[arin-ppml] quantitative study of IPv4 address market
Milton L Mueller
mueller at syr.edu
Tue Sep 4 16:42:52 EDT 2012
Like Owen, I found this message from Paul confusing. But very interesting.
Before explaining why, let me dip into the prior controversy raised by Owen and Joe. Note that our study of the IPv4 market identified TWO factors - not one - that stimulated a trading market in the ARIN region prior to the exhaustion of the free pool. Only one factor was the short 3-month time horizon for defining 'need.' The second factor was the stronger property rights that could be obtained via transactions with legacy holders. By itself, this seems to indicate that Owen's desire to lengthen the time horizon for needs assessment would address only part of the problem. (FWIW I tend to agree with Owen's recommendation to lengthen the time beyond the ridiculous 3 month period for the rest of the free pool.)
Now, it seems, Paul has identified a third factor, or more accurately, a variant of the second. If I understand what he is saying correctly (and I may not), there was an options market for future address space conducted privately among address holders in a way "not subject to ARIN rules." Providing more evidence for the conclusions of our study, Paul intimated that this underground options market was driven by the mismatch between the short time horizon of ARIN needs assessments and the longer time frame dictated by business needs (measured in "half-decades").
But then Paul seems to contradict himself by saying that these options are consistent with ARIN rules and could be exercised using the directed transfer rule. My understanding of the directed transfer rule was that it required ARIN to do a needs assessment with a 1-year time horizon (which was recently extended to two years). Paul makes it sound as if the transacting parties themselves decided who needs what. That may just be unclear writing, so correct me if the interpretation is wrong.
Anyway, Paul reaches the conclusion that the existence of this "grayish" market indicates that "ARIN has a workable system even at three months." To me, this is only true if the grayish market of which he speaks basically bypasses ARIN needs assessment, or if two years corresponds through some form of new math to the half-decades to which he refers. Or perhaps what Paul is really saying is that the 3-month time horizon is just a step removed from the dribs and drabs of the "last /8" policy and therefore we should keep it that way, because other needs are being fulfilled in the market. Clarification welcome, if I have misinterpreted your message.
i've now sat with several arin members who have told me privately that their business needs for ipv4 growth are measured in half-decades not years, and so they were optioning future address space through a grayish transfer market even before arin went to a three month regime. i say "grayish" because the option agreements are a private matter not subject to arin rules, and the space in question will inevitably be transferred to the recipient upon demonstrable need. i've been told that the directed transfer rule whereby resources can be transferred between parties without first returning it to arin and then reallocating it, was the only instrument they needed.
to me this says arin has a workable system even at three months, and that unless this community chose to forego any needs basis at all, there is no way to ensure that addresses are available to those whose real demonstrated need -- which will be demonstrated in terms of capital for the network and also capital for the options and ultimately the resources.
this community has reached consensus on three month allocation windows. that consensus could be changed by debate. i welcome such debate.
but in no sense is non-needs-based allocation (within the community's chosen window, currently of three months) definitionally a "legitimate" use, such that "disadvantaging" such use is "not a win". nor would any of us enjoy an internet in which policies of this kind are set in any way other than by community consensus.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the ARIN-PPML