> On Sep 19, 2021, at 05:06 , John Curran <[email protected]> wrote: > > On 18 Sep 2021, at 11:05 PM, Owen DeLong <[email protected]> wrote: >> ... >> We’re taking about the fact that if you’re an end user, this latest fee hike >> is pretty abominable, while it’s a pretty nice subsidy for most LIRs paid >> for on the backs of the end users. > > Saying “if you’re an end user, this latest fee hike is pretty abominable” > conveniently ignores the fact that the fees go down for more than 3000 of the > smallest end-users organizations as a result of the change. > >>> You can assert that ARIN's costs are predominantly the result of “LIRs” but >>> that doesn’t reflect reality – many of our services and functions are >>> equivalent for an entire address block and only a small set of them are >>> related to subdelegation functions. >> >> IRR >> RPKI >> SWIP/RDAP/WHOIS >> RDNS >> Frequency of updates >> Frequency of additional requests/transfers/etc. >> >> are all impacted more by LIRs than by end users. > > Quite possibly, but it is also true that organizations that have larger total > holdings result in more costs than those with smaller holdings – something > that the prior end-user fees did not reflect.
Frankly, I’d be fine with going to scaled end user fees so long as that scaling didn’t jack them up to be equal to LIR fees for the same amount of space as there is a clear difference in usage, revenue generation, and required ARIN service level. >> On the other hand, there’s also the fact that the LIRs are (generally) using >> their addresses directly for profit (i.e. their business _IS_ (at least in >> part) providing IP addresses to their customers). Essentially, they are >> reselling RIR services. > > In the ARIN region, LIRs provide connectivity services and the registry > services are yet another cost in operating a network. ARIN doesn’t provide > connectivity services, and customers don’t go to Internet service providers > because they view them as “resellers of ARIN services” but rather to obtain > Internet services. OK, so LIRs don’t charge separately for addresses, they include them with connectivity? Want to bet me my next 5 years of ARIN fees that’s not universally true? >> What you’ve done here, is to make everyone pay the same (within any given >> resource category ... > > Correct - everyone pays the same for a given service category, and these > fees double for every 4x increase in total number resource holdings. > Customers can utilize the registry services they wish (RPKI, IRR, etc.) or > not based on their perceived need and value without ARIN telling them how > ARIN thinks their operation represents as a business. A fee schedule that > ignores the size of total resource holdings (as the end-user schedule was) is > not fair nor is one that is linear to number of IP addresses represented by > the registry entries. No… Everyone pays the same for a given SIZE category. End users don’t aren’t getting the same services as LIRs. That’s the point here. As to your statement about the rate of doubling, that’s only true for IPv4. For IPv6, it doubles every 16x. There are certain ARIN services that are most definitely LIR specific and of utterly no use to end users. Making end users subsidize those services for LIRs is also unfair. >>> Furthermore, there are costs that ARIN incurs as a result of customers that >>> have no relation at all to the customers individual utilization of services >>> or their choice to subdelegate, but still must be recovered (e.g. costs of >>> responding to customers on mailing lists…) >> >> If the costs of responding to members on mailing lists is a significant >> fraction of the ARIN budget, something has gone horribly wrong somewhere. > > It appears to be just select members heavily utilizing the feature, so > hopefully doesn’t add up to a significant fraction - however, the imputed > costs must still be recovered and do not correspond to individual utilization > of registry services or their choice to subdelegate. i.e., It is inherent > that cross-subsidization occurs with any fee schedule and all ARIN can do is > have a clear and fair methodology for cost-recovery that is used for all > customers. Yes, let’s create one of those that is actually fair to end users and isn’t built on the backs of forcing them to subsidize the services only LIRs use. Perhaps keep the same overall fee structure, but bill end users at 0.6 * LIR or whatever multiplier makes sense based on the relative costs of the total array of ARIN services vs. the subset of services end users utilize. I.e. if the total ARIN costs for providing all services are X, the following multipliers could be derived: Costs fixed per customer (services that don’t scale with address size, such as office overhead, billing, etc.): Y — Multiplier for this fraction: Y/X = Multiplier F Costs for other services that only LIRs utilize: Z — Multiplier for this fraction Z/X = Multiplier L Costs for other services that every resource holder may utilize: W — Multiplier for this fraction: W/X = Multiplier E F+L+E should add up to 1. Work out the fee tiers such that EU*(F+E) + LIR*(F+L+E) = cost recovery, then bill end users and LIRs accordingly. That would be fair. The current rate hike to subsidize LIRs on the backs of end users is not. Owen _______________________________________________ ARIN-PPML You are receiving this message because you are subscribed to the ARIN Public Policy Mailing List ([email protected]). Unsubscribe or manage your mailing list subscription at: https://lists.arin.net/mailman/listinfo/arin-ppml Please contact [email protected] if you experience any issues.
