[arin-ppml] fee schedule

[Matthew Kaufman]

On 4/7/2013 8:57 PM, John Curran wrote:
> On Apr 7, 2013, at 1:35 PM, Matthew Kaufman <matthew at matthew.at> wrote:
>
>> It is just ARIN trying to figure out how to maximize its own revenue.
> To be clear, the change in the Revised Fee schedule is revenue neutral;
> please review the slides referenced earlier for additional information.
> <https://www.arin.net/participate/meetings/reports/ARIN_XXX/PDF/thursday/curran_fee_schedule.pdf>

I wasn't trying to claim that the revised fees aren't revenue neutral. 
All I was trying to claim is that ARIN gets more with the current fee 
schedule than they would with everyone paying $500. And they can do that 
because some organizations are more than willing to pay a lot more than 
$500.

>> It knows that if it charged everyone the same amount, there would be some large organizations who could easily afford to pay more but from whom they aren't extracting any extra revenue.
> Actually, we could easily go to a single fee for all ISP regardless of
> size. This would result in significantly higher fees for smaller ISPs
> (as the per-ISP fee would approximately $2800 per year.)  This does not
> seem to be the right direction if we're trying not to burden smaller ISPs.

I believe this just proved my point above. Not finding the few who are 
willing to pay more and charging them more would result in $2800/year 
across the board fees. Which would then result in a bunch of small ISPs 
not bothering to pay that at all, which *wouldn't* be revenue neutral. 
(And then I suppose it would be even more than $2800/year, wouldn't it?)

I'm not saying ARIN is evil... far from it, ARIN is doing *exactly* what 
I would expect an entity setting prices to do. It has segmented its 
pricing such that some can pay more and others can pay less and the 
number of customers is high. Making the price $2800/year for the 
smallest ISPs would result in fewer of them being able to afford the 
service, and yet the ones paying $32,000/year under the new schedule are 
clearly ok with paying that much so that there's a $500 price point 
available for others.

The problem is that the current set of steps is based on something we 
shouldn't be messing with. There shouldn't be a /48 or smaller step and 
there shouldn't be a /36 to /48 step, because the existence of both of 
those has pressured the numbering policy into once offering /36 as an 
option instead of /32, and now trying to find something even smaller to 
offer... only because to have all of the small ones at the /32 step 
would (in order to be revenue neutral) require all of them to pay 
~$1500/year.

Find something else to base the fee steps on before we do something else 
stupid, please.

Matthew Kaufman