[ARIN-consult] Fee Schedule Change Consultation

[Owen DeLong]

On Oct 30, 2012, at 07:38 , John Curran <jcurran at arin.net> wrote:

> On Oct 30, 2012, at 9:24 AM, Owen DeLong <owen at delong.com> wrote:
> 
>> Under the proposed fee structure, this would go from <$1.4M to nearly $4M, mainly due to the increase as a result of charging per ASN ($1M vs. what is likely now $0). Given the number of ASNs listed, however, I suspect that this charge is distributed among both End-Users and ISPs and that ISPs pay for each ASN in addition to their IP Address subscriptions. (10,416 ASNs vs. 3,500 End-user ORGs). (Page 14 of the same document).
> 
> Owen - ISP registration service subscription plans include maintenance for 
> all number resources including ASNs, i.e. the 10,416 ASN maintenance fees 
> are simply for end-user held ASNs. 
> 

OK... It wasn't clear from the document that this was to remain the case. I am surprised to see such a high ratio of ASNs to end-user organizations.


>> In other words, the entire structure is a proposal to provide relatively small discounts to most ISPs at the cost of putting a huge additional burden on end-users and ISPs that happen to fall into the wrong side of fee-bucket-realignment.
> 
> What you characterize as a "huge additional burden" is the establishment
> of a maintenance fee for each resource held, as opposed to a single $100 
> maintenance fee for unlimited number of resources.  
> 

Nonetheless, that is, in fact, a multiplier of fees for most organizations.

> Any change to the fee schedule which establishes fees which increases with
> resources held will have this type of impact, but the alternative is a fee
> schedule which does not fairly recover costs in proportion to resources held. 
> As the vast majority of end-user organizations have a small total number of
> resources held, this is almost always a modest increase of several hundred 
> dollars per year.  If indeed we include one of each IPv4/IPv6/ASN in the 
> initial $100 maintenance fee, then even fewer organizations will see such 
> an increase under the proposed revision to the fee schedule.
> 
> As discussed, here is the distribution of total number of IPv4+IPv6+ASN 
> resources held by end-users under an RSA agreement (excludes LRSA held
> resources as their total fees paid will change over multiple years) -
> 
> # - Count of End-User Organizations with that total number of resources.
> 
> 1 - 11658

> 2 - 423
> 3 - 669
> 4 - 135
> 5 - 97
> 6 - 40
> 7 - 26
> 8 - 15
> 9 - 8
> 10 - 11
> 11 - 11
> 12 - 6
> 13 - 4
> 14 - 1
> 15 - 5
> 16 - 1
> 17 - 1
> 18 - 2
> 20 - 2
> 21 - 1
> 23 - 1
> 28 - 2
> 31 - 1
> 32 - 2
> 35 - 1
> 38 - 1
> 45 - 1
> 54 - 1
> 55 - 1
> 59 - 1
> 75 - 1
> 
> The large number of organizations with just 1 resource are due to the number 
> of organizations which simply have a single AS# (and likely are either legacy
> resource holders or using provider-assigned address space for connectivity)

In other words, a badly skewed artificially high number while most conventional end-user resource holders fit into the other categories.

I also find it interesting that the slide claims 3500 end-user organizations, yet the count above yields more than 12,000.

> 
>> I will point out that for the most part, only 56 very large ISPs are in this category and their fees double from $16,000/year to $32,000/year (with the possible additional cost of $100/asn on top of that).
> 
> Incorrect - more than 500 ISPs (out of a total of about four thousand ISPs) see 
> an increase under the proposed revision to the fee schedule, and this enables 
> the reduction in the fees seen by the fifteen hundred xx-small and x-small ISPs 
> (to $500 per year and $1000 per year respectively); this is one of the explicit 
> goals of the proposed revision to the fee schedule.

Of the 500, most are not very large. There are only 56 very large ISPs that see their fees increased.

This is from your own slide. In other words, the majority of very large consumers of address space get a discount on the backs of smaller organizations under this proposal.

Owen