[arin-ppml] Policy Proposal: IPv4 Recovery Fund / history lesson #1

[Tom Vest]

On Nov 24, 2008, at 10:10 PM, Tom Vest wrote:

>
> On Nov 24, 2008, at 6:24 PM, Milton L Mueller wrote:
>
>>
>>> -----Original Message-----
>>> From: Tom Vest [mailto:tvest at pch.net]
>>>
>>> I will add only that my own use of the term "empirical" implies open
>>> testing of assertions against potentially disconfirming
>>> facts, and use
>>
>> which is why I encourage you to actually read some of the reports and
>> papers.
>>
>>> of observation/measurement methods with known limitations/biases --
>>
>> as opposed to your own theses, which cannot be tested...;-)
>
> I'm pretty sure that my predictions are familiar to most community  
> members by now.
> I continue to stand by them publicly. If the predictions turn out to  
> be right, perhaps you'll reconsider your assertion, and my "theses"  
> -- just as I will be obliged to if they turn out to be completely  
> off base.
>
> If you have any non-tautological predictions to make (i.e.,  
> "'markets' are good, therefore whatever resulted from a 'market' was/ 
> is good"), then I think putting them on the record publicly here  
> would be a fine way to putting them to the same test.

Did you have any thought on the second question?

> Here's a more serious historical puzzle for you. Once NAT was  
> invented -- about six months after CIDR was codified** -- why didn't  
> existing Internet operators just run out the remaining public IPv4  
> addresses as quickly as possible? Sure, many thousands of new,  
> equivalent "peer" institutions were created in the intervening years  
> (including 90% or so of current ARIN members), but so what? After  
> all, using NAT plus any arbitrary non-unique address space that can  
> only be routed within a narrow, segregated routing domain would have  
> provided all of the same opportunities then that a pure IPv6-based  
> operator could enjoy today. Why do you suppose that that  
> "opportunity" wasn't compelling back then, but it is today? What  
> makes a CIDR (like) solution, which turns a smaller prefix into a  
> globally valid (and potentially "provider independent") locator for  
> a mix of public and private end-point identifiers, any less  
> important and/or any less attainable today than it was a decade ago?
>
> **RFC 1519 - Classless Inter-Domain Routing (CIDR), September 1993.
> RFC 1631 - The IP Network Address Translator (NAT), May 1994.