[ppml] Policy Proposal 2005-8: Proposal to amend ARIN IPv6 assignment and utilisation requirement - Last Call

Edward Lewis Ed.Lewis at neustar.biz
Mon Apr 17 09:40:19 EDT 2006


At 12:00 -0400 4/16/06, Christopher Morrow wrote:
>On 4/16/06, Randy Bush <randy at psg.com> wrote:
>>  > so if the word "recommendations" replaced "guidelines", would it work
>>  > better for you?  i.e.:
>>  > The following recommendations may be useful (but they are only
>>  > recommendations):
>>  > - /64 when it is known that one and only one subnet is needed
>>  > - /56 for small sites, those expected to need only a few subnets over
>>  >   the next 5 years.
>>  > - /48 for larger sites
>>
>>  very much
>
>I agree, I'd like to see a move away from the 'classful' ipv6 and to
>something more reasonable. 'recommendations' seems like a start.
>thanks!

I was all for 2005-8 
(http://www.arin.net/policy/proposals/2005_8.html) until I heard 
comments from Izumi Okutani (JPNIC) and the mic during the discussion 
following the presentation on April 10.

The point was that there is an operation benefit observed in Japan 
from having boundaries.  The benefit is, of course, not apparent when 
one looks at the architectural design of the protocol nor writing 
code that make use of it.  The benefit is in the business models.

E.g., if ISP A is dealing /52's and ISP B is dealing /56's, what 
happens when ISP B buys up ISP A?  How do you merge customers used to 
having /52's into a billing model that is based on /56's?

My take, which is not what she was suggesting, is that as a consumer 
who has a choice between DSL and cable modem, if I want to walk from 
one to the other, it would be beneficial to me if both were marketing 
the same prefix length.

I won't say that this point is a show stopper, nor is it a compelling 
reason to move away from 'classless' networking, but when I heard the 
comment I stopped intending to vote for recommending 2005-8, (I 
didn't vote against it either) until I was clearer on this.

I think that having set classes of size has a benefit, although not 
to the protocol.  I don't think the RIRs are in the business of 
setting the classes, nor are the RIRs "better business bureaus" for 
the ISPs.

I suppose that I agree with (I think it's Randy's) "very much" above. 
The "classes" ought to remain in print, but make it clear that these 
are what the membership senses are the right sizes.  The policy does 
need to be clear on what ARIN uses as a measure of utilization.

Looking at the URL above, it seems like the policy does do this - 
with the change of guidelines to recommendations.  So I guess I am 
for the policy so long as the recommendations stay documented.

-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Edward Lewis                                                +1-571-434-5468
NeuStar

Nothin' more exciting than going to the printer to watch the toner drain...



More information about the ARIN-PPML mailing list