ARIN XX Public Policy Meeting Day One Notes - 17 October 2007

Meeting Called to Order

Speaker: Ray Plzak, ARIN President and CEO


Ray Plzak opened the meeting and welcomed those in attendance.

Ray began the first day of the ARIN XX Public Policy Meeting by making meeting-related announcements and introducing those present from the ARIN Board of Trustees, the ARIN Advisory Council, and the Number Resource Organization (NRO) Number Council. He also acknowledged those colleagues from the other Regional Internet Registries (RIRs) in attendance and introduced ARIN's directors. Ray noted that other features of the meeting included the Registration, Billing Services, and ARIN Election Help Desks, the Cyber Café, and a First Timer Lunch. He also announced this was the first time there would be an Advisory Council Information Desk for those attendees with questions for the ARIN AC.

He continued with the following announcements:

  • 203 total registered ARIN attendees, including approximately 78 first-time attendees. NANOG 41 had 452 attendees, and there were 123 individuals attending NANOG 41 and ARIN XX jointly.
  • Attendance, as analyzed the previous day, is comprised of 162 attendees from 30 different U.S. states and the District of Columbia, 10 attendees from 4 Canadian provinces, 1 attendee from the ARIN portion of the Caribbean region, and 30 attendees from 15 different countries outside the ARIN region.
  • Thank you to ARIN XX network connectivity sponsor One Connect IP.
  • Discussed the First-Timer Luncheon and held the First-Timer Luncheon Raffle, which was won by Brian Murphy-Dye.
  • Introduced John Curran, Chairman of ARIN's Board of Trustees, for the purpose of reviewing the rules of discussion to be followed throughout the meeting.

Ray reviewed the meeting agenda and the printed materials given to each attendee. He then introduced Steve Ryan, ARIN Counsel, to speak about some legal matters including ARIN's first affirmative litigation in the matter of the domain registration and the Legacy Registration Services Agreement.

At the beginning of the meeting, approximately 168 people were in attendance.

John Curran, as Chairman of the Board, moderated discussions throughout the day.

[ARIN offered the opportunity for remote participation throughout the meeting. Comments from remote participants are read aloud at the meeting and are integrated into the meeting report. There were approximately 17 individuals viewing the meeting webcast throughout the proceedings, and 13 registered remote participants.]

NRO NC Election - Candidate Speeches

Speaker: Ray Plzak, ARIN President and CEO

Presentation: PDF

Ray Plzak provided information on the election process of the ARIN region's representative to the NRO NC for a three-year term to fill the seat currently held by Sanford George. He provided a listing of the candidates, and noted the biographies and statements of support for the candidates are available on ARIN's website, at the Election Help Desk on-site at the meeting, and as a hard copy in the folder given to meeting attendees. He concluded his presentation with an explanation of the method of voting, the qualifications for voting, and when voting could occur.

Ray Plzak introduced the candidates for the NRO NC, displaying the candidates' statements describing their motivation to serve and their biographies. Candidates present at the meeting were given an opportunity to speak.

  • Andrew Dul - present, introduced himself and offered a statement on his qualifications
  • Vikas Khanna - not present, motivation to serve and biography presented
  • Jason Schiller - present, introduced himself and offered a statement on his qualifications
  • Michael Smith - present, introduced himself and offered a statement on his qualifications
  • Sam Weiler - present, introduced himself and offered a statement on his qualifications

Ray highlighted the URL to more detailed biographical information and statements of support and asked that the meeting attendees review this information before voting. He also explained that the Election Help Desk was available for attendees with questions about voting.

Regional PDP Report

Speaker: Einar Bohlin, Policy Analyst

Presentation: PDF

Einar Bohlin presented a summary of policy developments and discussions for ARIN and the other RIR regions. Policy issues common among more than one region include:

  • Global policies
    • End Policy for IANA IPv4 allocations to RIRs, "N=1"
    • Global Policy for the Allocation of the Remaining IPv4 Address Space, "N=2"
    • IANA Policy for Allocation of ASN Blocks to RIRs
  • IPv4 policies
    • 12 Month Allocation Period
    • eGLOP (Multicast Address Space)
  • IPv6 policies
    • End sites can request allocations
    • Amend IPv6 assignment and utilization requirements (.94 and /56)
    • Provider-independent assignments for end users (PI IPv6)
    • Removal of plan to assign "200 /48s"
    • Removal of "interim status"
    • ULA Central

Einar also explained how AFRINIC was currently discussing adding a moderator group to their policy development process, which would serve a function similar to that of the Advisory Council in the ARIN region. He concluded by noting that an RIR Comparative Policy Overview was available on the NRO website, and that it included comparisons of the Internet number resource policies in all the RIRs.

Discussion Overview [Transcript]
  • An attendee inquired if this presentation would soon be available on the ARIN website. Ray Plzak replied that yes it would, and noted to newcomers that this was a standard element of ARIN meetings.

