[arin-ppml] Support proposal 2010-2: /24 end-user assignments
tedm at ipinc.net
Mon Apr 19 13:53:34 EDT 2010
Are you sure your not referring to 2010-5 since it uses the same
I can't speak for 2010-2 since I wasn't involved in that one but I did
author policy proposal 102 in conjunction with Chris Grundemann which
was reworked into 2010-5
Neither Chris or I considered the issue of the time horizon for the
beneficial effects for the simple reason that the return-and-renumber
requirement is a moral obligation more than an attempt to clean the
dfz. In the case of 2010-5 it would be a wash anyway, because that
opens the door to small ISPs who currently DON'T qualify to obtain
numbering - thus it's kind of, well this year we will allow extra
fragmentation but next year when your bigger we will take it back.
For 2010-2, even though the goal, is to defragment
the dfz, I feel that a much more overriding issue is that the community
has a moral obligation to Do The Right Thing, as it were.
From a forest-vs-tree perspective, it may not result in a
few additional percent of fragmentation in the dfz to allow small
holders to make non-contiguous, fragmented advertisements. I think
what your getting at here is the argument that this polity institutes
a rule (return-and-renumber) that costs a few members greatly, but
saves the vast majority nothing.
But from a tree perspective this is the same argument that well we
should just go ahead and allow all families with 13 or more children
free admission to Disneyland since families that large are cash-tight
and there's so few of them that this won't increase ticket pricing.
It isn't morally right to allow a small ISP to "get away" with
fragmented advertising of multiple small blocks, when that ISP, with
a bit of discipline, and a bit of work, could renumber into a single
larger block. It doesen't matter that there's so few of them in
this criteria that you can make a macro economic argument that it's
not worth the trouble to the community to enforce this.
My personal view is that this should have been done a decade ago
for that reason. The fact that it wasn't shouldn't be used as a
justification that well we let them "get away" with this for
so long, it's futile now to close the barn door, the horse was
stolen a decade ago and has gone to the glue factory by now. I
also don't see that renumbering is that much of a burden - I've
done it with a far larger network than what is being discussed
by this proposal and it's just like eating an Elephant. It's no
problem if you do it a bit at a time.
On 4/18/2010 10:57 AM, Milton L Mueller wrote:
> I have a question about the anticipated time horizon for the beneficial effects of 2010-2. This policy is about cleaning up the ipv4 space by consolidating smaller assignments for multihomers and involves returning and in some cases renumbering. For renumbering cases a one-year cycle of dual assignments is involved. The question: how many years do you think it would take for this policy to have an appreciable effect on aggregation? And do you think would that effect be one of slowing table growth or actually reducing its size? Please note: I am not questioning the beneficial effects, just interested in their anticipated magnitude and the speed with which they would happen.
> Milton Mueller
> Professor, Syracuse University School of Information Studies
> XS4All Professor, Delft University of Technology
> Internet Governance Project:
>> -----Original Message-----
>> From: arin-ppml-bounces at arin.net [mailto:arin-ppml-bounces at arin.net] On
>> Behalf Of William Herrin
>> Sent: Saturday, April 17, 2010 8:05 PM
>> To: arin-ppml at arin.net
>> Subject: [arin-ppml] Support proposal 2010-2: /24 end-user assignments
>> Hi Folks,
>> Yesteryear's attempt to restrain the mess the IPv4 BGP table has
>> become by setting RIR minimums to /19 or /20 within particular /8's
>> utterly failed, largely because small multihomed end users had no
>> choice but to get /24's from their ISP -- in the middle of the
>> restricted /8's. Though late in the day, proposal 2010-2 largely
>> corrects this grave mistake for new assignments -without- expanding
>> who can introduce BGP routes into the table or how many IP addresses
>> anyone qualifies for. It is a major step forward and I strongly
>> encourage you to stand in support of it at the ARIN meeting next
>> For your reference: https://www.arin.net/policy/proposals/2010_2.html
>> Bill Herrin
> You are receiving this message because you are subscribed to
> the ARIN Public Policy Mailing List (ARIN-PPML at arin.net).
> Unsubscribe or manage your mailing list subscription at:
> Please contact info at arin.net if you experience any issues.
More information about the ARIN-PPML