debates we've recently covered.

According to it's
announcement
, the AC abandoned policy proposal, ARIN-prop-132 ISP Sub-assignments Do Not Require Specific Customer Relationships,which would have allowed ISPs to lease excess address space without going through ARIN to get its approval on a case by case basis. The AC claimed:

– the policy encourages deaggregation and accelerates routing table growth.
– it creates a situation where the ISP providing the address space is not in a position to remediate security and abuse issues which may have significant impact on the larger internet community.

So, it appears that the AC believes ARIN is best positioned to centrally manage how address space should be aggregated and that ISPs, whose business actually depends on efficient use of addresses and routing tables, are not incentivized to monitor and address this issue. The second objection seemingly is tied to the alleged problem that ISPs would not be able to maintain accurate contact information about reassigned space, and therefore be unable to “remediate security and abuse issues.” In other words, ARIN is encouraging addresses to be used as leverage for law enforcement.

A second proposal, ARIN-prop-133 No Volunteer Services on Behalf of Unaffiliated Address Blocks, was abandoned by the AC because:

there was not significant support for it on PPML and because the proposal as written is not something that could be implemented under the existing circumstances (ICANN MOU, ARIN structure, and the fact that there is no provision for “alternate” registries in any existing documents, RFCs, or other approved structures).

This is disappointing because it is saying, in effect, that a needed policy innovation shouldn't happen because it is different from what we have been doing (duh). ARIN is thus keeping in place the current, unsatisfactory situation, whereby legacy holders sort of are, and sort of aren't, part of the regime. This encourages legacy holders to secede from the regime altogether and maintain a grey market for IPv4-related services. 

A third proposal, ARIN-prop-136 Services Opt-out Allowed for Unaffiliated Address Blocks, sought to let legacy space holders opt-out of the ARIN regime, while maintaining obligations to maintain accurate Whois information. The AC provided following rational for abandoning the proposal:

– Overwhelming lack of support within the ARIN community, as evidenced by discussion on the PPML.
– ARIN staff has indicated that they do not believe that it is possible to implement this policy due to existing agreements, namely they can not not-serve our community.
– Today folks have the option of opting out of SWIP by using RWHOIS and providing ARIN information about the location of the RWHOIS service to publish in the SWIP record.  This hasn't always gone so well, not all RWHOIS services are kept up to date and publicly available.  Providing another mechanism for folks to opt out of having records published publicly in a centralized location, is not in the best interest of the community, as it would complicate abuse, security and law enforcement investigations.

The first point highlights a shortcoming in the ARIN PDP and more generally of Internet governance processes which draw sparse participation.  The draft policy steps of the PDP are rather informal, there are no requirements or defined process for summarizing comments received.  A simple count of “for” and “against” emails to a list is a nice bellwether indicator, but it is certainly not representative of the potentially affected parties opinions.  The second argument, once you parse it, really makes no sense.  Similar to ICANN and DNS industry, ARIN and IP address allocation is a contractual regime. How could ARIN have an obligation to provide services for resources it did not allocate?  And the third point again highlights the growing influence of LEAs in shaping ARIN policies (although we don't recall seeing LEAs participate in the “bottom up, multistakeholder” PPML discussion around these policy proposals).

All of the above proposals focus on the relationship between legacy space
holders and ARIN.  While abandoning the proposals short circuits any attempt to arrive at mutually agreed upon policy for now (there is an appeal mechanism), the AC's actions certainly clarifies the position of ARIN.  ARIN wants legacy space holders to come under their contracts so that it can maintain control of a single registry of IP address resource allocations and the identify information associated with those resources.

3 thoughts on “ARIN says “No!” to competing registries, independence of legacy space holders

  1. Brenden –
    The ARIN Advisory Council (composed of 15 people from the Internet community elected at large) has abandoned many policy proposals before, so it is somewhat surprising that these actions are deemed noteworthy.
    I just want to take one issue with respect to your characterization of the abandonment of ARIN-prop-133 and ARIN-prop-136. ARIN performs Internet number resource management for region composed of the US, Canada, and parts of the Caribbean. This region used to be much larger, consisting of all non-APNIC/non-RIPENCC locations, but we worked with the Latin America and AfriNIC communities to assist them in forming their own regional registries). Within ARIN's service region, we have an obligation to provide registration services with due consideration to the principles of aggregation, conservation, and accuracy.
    ARIN-prop-133 and ARIN-prop-136 did not propose competing registries, but instead proposed abandonment of ARIN's current responsibilities without any actual proposal for how registration services would reliably be provided in the region. Note that the registration system doesn't just serve the address holder, but also serves the entire Internet community, and changes to the structure of the system should be clearly documented and discussed on a global basis. This could be done by proposing an alternative structure to the ICANN's ICP-2 global policy. It is conceivable that the overall goal of the Internet registration might be fulfilled successfully by some alternative structures to the present system, but there's no way to know that until there are actual proposals and discussion of same. As you are aware, this is how Internet governance is actually done, i.e. by the public discussion of the merits of actual proposals, and the presumption of ARIN-prop-133 and ARIN-prop-136 that the overall structure and operation of an alternative registry system does not first need to be documented is obviously incompatible with public discourse of the actual pros and cons of such alternative schemes.
    ARIN can't set aside any of its responsibilities in the registry system unless and until there global community consensus on some viable alternative structure which fulfills the same needs, and hence any proposals to abandon services on an ad-hoc basis will be deemed out-of-scope.
    /John
    John Curran
    President and CEO
    ARIN

  2. Hi Brenden,
    In addition to John's valid comments, I'd like to point out that you seem to be conflating ARIN with the ARIN AC. They are not one and the same.
    How else would you like to measure support or opposition to policy proposals if not by what is said on the PPML? To quote someone from the thread on prop-136:
    “If they are unrepresented, it is because they choose not to represent themselves. The policy process is completely open to them.”
    In terms of contractual obligations, when ARIN took over the administration of IP address registrations from the InterNIC, it also took the registrations from the legacy holders, and has been continuing to provide services for them. Abandoning these legacy holders would be irresponsible, as I am sure the IGP would be the first to point out if that came to pass. Perhaps you overlook how often voluntary services are provided by the Internet technical community for the good of the Internet (rootserver operators spring to mind).
    Have you actually read ICP-2? It is a very useful document to have around.
    rgds,
    McTim

  3. Just to echo McTim's comments, there was very little community support for the propositions that you listed. In this case, these were props that you support, but the comments by those who did participate show that they didn't.
    ARIN's policies are community driven. If there are people you want to participate, encourage them to do so. If you think participants don't understand the positive value of the policies, arin-ppml is a forum to promote it.
    I understand that just because the participants in the community don't move a proposal forward that it was “right” (in the absolute sense of the word), and that's why there's multiple opportunities to petition them. But don't blame the ARIN AC, because if they had moved those props forward they would have not been representing the participants.
    Frank

Comments are closed.