Internet Number Resource Status Report

Speaker: Leslie Nobile, ARIN Director of Registration Services

Presentation: PDF

Leslie Nobile presented an NRO report on the current status of all IPv4 address, IPv6 address, and Autonomous System number resources as of 30 September 2007. This report is updated quarterly by all the RIRs and provides up-to-date statistics on rates of IPv4, IPv6, and AS number resource consumption.

Ray Plzak added that this report is also given at every RIR meeting, in addition to ARIN meetings, and that a portion of it is presented at every ICANN Board meeting as part of the ASO report. He went on to say that each RIR updates its statistics regarding the allocation of a number of resources on a daily basis, and that quite a few researchers use these statistics regularly.

There were no questions or comments from the floor.

2007-19: IANA Policy for Allocation of ASN Blocks to RIRs

Speaker: Axel Pawlik - Proposal Author

Introduction: PDF
Presentation: PDF

Ray Plzak presented an introduction to the proposal. Highlights included:

  • Advisory Council shepherds: Dan Alexander and Heather Schiller
  • Introduced on PPML on 25 July 2007, designated as a formal policy on 28 August 2007. This will be its first discussion at an ARIN meeting. No revisions have been made.
  • Presentation of Staff Impact Analysis, Legal Review, Staff Comments, and overview of PPML discussion activity. [see presentation and transcript links above for details]

Axel Pawlik, as the author, continued with the presentation of the proposal. He provided details on the text and rationale of the proposal.

There were no questions or comments from the floor.

End of Discussion:

John Curran took a straw poll to determine the sense of the room. After the tally was conducted, John stated that this information would be provided to the Advisory Council for use in its deliberations.


Speaker: Ernest Byaruhanga, AFRINIC Registration Service Manager

Presentation: PDF

Ernest Byaruhanga began his presentation by discussing the membership and resource allocation growth at AFRINIC. Ernest continued by presenting information on AFRINIC's IPv6 efforts and fees, training efforts, the AFRINIC LIR portal, policy developments, and AFRINIC Public Policy Meetings.

  • From 2005 to now, AFRINIC has more than 300 members, excluding the legacy members that were brought over from the other RIRs.
  • AFRINIC's Board passed a resolution noting the issue of IPv4 exhaustion, and that efforts to deploy IPv6 should be intensified. In addition, the IPv6 fee schedule was reviewed by the AFRINIC Board, AFRINIC staff has stepped up outreach efforts relating to IPv6 including holding a press conference, and staff has launched an IPv6 information page on the AFRINIC website and planned for more IPv6 training.
  • MyAFRINIC, a web portal to allow LIRs to easily interact with AFRINIC and manage their resources, was released during the AFRINIC-7 meeting.
  • Next AFRINIC meeting will be AFRINIC-8 in Rabat, Morocco from 31 May to 6 June 2008, and will occur back-to-back with an IPv6 conference, AfNOG, and INET Africa meetings.

There were no questions or comments from the floor.

Panel: "IP Markets?"

Moderator: Paul Vixie, ARIN Trustee
Panelists: John Curran, Ben Edelman, Tony Li, Steve Ryan


Ray Plzak introduced Paul Vixie as the moderator of the panel. Paul began with an introduction of the scope of the panel discussion and introduced the panel members. Paul made an opening statement, and then invited the panelists to make opening statements on their own views about what they see as the issues regarding the shrinking pool of IPv4 addresses and the possibilities and consequences of markets developing for the purpose of buying and selling IPv4 addresses.

The panelists' opening statements may be read in the meeting transcript. A discussion period then took place.

