[arin-ppml] ARIN-prop-172 Additional definition for NRPM Section 2 - Legacy Resources

[Owen DeLong]

On Jun 13, 2012, at 8:04 PM, Milton L Mueller wrote:

>> -----Original Message-----
>> If I put the property religion aside, can we explore one of the points
>> you make below? 
> 
> [Milton L Mueller] Sure...
> 
>> [is it] advisable for all the ISPs, End user network operators, content provider
>> networks, Universities, and everyone who operates the Internet to pave
>> the way for post-allocation services and interact with other addressing
>> registrars.
> 
> [Milton L Mueller] For some of them the answer is emphatically yes, for others, it will make little difference either way. 
> There is no homogeneity in this community with respect to their interests on this matter. Tom Vest's attempt to poison the well notwithstanding, we don't have to have an abstract debate about methodological individualism in the social sciences to agree on that simple fact.
> 

That simple fact isn't. You state that some have an emphatic interest, which I believe, but, you also assert that no harm is created by destroying the existing system of policies in favor of this new zero-policy regime. Here, you are off the rails when you claim that is simple fact. Indeed, there is no basis for such a claim and ample evidence to the contrary.

>> My question would be why? As with any value proposition, do the benefits
>> they receive outweigh the price they have to pay? 
> 
> [Milton L Mueller] That depends on how big the price is and how big the benefits are. I suspect that both will be a function of the length of time it takes to get to IPv6, if we get there at all. If we enter into an extended period of IPv4/v6 coexistence, say 10 years or more, increasingly scarce v4 resources will need to be carefully managed and husbanded. In that case, the value of having specialist address registrars and a competitive post-allocation services market would be quite enormous, and extend to quite a few operators. And the value of having ARIN and the other RIRs develop an ability to integrate multiple private, competing, efficient registrars with a single, global authoritative database would be potentially quite beneficial also - even if the transition to v6 was not so long. 

The price includes, but is not limited to:

	The ability to have any form of homogenous address management policy administered by the community.

To me, that's a pretty steep price to pay given that the natural outcome of such a situation is corporate darwinism and the golden rule (he who has the gold rules).

> 
>> All the networks around
>> the globe benefit from the central coordination of the RIR system as a
>> mechanism to participate in the Internet.
> 
> [Milton L Mueller] problem is, it isn't really centrally coordinated. There are 5 RIRs and you have to search separately in each one. The RIRs and their Whois are actually quite antiquated in some respects. 
> 

Actually, this isn't really true. Each RIR contains referral entries which are easily parseable pointing to the other RIRs records...

You can search IANA and get a pointer to the applicable RIR for each /8 (and in a few cases, smaller blocks).

ARIN and LACNIC also return pointers to the correct other RIR in response to queries for out-of-region resources.

You should never need to search more than 2 registries to get the data you need, even if we added more registries.

Compare this to DNS, where you actually have to recurse at least 3 servers, just to get an authoritative record for a second-level domain (root-server, TLD-server, zone authoritative server).

It would not actually be hard to implement a tool which would automatically recurse the whois hierarchy, but, I suspect nobody has found it to be worth the effort since the (maximum) 2 queries for an address you don't know where to find is really not that hard and I suspect most people rarely end up searching for addresses in unknown regions.

>> Because some address holders or individuals wish to trade IP address
>> space for money, it is not done without a cost. To operate under such a
>> model, those costs are pushed out to all those entities listed above
>> that currently benefit from the central coordination. Why should all the
> 
> [Milton L Mueller] I have been saying all along that you can have central coordination AND competition/multiple registrars; we do it in DNS and we do it in stock trades. So are you assuming that the model I am discussing involves eliminating coordinated uniqueness? No wonder you think you're against it!  
> 

DNS is effectively the wild west unless you run afoul of WIPO's interest and/or the law. I would not hold it up as a shining example of how this should work, but, rather point to it as the fetid pit that will result from this proposal.

As to stock trades, I'll leave the relative merits of that system to the debate between you and TV as Tom seems far more informed that I would ever want to be on the subject.

>> network operators across the Internet incur the costs of having to refer
>> to multiple regristrar systems and the inevitable disputes that will
>> result beyond those that already occur in the current model?
> 
> [Milton L Mueller] Like Paul V., you see this as something that only a few brokers want. I suspect that it is operators and ipv4 address holders that will drive this, the brokers are only responding to that market. If the operators don't want it, or if it harms operators or increases their relative costs, it won't happen. On that we agree. 
> 

Actually, I disagree. I believe that in the long run ASN trading will harm operators and increase their costs. The same is true with IPv4 trading, but, not allowing it would have done more harm. However, what matters is whether or not the operators (and others) that participate in the ARIN policy process perceive that harm at the time of policy development.

I will point out that the last ARIN PPM had roughly 10x by my count the number of lawyers as any previous ARIN meeting I can remember (and I have attended the vast majority of them at this point). The financial interests looking to profit from the creation of this market (I believe they are trying to create it for their own benefit more than responding to its existence, in fact) are becoming more active in the policy process as they see this as the best way to achieve their goals.

>> It is only my opinion but the property argument is misleading. Some are
>> being led to believe that they can pay xxx amount of dollars for an IP
>> block and it will work across the Internet. Those dollars spent, or even
>> a US court order do not guarantee global connectivity to the Internet.
>> Correct me if I'm wrong, but there is a pretty large portion of the
>> Internet that does not care about what a US judge might decide. I think
>> we really need to weigh the costs and the benefits of the road we are
>> heading down.
> 
> [Milton L Mueller] I think if RIRs or whatever registration system evolves out of them accurately reflect the results of trades, then yes, damn right, someone ought to be able to pay xxx dollars for an IP block and those addresses should be as easily routable as they were before the trade. 
> 

If the block is the same size as it was before the trade, perhaps, but, there are other factors that I believe should govern the transfer of addresses.

However, often in these trades, blocks are subdivided from previous blocks and smaller blocks are, necessarily, harder to get routed than larger ones until the routing system is fundamentally changed. I doubt anyone would invest the effort necessary to do so for IPv4. I am hoping that this will get done for IPv6. It is sad and frustrating that IETF dropped the ball on this when developing IPv6.

Owen