Here is more evidence of how the emergence of a private market for IPv4 addresses is creating concern at ARIN and driving policy change. Note that in my interview with private address broker Addrex, published here a week ago, its President Charles Lee noted that large ISPs in the Asia Pacific region were showing strong buyer interest in legacy address blocks currently sitting idle in North America. This poses a problem for the current RIR system. The RIRs are organized around exclusive territorial regions and their policies basically prohibit inter-regional sales of address blocks. A private broker dealing in legacy address space, however, is not subject to those constraints. Once again, an emergent market subtly disrupts the establish institutional regime.
ARIN is, evidently, scrambling to catch up. A new proposal was submitted to ARIN's policy development process just yesterday. Entitled “ARIN-prop-156 Update 8.3 to allow inter-RIR transfers” it would allow inter-regional market transfers. Here is the specific language:
“…IPv4 number resources may be released to ARIN by the authorized resource holder, in whole or in part, for transfer:…to another RIR, for transfer to a specified recipient in that RIR's service region who demonstrates plans to deploy the resources for the justified purpose within 3 months, as long as the request meets the policy requirements of both RIRs, and the recipient (and any organizations to which they have transferred or reassigned space) can show efficient utilization of all prior allocations, assignments, and transfers according to the current policy requirements of both RIRs.”
But adjusting the incumbent regime will not be easy. Note how complicated this process is: you transfer not from buyer to seller but to another RIR who controls the buyer's region. Note the rather severe restrictions attached to the transfer: the allocation is still contingent upon demonstrated need, the resources must be deployed within 3 months, the request must meet the policy requirements of both RIRs, and the recipient must run a gauntlet proving that they are efficiently utilizing all prior allocations, assignments and transfers that have ever been given to them in the last 15 years.
Even with all this bureaucratic baggage thrown in, it is unclear how the proposal could be implemented in the current RIR system. Paul Wilson, the Executive Director of the Asia-Pacific RIR (APNIC), popped up on the ARIN policy discussion list within hours of this proposal being posted and asked:
* “Has there been any consideration of the cost and practicality of this approach?
* “Is it possible that a recipient would be asked to pay ARIN to cover the cost of this service?
* “Would a registration services agreement be required with ARIN?”
To which John Curran, Wilson's counterpart at ARIN, responded: “Without additional clarity (either in the policy proposal or filled in during the staff assessment), it is very uncertain how proposal-16 would be implemented.” The proposal has sparked an interesting debate about “trade reciprocity” in address blocks, which is archived here.
What's at stake here is not really implementation detail, but a structural tension between the regionally exclusive nature of the existing RIR address allocation/regulation system, and the existence of a global market for IPv4 addresses (and internet services generally). The problem could be avoided if the RIRs were integrated into a more globalized system, and/or by dumping the insistence on needs assessments, which creates a layer of bureaucracy and inspection that adds little of value to the Internet marketplace but does add huge barriers to trade. But while those barriers do little for users and suppliers of internet service, they are perceived by each RIR as absolutely necessary for them to retain policy control.
3 thoughts on “International trade in IP address blocks”
While I'd agree that the current differences in each region don't necessarily contribute to easy transfers in unified global market, ARIN and the other RIR regions has been talking about how to do inter-regional transfers for a couple of years.
The reason that we don't have inter-region transfers is that collectively the regions have not agreed on a common approach. Until this occurs current policy does not permit these types of transfers. While it would be “nice” to just wave a magic wand and have global policies just be easily adopted, the multi-stake holder process in place now requires considerable time and coordination to create global IP number resource policies or globally coordinated policies. It doesn't seem like we should just ignore the current policy process just because it isn't “fast enough” or doesn't produce policies that don't achieve the current standing requirements of a consensus driven policy creation process. The process isn't perfect, but I believe it does produce good policies when there is significant consensus between regions.
The RIRs continued the process of “demonstrated need” in the transfer market not to be difficult, but to reduce the amount of hording and speculation in the new marketplace. One could argue that this is not the best method to deal with hording and speculation, but this is what the community achieved consensus on.
Good points all, Andrew. Yes, the difficulties around coordinating regional policies have been mounting for about 2 years, this blog post could not go into all of them. The RIRs have been unable to agree on how to handle blocks returned to the IANA, for example, and they also had trouble agreeing on a coordinated approach to RPKI.
Of course it's true that we can't wave a magic wand and do away with these problems; the RIRs exist, they have their virtues and strengths, and at any rate they can't be wished out of existence leaving a clean slate. On the other hand, the “community's” tendency to achieve consensus on dumb things, like “needs assessments,” will either drive the whole process into the arms of external private actors or freeze things in a less than optimal place
Unfortunately, your rhetoric obscures the facts for the readers.
1) “ARIN is, evidently, scrambling to catch up. A new proposal was submitted to ARIN's policy development process just yesterday. Entitled “ARIN-prop-156 Update 8.3 to allow inter-RIR transfers” it would allow inter-regional market transfers.”
This is not the first Inter-RIR transfer policy proposal, it is an additional one just submitted in the ARIN region. There has already been discussion since late 2010 on ARIN prop-119 (now Draft Policy 2011-1) https://www.arin.net/policy/proposals/2011_1.html
which would provide for transfers of number resources globally.
2) “Even with all this bureaucratic baggage thrown in, it is unclear how the proposal could be implemented in the current RIR system. ”
To be clear, no assessment of new proposal has been done, and many implementation details are unclear.
You conveniently omit the rest of my remarks from the ARIN public policy mailing list: “This is in contrast to draft policy 2011-1, which when processing transfers would simply require ARIN to confirm the transfer with the recipient's RIR. ” i.e. The other existing policy under discussion is quite straightforward in implementation if adopted.
3) “The problem could be avoided if the RIRs were integrated into a more globalized system, and/or by dumping the insistence on needs assessments, which creates a layer of bureaucracy and inspection that adds little of value to the Internet marketplace but does add huge barriers to trade. But while those barriers do little for users and suppliers of internet service, they are perceived by each RIR as absolutely necessary for them to retain policy control.”
It's up to the community in each region to determine what policies are necessary. While there is a global framework set out in ICANN ICP-2 and RFC 2050, the community in each region determines the appropriate values and decides policy based on open and transparent discourse. The RIR staff implement these policies. Are you advocating the RIR staffs should “guide” the community to your desired free market or are you advocating that RIR staff should disregard the output of the community process? It is difficult to ascertain what process for Internet governance you are suggesting.
President and CEO
Comments are closed.