[arin-ppml] New IPv4 Transfer policy

[Owen DeLong]

On May 10, 2011, at 12:23 PM, Stephen Sprunk wrote:

> On 10-May-11 13:49, Ted Mittelstaedt wrote:
>> On 5/10/2011 10:06 AM, David Farmer wrote:
>>> My comment was directed at what I felt was your accusation that you
>>> felt the LRSA was misrepresented and was only about the fees that a
>>> legacy resource holder should pay.
>> 
>> OK, but that still doesn't really answer my question of what about the
>> LSRA is so much better than the RSA for a company like Microsoft,
>> which probably spends more in it's annual budget for office birthday
>> parties than it would on ARIN fees?
>> 
>> In other words, if the fee difference isn't the issue, what is it?
> 
> The only other obvious difference is that resources under LRSA cannot be
> revoked.
> 
Cannot be revoked for non-utilization. They can still be revoked for
non-payment or virtually any other issue, identical to RSA. The
LRSA provides a specific exemption from utilization criteria.

> That's what dangerous about allowing LRSAs for NRPM 8.3 transfers: if
> the justification is discovered to be bullshit after the fact, there is
> nothing ARIN can do about it.
> 
Agreed.

>> The fee structure is, fundamentally, based on the notion that a
>> commercial money-making LIR should pay more than something like a
>> public university even though they may consume fewer IPv4 addresses.
>> 
>> Thus it is very irritating when a commercial entity like Microsoft
>> appears to get off scott free without making any kind of additional
>> compensation to ARIN for enabling it to essentially make a ton more
>> money than it's competitors.
> 
> I don't see how you get that from the fee structure.  Any org, public or
> private, getting an assignment pays $100 per year, and any org, public
> or private, getting an allocation pays a lot more.
> 
> Supposedly, this is because allocations imply a lot more work for ARIN
> processing sub-allocations and sub-assignments, whereas assignments
> cannot be sub-allocated or sub-assigned.
> 
Allocations also include membership which would cost an assignee
$500/year. That narrows the gap between an assignment
with membership ($600/year) and an X-Small ISP ($1250/year) to
just barely over 2x the ($650/year). I believe ARIN does more than
twice the work annually for an X-small ISP on average than for
most end users, so, I think this is reasonable.

I think the $500/year for membership is more than reasonable
when you consider what it includes. As an end-user organization[1],
however, I do not elect to join ARIN at that price, but, would do
so if it were less than $200/year. However, my primary concerns
with ARIN are the policy process which is fully open to me as
an end-user. I am less concerned about voting in the ARIN
elections. However, I am also currently entitled to do so as the
DMR for two subscriber member organizations.[2]

[1] End user Org ID's DELONG (RSA) and DELONG-Z (LRSA)
[2] Day job -- Org ID HURC and a consulting client that I am not
	at liberty to name here.

>> Do we really want an Internet with a few large LIRs and all the
>> smaller ones gone out of business?  Because that is what the ARIN fee
>> structure encourages.
> 
> That's a matter of the decreasing cost per IP for larger allocations. 
> It has nothing to do with how wildly divergent the allocation and
> assignment fee schedules are.
> 
Since there is no cost per IP, but, rather a cost per IP growth per year,
sort of, I think the entire topic is rather specious in nature.

Owen