[ppml] RFC 1744 and its discontents

Geoff Huston gih at apnic.net
Wed Apr 16 19:36:36 EDT 2008

k claffy wrote:
> On Tue, Apr 08, 2008 at 11:08:29AM +1000, Geoff Huston wrote:
>   Tom Vest wrote:
>   > "Address Ownership Considered Fatal"
>   > Yakhov Rekter
>   > March 31, 1995
>   > 
>   > ftp://ftp.ietf.org/ietf-online-proceedings/95apr/area.and.wg.reports/ops/cidrd/cidrd.rekhter.slides.ps
>   Both this reference and RFC1744 are illustrative of the observation
>   that study of this topic and the issues relating to the "role" of
>   addresses in the context of the Internet is not a recent occupation,
>   and that folk have been thinking about this for some years. So in
>   response to KC's comment in the policy session at the ARIN meeting
>   today, one can point to such studies (and related studies associated
>   with thge PIARA activity in the IETF in 1996) and say that there is
>   some level of longer term work in this area. Also, as far as I can see
>   both RFC1744 and the Rekhter document both are supportive of a lease
>   concept (Corollary 3, page 12 of the Rekhter presentation, RFC1744,
>   section 3, para 4).
> geoff,
> we're proposing to turn the economic architecture of the Internet 
> addressing (and as we know, that includes routing) system upside 
> down, and as backup material we're using two essays (ok one essay 
> and 15 slides) reflecting observations of two individuals, written 
> last century when circumstances were substantially different, with 
> no formal peer review, nor references to any related work in 
> economics or market/regulatory transitions?

I'm not sure that its a case of providing direction as you infer, or
following behind the actions of industry players. The transfer proposal, 
at least in the APNIC region, is a proposal to recognise in the registry
the outcomes of actions between address holders that result in the movement
of address resources.

I agree that the proposition of IPv4 exhaustion is one with many
unknowns, and it seems that many folk are looking at what others
are going to do before they react, and the resultant situation looks
a lot like an uncomfortable paralysis, at least today. That's
not altogether heartening to see.

> 1744 is a nice essay and would make nice phd thesis proposal, 
> but someone (or several) ought to go off and do the research 
> and write a few theses.   i believe ben edelman is doing good
> thinking and his writeups will be useful, but if we think
> that's sufficient, we're in a heap of denial. the kneejerk
> "but there is no truly related work; this is a whole new world!"
> position i keep hearing from some of the smartest people i know
> has me wondering how many days till men in suits come in and take over.
> i've only spent a few hours thinking about this, but i see at
> least 5 specific research questions that the registries should
> [get icann to use some of their $57M/yr budget to, or do themselves 
> if icann won't] sponsor and guide an interdisciplinary working group 
> to rigorously study and get peer reviewed publications for the 
> community to learn from before even launching any proposals:
> 	(1) comparison of ip address allocation and spectrum 
> 	allocation, including different models used in different 
> 	countries, and metrics for evaluation of efficiency
> 	and consumer surplus generated 
> 	(2) comparison to other industries privated in this country: 
> 	electricity, natural gas, trucking, airlines, telecom 
> 	(3) comparison to other industries privated in other 
> 	countries, esp G7
> 	(4) comparison to other industries privated in our own 
> 	field: ip transit, dns
> 	(5) comparison to other market reforms in last 200 years:
> 	russia, china, india, latin america.  
> for each comparison, the similarities and differences to address 
> markets should be compared, metrics of success proposed/described, 
> data gathered/analyzed, models built.  
> otherwise this exercise looks like promoting blatant cyberlandgrab, 
> which i don't believe is what any of the registries intend.
> (good intentions are not sufficient here, we also need good research.)

Yes, this is a significant problem, and it involves a broader cast of
folk that just the addressing community. As I said at the APNIC meeting last
September I believe that this does involve many interests, of which
the address community is but one grouping. There are legitimate public
sector interests, private sector interests, investment intersts, and so on,
and I suspect that part of this issue involves some very fundamental
questions about the nature of deregulation and the ability to phrase
longer term common beneficial interests in terms of shorter term competitive
factors in today's market for internet goods and services. So I agree 
wholeheartedly with the proposition that this topic would benefit from
further study and, in particular, comparative studies as you suggest here.

But, as you are aware, the processes that will lead to IPv4 unallocated address pool
exhaustion are inexorably grinding away, and we have very little time left if
we believe that there is a benefit in taking actions before the event.

I suspect that the post mortem analyses of this event will be many and varied, but
the question before us at the moment is a more pragmatic question of what can we do
as address registries in the interests of a cohesive single Internet that 
would be of common benefit here that would not endanger the internet at the same
time, and be achievalbe within the time available to us?



More information about the ARIN-PPML mailing list