[arin-ppml] About needs basis in 8.3 transfers

[Owen DeLong]

On Jun 6, 2014, at 7:57 AM, Mike Burns <mike at iptrading.com> wrote:

> Hi Larry,
> 
> "Personally I suspect that without needs testing the "haves" would have had it all a long time ago.
> I have felt the same frustration, as a small provider, trying to meet the 80% requirement can be almost
> impossible without gaming the system due to numerous small holes in a small allocation.
> That said, I worry about any company that could purchase a couple of Billion dollars of IPV4."
> 
> What is being expressed here is a fear of hoarding in the absence of a needs test. A couple of billion dollars of IPv4 at current prices would yield about 15 /8s. Even if some company wanted to risk those funds with IPv6 transition threatening to erase them, there is no single seller who could offer 15 /8s, nor would the sequestration of 15 /8s destroy the market, since this amount represents just 25% of the estimated market size of 1 billion addresses. Since the buyer and seller will be disclosed and registered under any no-needs policy, there is little threat of a stealthy move here, and any buyer seeking to corner or manipulate the market knows he does so at the peril of forcing the IPv6 transition.  The best protection against this is continued work towards IPv6 and the establishment of a reliable, open, and global IPv4 market with at least the same level of transparency into registration as we have now.  As a broker, I actively sought out speculators to bid for addresses in the Nortel sale. This was a prime opportunity to acquire 660,000 addresses at the floor of the market, but it was regarded as too risky by almost all. In the intervening years I became aware of other opportunities to acquire address space without needs tests, but I never found anybody interested in buying addresses on pure speculation. In any case, this fear can only reasonably be expressed in the context of a complete removal of needs tests, and could not be applied to a more limited removal of the needs test such as that proposed in 2014-14.

Finding 4 actors who want to corner the market probably wouldn’t be very hard. Since 25% is 1/4 of the projected market size, I would say that the rest of your argument is on pretty shaky ground.

> "Many of us fear that if need is not considered in the transfer market the little guys will find that none is available at any price."
> 
> We are three years into the open, post-Microsoft/Nortel market and there is no evidence of hoarding in my experience. I have never fielded a phone call or email from any company or individual seeking addresses they didn't plan to utilize at some point, although I have fielded plenty from people seeking addresses that for whatever reason ARIN policy would prohibit them from registering. Perhaps other brokers on the list might report on their experiences.

Address space is still available nearly for free from ARIN, especially for smaller organizations, so this isn’t a real test of what will happen post runout and any claim that it is is absurd.

> "Like it or not the big guys have an advantage. Let's make sure that "cornering the market" isn't one of them."
> 
> Little guys benefit from the dropping of needs test for small transfers. No need to navigate the ARIN process if you just need a /24 and you can't get one from your upstream, or not at a reasonable cost, or because you feel more secure with your own space, or because you don't wish to game the system. Support of 2014-14 would allow small companies to  have this option while preventing hoarding or speculation through limits on size and number of transfers.  Perhaps you care to comment on whether you might consider support of 2014-14 at the current size of /16 or at another size that you might feel more appropriate?

Again, this is your perspective, but it’s not necessarily entirely true. It’s only true if you assume that the market will have a continuous supply and demand will be lower than supply. I would argue that the number of inter-RIR transfers to the APNIC region which have already been processed would indicate that after exhaustion, this is unlikely to be a persistent state or even last very long at all.

Owen