[ARIN-consult] discounting registration fees for IPv6 assignments

Jesse D. Geddis jesse at la-broadband.com
Tue Oct 30 06:55:14 EDT 2012


John,

	Thank you for shooting that link over. Although it doesn't address the
kind of detail I was looking for as far as what certain classes of
customers cost vs what they contribute. However, with a flat fee this
really doesn't matter. As far as the tickets, the two tickets I used as an
example of mine were an allocation request and one related to my arin-rr
entries getting mucked up when the auth mechanism was changed by ARIN.
There are also several other questions it doesn't answers, like why does
ARIN have both an end user and service provider category when ARIN's model
is said to be a delegated one?

	Jimmy, with regards to what you're saying about block reservations I
think that's more relevant for IPv4 and much less so for IPv6. I don't
think ARIN is doing that anymore for IPv4 anyway given the resource pool
situation. My blocks certainly aren't contiguous. So I think that may be a
moot issue.

	What I'm wondering is if you were to take the current total count of
/48's issued (both end user and ISPs together) and the total ARIN is
projected to receive for IPv6 fees based on the proposed fee structure
(not including discounting) and divided them by each other what would that
number be per /48? n/48 / $ipv6 fees. Would the total be lower than the
perceived barrier for small organizations? I don't think the cost to
support those allocations will scale linearly. I think the cost will go
down. So I'm guessing with each successive year we are at this with a flat
fee that flat fee will become lower.

	High level, this is where I'm going with this. I think we could eliminate
the service provider category. I think we can eliminate the tiers. I think
it is certainly true for IPv6 more so than v4. I wonder if we can get to a
point where every end user has their own block that can be ported like a
phone number from carrier to carrier. I think we can get there. From a
technology standpoint I don't see any reason why that isn't possible
today. What makes it impossible is this delegated model where the carriers
are in control of the majority of end user address space. If our goal is
to get IPv6 out there I can't think of any better way than to eliminate
the very middlemen who have been dragging their feet on IPv6 rollouts for
friggin years from the process of getting the address space to the end
users and content producers.

	Jimmy offered a conspiracy theory of ARIN being beholden to carriers
because they'd pay too much in a flat rate model. However, the complete
opposite is true. Going in this direction the carriers lose their
relevance in IP management and they contribute very little.

-- 
Jesse D. Geddis

LA Broadband LLC
AS 16602




On 10/29/12 7:51 PM, "John Curran" <jcurran at arin.net> wrote:

On Oct 28, 2012, at 10:08 PM, Jesse D. Geddis <jesse at la-broadband.com>
wrote:

> Perhaps John could shed some light as to how these numbers have been
>arrived at.

Jesse - 

ARIN's core registry services could reasonably be considered proportional
to the number of prefixes served; costs which are comparable between IPv4
prefixes and IPv6 prefixes, and most definitely not proportional to the
number of IP addresses. Note also that core registry operations costs are
smaller than the policy & registry development costs (which benefits all
customers.)

A cost breakdown was presented last year at the ARIN 28 meeting in Philly -
https://www.arin.net/participate/meetings/reports/ARIN_XXVIII/PDF/friday/cu
rran_cost_breakdown.pdf)

> A way to find this out is to compare customers. In 2012 LA Broadband
> generated a total of 2 ARIN tickets. How much did that cost ARIN? How
>many
> tickets, on average, did Orgs with over say 5 million IP's create? How
> much time did they take to resolve? What is the dollar amount associated
> with maintaining those Orgs vs my 2 tickets? Did my org really cost 100x
> more per IP to maintain? My fee last year was $2,250. Did two tickets
>cost
> that? Unlikely.

You are likely to be surprised how few tickets many larger organizations
generate, as their personnel are often more experienced interfacing with
ARIN.  Finally, it's important to remember that tickets are generally
related to requests for issuance or transfer of number resources, and
wouldn't result in costs actually proportional to number of IP addresses
involved.

FYI,
/John








More information about the ARIN-consult mailing list