[arin-ppml] Set aside round deux

[Andrew Dul - andrew.dul]

On Fri, 30 Jul 2010 09:56:07 -0400, "Hannigan, Martin" <marty at akamai.com>
wrote:
> On 7/30/10 1:18 AM, "Andrew Dul" <andrew.dul at quark.net> wrote:
> 
>>    Has anyone considered that maybe what we should do is just reserve
the
>> block for future use and not try to predict the future too much at this
>> point
>> and instead comeback after IPv4 exhaustion and then write an
appropriate
>> policy.  Yes this does kick the ball down the field, but it also allows
>> the
>> ability to have space to work with in the future to use for an
>> appropriate use
>> after IPv4 run-out.
>>  
>>  Andrew 
>>  
> 
> 
> Hi Andrew,
> 
> Guess I'm not sure what would be speculation or not. I think that there
are
> some facts at hand:
> 
> -  ipv4 addresses will run out at the IANA this year
> -  at least three RIR's (ARIN, RIPE, APNIC) will run out of ipv4 address
>    space next year
> -  networks will need to find ipv4 addresses in order to grow and
> transition
> 

I agree with your facts here.

> 
> We could modify such a proposal later if it were posed and succeeds, and
> finally, if something is broken horribly and there's consensus the BoT
> could
> act (and probably not be vilified).

I'd prefer the community to develop policy rather than having a policy
which pushes the decision to the BoT.
 
> Setting aside a block for future use doesn't seem prudent to me. The
time
> to
> take it from "reserved" to "in use" would be significant. The fact that
we
> don't have policy already that sets a direction for transition is
troubling
> and I see the opportunity to do that with a solid set-aside proposal.
> 

I see the current NRPM 4.10 policy as the policy which should be part of
the transition process.  While 4.10 is probably not a perfect policy it
does seem like it gives an option to a set of orgs after IPv4 run-out.   

Mostly what I was seeing in my reading of the mailing-list was not a lot
of consensus on what should be done right now in the use of the last /8. 
Granted "reading" the mailing-list is an imperfect method of determining
consensus, and maybe you have a different opinion that consensus is
building around a particular policy.  So I'm not specifically opposed to
your policy proposal, just that maybe now isn't the best time to figure out
how to slice it up since we don't have a large number of organizations
deploying transition technologies...yet...


On Thu, 29 Jul 2010 22:28:28 -0700, Scott Leibrand
<scottleibrand at gmail.com> wrote:
> Joe proposed that for the whole /8, and it didn't get much support. How
> big a block were you thinking?
> 

Today, 4.10 sets aside a /10 or 1/4 of the last /8.  If we think we needed
additional policy now to use some of the remaining /8 in a different way
other than 4.10, using another /10 now seems reasonable and reserving the
remaining /9 for the future.  We could even codify that the remaining part
of the last ARIN /8 is released after 4.10's current /10 is used up or some
other timing mechanisms.  Since 4.10 is rate-limited to 1 block of a small
size every 6 months, then its use will probably be pretty consistent once
we start allocating/assigning from the 4.10 block.



On Fri, 30 Jul 2010 00:17:43 -0700, Owen DeLong <owen at delong.com> wrote:
> I think to some extent 4.10 attempted to do that.
> 

I agree 4.10 is what we have now to assist through the transition phase.  

> I think that 116 is an attempt to do a little more of that. As it
> currently stands, there are no criteria
> for 4.10 so arguably ARIN staff could either not be able to issue space
> for any technology under
> existing 4.10, or, they may not be able to make any determination that
any
> particular claimed
> usage is not transitional.

I understand your concern here, but I also think that the current 4.10
text is OK.  I haven't talked with ARIN staff specifically about the 4.10
language and their concern so perhaps you have more insight here, but given
that the usage is rate-limited and small blocks I don't necessarily believe
that what is there now won't work.

> As such, I think we need to do something to shore up 4.10. 116 was an
> attempt to do that while
> not restricting the policy only to the technologies known now.

Having just gone back and read 116 and 4.10, I can see why you want to add
a few more restrictions to 4.10, but 4.10 seems reasonable to me.

I guess to me 4.10 seems "good enough" right now.  

I have a little bit of a problem with the 116 text in that it attempts to
specify the technologies, but then has "etc" and "other transitional
technologies at the time of proposal" as escape clauses.  Seems to me that
the current 4.10 language of some examples and staff discretion is OK.

Andrew