[arin-ppml] About needs basis in 8.3 transfers
Morizot Timothy S
Timothy.S.Morizot at irs.gov
Tue Jun 10 17:54:57 EDT 2014
> I certainly do understand why those large Legacy holders would just watch and
> wait to see what happened in the marketplace once all of the RIRs run out of
> resources. It is in their interest to do so.
> but there have been cases of good net-citizenship where entities have
> returned large blocks (even entire /8's) in recent years.
Indeed. However, just from knowing the sorts of organizations that received many of the larger legacy allocations, I would wager that those who have not already taken some sort of similar action are probably not interested in either returning or selling those allocations. There are a number of smaller legacy allocations that will likely find their way onto the market. And there are a number where the registered recipient may be defunct.
Our /13, /15, and /24 don't really qualify as the sort of "large" legacy allocation under discussion, but I know we have no interest in selling them. And we also have no interest in returning them during any timeframe in which there would still be a demand for new IPv4 on the Internet. Out of the legacy holders, only commercial entities are likely to be significantly swayed by market demand for IPv4. And out of the commercial holders of large allocations of legacy space, most of those remain large companies. They are unlikely to treat their legacy holdings any different than their non-legacy IPv4 holdings. I've seen, for example, Lee Howard's analyses. Legacy or not, commercial entities are most likely to sell IPv4 only if and when they bring more on the market than technologies like CGN cost. There may be movement in parts of the smaller legacy allocations when there's a financial incentive to do so, but I don't anticipate some large-scale stampede to the v4 market.
Anyway, given that IPv4 will be a tightly constrained resource during the initial phase of exhaustion where inequalities of scale definitely come into play (large players are better positioned to buy more than they need in an effort to disadvantage smaller players) I'm not at all convinced that "market magic" will naturally ensue if needs basis is eliminated for transfers. I remain unconvinced by the arguments presented.
And if limiting markets by continuing to require a needs basis leads some organizations to do things that end making IPv4 even less attractive moving forward, I'm not convinced that's a negative side effect. I don't believe there's any stopping the Internet's switch to IPv6 at this point. I was concerned for quite a while, but the major deployments by access networks this year and the major upward trend in IPv6 traffic (in the US) have done a great deal to alleviate those concerns. By this time next year, I expect we'll be having very different discussions.
More information about the ARIN-PPML