[ARIN-consult] Comments of the Internet Governance Project

[Milton L Mueller]

Comments of the Internet Governance Project (IGP) on the proposed Transfer Listing Service

ARIN is to be commended for making this first attempt to implement a Transfer Listing Service (TLS). In effect, ARIN is setting up a market intermediary or exchange. In our view, ARIN should strive to make this market as liquid, open, efficient and responsive to supply and demand as possible.
General comments
To design it correctly, ARIN must have a clear understanding of two things: the purpose of the Transfer Listing Service (TLS) and its own role in the transfer process.

Regarding the TLS's purpose, in our view the design should be based on the following premises:

*         There will be legitimate needs for IPv4 addresses well beyond the exhaustion of the free pool

*         Prolonged scarcity of ipv4 addresses means that there must be a mechanism to move ipv4 resources from lower-value to higher-value uses;

*         It is vital to retain ARIN's function as a comprehensive and accurate address registry, and transfers that take place without its knowledge or recordkeeping risk undermining this function;

*         It is more important to get parties who hold address blocks they don't really need to release them for others to use than it is to ascertain whether these parties actually needed the block they currently hold, or whether they are signatories to ARIN contracts. Those receiving resources can be brought into the contractual regime; it is less important whether those releasing resources meet rigid guidelines. (Obviously, sales of hijacked or illegitimately claimed resources must be screened out.)

Regarding ARIN's role, what ARIN calls a "Transfer listing service" implies four conceptually distinct functions:

1.       Verification of who is a legitimate offerer of ipv4 address assets and verification, through "needs assessment," of who is a qualified seeker or bidder.

2.       A listing process which publicly identifies seekers (bidders) and offerers (askers) of ipv4 address assets;

3.       A process for matching of bid/ask prices;

4.       A clearing house which formally registers the transaction once verifying that appropriate payments are made (including any transaction or listing fees imposed by ARIN), and officially alters the title to the resources.

In our view, ARIN only needs to do #1 and #4. Our understanding of the proposed TLS is that ARIN has devoted most of its attention to the first and second functions, has completely ignored the third function, and has not fully thought through its approach to the fourth function.

Ideally, ARIN policy should foster an efficient and decentralized market in which its role is restricted to authenticating legitimate sellers and transferring titles to legitimate buyers once transactions are concluded. In this view, the second and third functions could emerge in the market based on authoritative information supplied by ARIN. However, setting in place the infrastructure for a more robust and competitive exchange system could be more complicated than a more closed and monopolistic system. Due to the need for immediate action, it is important to get the TLS up and running in a simpler form and consider ways to improve it in the second and third iterations.

ARIN needs to specify whether the listing service will list asking price (we think it should) and whether seekers can also specify their initial bid price (we think they should be able to). The market will be more efficient and prices for seekers will be lower if this information is public. Askers and bidders should be able to modify their listed ask/bid price easily and quickly. ARIN should publish the clearing price of address resources, both to facilitate research and assessment of the transfer policy and to aid industry planning and operations. (Organizations that may not yet need addresses but think they might in the future have a legitimate interest in access to information about the volume, price and specific blocks traded.)
Questions posed by the Consultation:


1.       How long should ARIN offer a Transfer Listing Service and what, if any, conditions should be tied to its availability?
ARIN should maintain the TLS indefinitely. The listing service should be offered as long as ipv4 addresses are in use and for the duration of a full transition to ipv6. Because no one knows when ipv6 will completely supplant ipv4 it is unwise to set an arbitrary time limit on the TLS.

ARIN should, however, periodically reconsider the design of the TLS. ARIN might commit itself to a review of the service design every 18 months or so.

2. As proposed, there is a usage agreement for participation. Is this a fair and reasonable requirement?

As (vaguely) described in the consultation, the usage agreement specifies that parties can only participate in the TLS "for the purposes of meeting the requirements of the Transfer Listing Service participants." We don't know for sure what this means. If it is simply an anti-spam agreement, we have no objections, although misbehavior of that sort might already be covered by existing laws. If it is an attempt to strongly restrict who can look at listings of seekers and offerers, for example by limiting access to parties who actually hold address resources or who have contracts or "demonstrated need," we think it is a bad idea.

Listing information should be public if it is to serve its primary function of facilitating transfers among parties. Transparency and liquidity go hand in hand. Brokers and intermediaries who could facilitate the effective functioning of the transfer market would have a legitimate interest in access to the information, as well as academic and industry researchers.  Once the information is public, ARIN is not in a very strong position to control subsequent "uses" of the information, which could be compiled, stored, exchanged and analyzed by anyone who possessed it. Thus, we question the need for such an agreement.

3. Under what conditions should ARIN disallow or remove either an organization listing or seeking IPv4 address space from the service? For example, if an organization has a Registration Services Agreement ("RSA") with ARIN and is not current on payments with ARIN, should they be allowed to participate?

ARIN must clearly define its priorities with respect to the proposed listing service. In our view, the most vital purpose is to facilitate the movement of resources from parties with unneeded or larger-than-needed assignments to those with too few address resources.  An equally vital function is to preempt the development of underground trading of address resources. Those two purposes should take priority over all others.

ARIN can and should insist that those receiving resources under the policy (successful bidders) be under an RSA and able to demonstrate technical requirements or "need" for them. But ARIN should be much more liberal about regulating legitimate address block holders who are willing to release resources through the listing service. The movement of resources from legacy, non-contracting parties to contracting, fee-paying parties benefits both ARIN and, presumably, the public by bringing them under a consistent policy regime. ARIN should do everything it can to encourage participation by legacy holders and remove all obstacles to their participation. While obviously ARIN should not allow hijacked or illegitimately identified resources to be transferred, the listing service should not be confused with some kind of tool for retroactive punishment of past overconsumption, or as a way to coerce organizations into signing RSAs. Such an approach would simply discourage participation and undermine the net benefits that could be gained from efficiency-enhancing transfers. Attempting to leverage the listing service to achieve unrelated objectives produces no corresponding benefits, as parties could simply avoid it.

With respect to unpaid ARIN fees, it makes no sense to bar an organization from participation in the transfer market on that basis - but it would be justifiable to make the final settlement and re-registration of the address resources contingent upon the back payments. Buyer payments to the seller could be tapped for this purpose during the settlement stage.

4. Should there be a fee associated with the Transfer Listing Service? If so, on what factors should the fee be based?

ARIN's primary purpose is to facilitate liquidity and the matching of supply and demand for v4 addresses. Fee mechanisms and fee levels that discourage participation in the listing service should be avoided.

But we have no objection to flat, simple listing fees that cover the costs of providing the service. Bear in mind that issues related to the size of fees become less difficult if and when competition enters into the picture. As noted before, TLS designs that allow competing third parties to operate exchanges should be explored going forward.

5. Are there any other factors that ARIN should consider in providing this service?


*         How will the success or failure of a particular TLS design be assessed?

*         Will it be possible for third parties/consultants or other intermediaries to provide gateways into the ARIN listing service?
Milton Mueller
Brenden Kuerbis
Lee McKnight
Michel van Eeten
(IGP)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-consult/attachments/20100511/f8ba3695/attachment.html>