On Oct 6, 2013, at 3:22 PM, William Herrin <[email protected]> wrote: > There is a subtlety here that is escaping me. If the addresses are > assigned to hosted infrastructure within the region then how are they > not used within the region?
Bill - Apologies in advance for length - the community is entitled to a complete understanding of how we administer number resources, and that appears to take more words than I expected at first.. Technical infrastructure can be used to justify addresses (for example, network routers, POP equipment, concentrators, management systems, IT systems, data center servers, etc.) We're very experienced with this, and often ask for equipment lists, invoices, etc. to confirm existence of the asserted equipment. When technical infrastructure grows, more IP addresses are needed for assignment to the technical infrastructure. Technical infrastructure exists in a region based on the location of the physical devices. Customers growth can also justify addresses (e.g. you assign addresses to an single customer, or customer growth requires that a certain dynamic pool assigned to customers be expanded) We are also quite experienced with the verification of customers, including obtaining customer lists, contact info, etc. When the number of customers served grows, more IP addresses are needed to assign to the new customers, or to assign to the dynamic pools for those customers. Customers exist in a region, based on their actual location of each customer. For example, if someone needs more addresses because they are adding new VPN concentrators (i.e. more actual equipment itself), then that is indeed technical infrastructure need based on where the equipment will be installed. i.e. if you buy stacks of new VPN concentrators and have no addresses to connect them to your infrastructure, it's a simple matter to request (per policy) additional addresses needed. We do this all of the time with the infrastructure side of DSLAMS, cable head-ends, wifi nodes, etc. If someone needs additional addresses for their _customers_ (including to expand a dynamically assigned pool), then we verify the customer growth as noted earlier, including the location of the customers. Requesters have to show that they have additional equipment or additional customers to obtain addition addresses. Both equipment and customers are associated with real-world addresses, and hence are within some region of the registry system. > What aspect of the draft policy would change that evaluation? ... The draft policy 2013-6 reads as follows: "In addition to meeting all other applicable policy requirements, a plurality of new resources requested from ARIN must be justified by technical infrastructure or customers located within the ARIN service region, and any located outside the region must be interconnected to the ARIN service region." The requesters who were previously discussed (i.e. "the 52" who have received allocations since mid-year), often have some existing infrastructure, are not adding anything more, and do not need more addresses due to their _technical infrastructure_ growth. They do have remarkable customer growth, and hence need additional addresses. Under the policy change proposed by 2013-6, we would only consider customers within the ARIN service region for this purpose, and would provide an allocation which they could justify based on that in-region customer demand. The proposed policy change makes clear that customers must be in-region to be considered. (The present policy does not have such a constraint, hence the approval of recent requests where the vast majority of customers are known to be outside the ARIN region.) We don't consider virtual "technical infrastructure" for assessing the need for addresses, even though service providers may use such when adding customers. The additional customers driving such virtual growth are readily verifiable. The alternative would be to consider virtual equipment (e.g. VM's) as actual technical infrastructure and that would effectively open the justification of unlimited resources by any party without any actual equipment or customer growth. If the desire was to simply clarify existing policy (without actually changing the current practice), it could be accomplished by dropping the "in region" requirement for customers as follows: "must be justified by technical infrastructure within the ARIN or by customers located anywhere as long as the addresses are managed and interconnected in the ARIN service region." (Note that such a change would make rather clear that nearly any heavily-utilized customer-driven IP address pool interconnected in the ARIN region can qualify for additional resources, which may or may not be a desirable outcome depending on one's individual perspective.) I hope this helps in clarifying what ARIN-2013-6 would be change to the existing policy, as it would make clear that customer growth in-region would be necessary to justify requests, whereas presently we have some folks requesting resources using nominal equipment within the region and backed by extensive customer growth external to the region. Please do not hesitate to ask if you have any further questions - it is indeed very important that folks understand how we implement the policy in order to fully consider any proposed policy changes, so thank you for your diligence on this topic. /John John Curran President and CEO 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: http://lists.arin.net/mailman/listinfo/arin-ppml Please contact [email protected] if you experience any issues.
