[arin-ppml] About needs basis in 8.3 transfers

Owen DeLong owen at delong.com
Wed Jun 4 19:23:08 EDT 2014

On Jun 4, 2014, at 8:16 AM, Mike Burns <mike at iptrading.com> wrote:

> Hi Blake,
> We can be wistful for the lack of progress of RPKI or the fact that addresses are regularly routed for customers who are not the Whois registrants, but we are powerless to change those things, community-wide.
> We are a community of private network operators for the most part. We are stakeholders tasked primarily with maintaining a registry of uniqueness of IP addresses. We need to ask ourselves whether the purported benefits of maintaining a needs-test for every change of registrant in Whois is worth the risk to the registry and the expenditure of fungible ARIN staff resources.
> I elucidated one such risk, which is the risk of un-registered acquisitions of shell corporations which are incentivized by the lack of a needs test. John Curran acknowledged this risk.
> I offered an example of one of the few publicly demonstrable cases of this in Whois, related to the public information surrounding the Microsoft/Nortel deal. I am aware of many more but can not disclose them.
> People seek to frame this issue as if it were this question: "Should we change the rules just because some people will break them?"
> My answer to that is yes, of course we should, unless the rule provides some overriding benefit.
> So my question for the community is "What is the benefit we realize by insisting on ARIN team review of every single transfer, down to /24, and is it worth ARIN ticket time delay and the risk of decreased Whois accuracy?”

The benefit is preserving addresses on a fair basis for those who actually have legitimate and quasi-immediate use for them.

Yes, this benefit is worth the ARIN ticket time and delay.

There has not yet been any actual evidence presented to show that the risk to whois accuracy is any greater without this proposal than it is with it. Those that would ignore ARIN policy to effectuate a transfer are just as likely IMHO to ignore whois as not.

> And secondarily, what size of un-needs tested transfer would be an acceptable balance between the benefits of the needs test and the costs of the needs test?

/24 seems like a perfectly reasonable balancing point to me. I’d be willing to conduct an experiment on a temporary basis at /20 for a limited time (12 months).


More information about the ARIN-PPML mailing list