[ppml] FYI: LACNIC micro-allocation policy approved and implemented

[william at elan.net]

> >
> > Well, that is the risk you take. Like William said, ARIN VOTED for it
> > and THE AC went thier own way...
> That is a major distortion of the truth.
I was being heavely ironic in my original email and you're probably taking
it as hard hit, but nevertheless my email did not have any "major 
distortions", see below

> The only straw poll taken with respect to policy proposal 2002-3 at the
> Memphis ARIN meeting was 15-4 in favor of the ARIN AC looking into reducing
> ARIN's minimum allocation size (you can verify this for yourself in the
> meeting minutes,
This was the only question asked after the presentation of proposal and as 
such people are automaticly assuming this is regarding current proposal 
and in fact I've asked several people present at the meeting as to when 
they were voting for whois aup proposal and for 2002-3 if they considered 
that they were voting for original proposal (with modifications as necessary)
or for something else and everybody said they did in fact undertood it as 
continuation or finish up of work of original proposal and not something new. 

I did perhaps exadurate on concensus slightly but I clearly remembered 
that overwhelming majority of those who voted were for and per standard 
system if somebody does not vote, his voice is not really counted as being
for or against and concensus is estimated from those who do participate, so
if figures number of votes you showed is correct, we can say that 80% were 
for the proposal. Not sure what iis requirement for rough consensus but I 
believe 80-90% is somewhere close and 80% is overwhelming majority if 
somebody wants to put it this way.

> Furthermore, the ARIN AC has specifically NOT abandoned the issue of 
> reducing ARIN's minimum allocation size.  In fact, the AC has established a 
> discussion group that is actively looking into this issue, and has been 
> tasked with forming a proposal that addresses the concerns and wishes 
> raised by those on all sides of the issue (you may find evidence of the 
> formation of this group in the April 8 AC meeting minutes, and a report on 
> the progress will be in the minutes of our last meeting (5/8/2003), once 
> they are approved).

This is not how it was presented at the end of Memphis meeting, you actually 
said you're abandoning proposal and will instead work on proposal that will 
reduce minimum allocation size for multihomed organizations to /21 (currently 
/20) and afterwards I had private discussion with person from AC and he said
you were going to reduce to /21 keeping /22 as being minimum justification.
So my email was right on these issues and on what AC has proposed.

As for discussion group, we do have system of "open discsussion" and
this is not how AC operates so AC discussion group would not in my opinion 
be a good substitute for working group as a way to get final version of the 
proposal, instead if you want separate discussion of this proposal from 
everything else, you should establish it as open mailing list and invite 
people there but keeping it within AC and calling it a "discussion group" 
is not appropriate, you might as well just say AC wants to do it on  your 
own not taking into the account standard procedures for establishment of 
working group and proposals that are done on this mailing list. 

And no I will not consider AC doing on its own to be a substitute for what 
people have assumed would be a continuation of work on original proposal.

> Once again, I must emphasize that the AC is a group of elected volunteers 
> who are trying to meet the needs of ALL entities in the ARIN public policy 
> community. 
You can not meet needs of everybody on this particular issue, we'v had
discussion on this topic before - there will always be some against it no 
matter what, especially from large ISPs. If you try to meet needs of 
everybody on every issue you will undoubtfully end up having each issue
been under consideration forever as this one have been already. At some 
point you just have to make a choice if there is majority support on the 
issue or at least not a majority opposition (and no majority opposition 
is what ARIN usually uses as far as I can see for other policies being 
approved).

Again going into the issues, the reduction of mimimum assignments by one 
bit as had been proposedby AC is not a substitute for micro-assignments, 
this is completely different. Micro-assignments are really blocks of /24 - 
/21 as had been defined by all other RIRs in their micro-assignment policies.
The reason I supported /22 is that I thought if we make it too easy for 
very new company with no internet experience to get /24 it would backfire 
as there would be many who do not have good idea on how to use the space,
many who get the space and then abandon it and higher potential for abuse,
so I though that either we have to have some kind of a check system (like 
sponsorship I'v proposed - though I've to admit it probably looked way too 
complicated and may not be practical in its original form) or without that,
we'd have to go for higher block size and I was the one who originally 
proposed /22 as compromise for such micro-assignment policy.

But /21 is nowhere close for micro-assignments, at least for /22 with /23 
minimum we do actually have companies that need space and have already shown 
some knowledge of using bgp and ip allocations (by utilizing 512 ips which 
would be checked by ARIN) that can get their own block, but /21 is not 
like that and this would not be supported by majority who want 
micro-assignments (most of these already consider /22 to be too much of a 
compromise). 

Now I was privately asked if reducing both requirement (currently /22) and 
qualified minimum allocation size (currently /20) by one bit would be 
acceptable - it would be as as it would result in /23 requirement to get 
/21 block. But I see no reason to assign somebody /21 block immediatly, 
instead it would just be better to go with original proposal 2002-3 and
have /22 being assigned when /23 had been utilized. But if AC wants to 
propose just one-bit reduction everywhere (becoming /23 requirement when 
multihomed to quialify for your own block) then I would support that provided
that ARIN also looks into new schedule of fees for small assignments.

William