Discussion Overview [Transcript]
  • An attendee stated that he believed it was in the best interests of everyone to avoid this becoming an adversarial situation. He then inquired of Tony Li whether, in his estimation, NAT IPv4 or NAT-PT was most likely to cause routing table growth. Tony replied that he didn't believe the type of NAT necessarily mattered; the results would be the same. He went on to say that problems could also arise with the IPv6 routing table, as the routing policies aren't any different. John Curran noted agreement with the attendee, stating that a cooperative solution was very important.
  • An attendee asked for clarification on how well a white market in IP addresses would control deaggregation and routing table growth if there was still a black- or gray-market where organizations could go to avoid the oversight implied in a white market. Ben Edelman responded that it was his opinion that if the rules were fair, they would be followed, noting that it was generally only when rules were perceived to be unfair that people didn't follow them. Paul Vixie added that, on the issue of ARIN as an organization possibly ignoring whatever market may currently exist, the documents that provide the framework for the organization operationally are all in agreement that IP address are not property. He pointed out that while those documents can be changed, under the current restrictions of those documents, ARIN stance on the issue is clear.
  • An attendee, noting his concern that a move to market-based handling of IPv4 addresses could mean increased involvement and regulation by many governments, asked the panel for their views on whether this would also mean increased government oversight of IPv6 addressing, and a move away from the current paradigm of a self-governed community. John Curran replied that he thought the two protocols were different enough, due to the much larger pool of IPv6 addresses, that the same possible market-based dynamics would not occur. He went on to add that he believed that the current situation is, in his opinion, a temporary one and that much of this may be moot in twenty years. Paul Vixie followed-up with a question for Tony Li, asking if he believed that with greater adoption of IPv6 we'd see a move away from a market-based approach to addressing. Tony replied that while IPv6 had many more addresses than IPv4, they could still have value and the value of routing slots in IPv6 would be even higher.
  • An attendee stated that there is empirical evidence that small and medium size blocks have been traded more or less openly, but asked what about large blocks? He went on to ask the panel if they had seen any evidence that legacy holders who have large blocks would actually be willing to trade them as large blocks, or would they simply split them up into small blocks? Ben Edelman noted that yes, it was his opinion that large blocks would have a disproportionately large cost premium, and there will likely be pressure to break the blocks up. He concluded by saying he did not see a future for operators that need more /8's to operate, and that there would be greater pressure on large operators to move to IPv6 or NAT. John Curran added that because of how the addresses would likely be used, there would definitely be more value in splitting a large block into smaller ones, as in the presence of a market economy for IP addresses, the monetary return would be much larger. He also noted that while we currently have the potential for a market in IP addresses, there is no real expectation currently of a market in routing table entries, and that could setup an inherently unbalanced market.
  • An attendee asked the panel for their interpretation of need-based allocations, specifically how need is evaluated. John Curran replied that the basis for the RIR policies is RFC 2050, where documented need is described as an ability to use the IP addresses to number network components for the purposes of carrying traffic.
  • An attendee stated that he thought that the vast majority of the people in attendance, when they hear "market," are not thinking about an exact continuation of our current allocation policies, and asked the panel to talk about the difference between the utility value and the speculative value of IPv4 addresses and how we might keep addresses available at or near the utility value. Ben Edelman responded that the speculative value should reflect the expected long-run utility value, but that as a long-lived asset, an IPv4 address should last for as long as anyone is using IPv4. He added that the tempering effect would be that there would still be a relatively large and fluid supply, and so there would be price competition to push prices down, as long as rules on transfers didn't severely restrict the market.
  • An attendee asked the panel if they were aware of any technologies that could have an impact in this area. Tony Li responded that we have a routing architecture that is largely hierarchical, and that he didn't see any proposals or technology that would substantially alter that in any significant way.
  • An attendee asked the panel if they thought that continued emphasis on trying to extend the IPv4 address pool through policies and procedures was going to invite any kind of regulatory pressure from governments. Steve Ryan responded that, at least in the case of the U.S. government, their involvement has been benign and that it could reasonably be assumed that governments will either use existing laws or create new ones to regulate our industry. In the course of later discussion, this issue was raised again when an attendee disputed Counsels claim that the U.S. government involvement in the Internet has been entirely benign and asked what metrics such an assertion was based on. Paul Vixie replied that ARIN Counsel was not in the room and able to respond, but that Counsel was stating his own views, and was likely speaking in relative terms to how the government has regulated other industries. Paul suggested to the attendee that they contact Counsel themselves to request that data.
  • An attendee noted that it appeared that the longer the transition from IPv4 to IPv6 takes the more important all the issues become, and asked the panel their views on initiatives to encourage adoption of IPv6. Ben Edelman responded that the current microeconomic incentives for IPv6 were troubling as they appeared not to do enough to encourage broad adoption, which is a requirement for complete IPv6 deployment.
  • An attendee asked Ben Edelman to reconcile the apparently contradictory views that the creation of an IPv4 white market would both help solve the issue of IPv4 depletion and help encourage organizations to deploy IPv6, adding that in his opinion the IPv4 market would discourage deployments of IPv6 and that it would force the RIRs to assume more powers and more duties of enforcement. Ben responded that either way, the RIRs will have more work to do -- either to keep a market "white" or to forbid gray or black market transactions. John Curran added that, in his view, the community is faced with either running out of freely available IPv4 address space and things get exciting for those still using IPv4, or that the IPv4 run-out occurs however, IPv4 addresses are still available through an accepted mechanism as yet to be determined. He added that this would in fact mean that IPv4 address space could continue to be obtained, but at a higher cost, and there are possible side effects to that, including a perception that IPv6 deployments could be put off for another time.
  • An attendee asked in the theoretical situation of a constrained market, is there some point where we could actually set the prefix boundary, that given the amount of likely unused available legacy space out there, we could avoid seeing routers become overwhelmed? Paul Vixie replied that he viewed this as a tuning parameter and that it could be used to control the churn-rate and the rate of new entries in the routing table. John Curran added that in what Tony Li proposed, if routing slots were subject to a market economy as well, using the prefix boundary as a tuning parameter would no longer be necessary. He concluded by saying that if this would be a system with an inherent feedback loop and that absent such a feedback loop, we'd be left with only these artificial boundaries to create back pressure on the routing tables. Bed Edelman commented that it would be useful to perform a thought experiment where if legacy holders were willing to return /8s right now, would we be as worried about routing table growth if it were handed back out by the RIRs in the usual sizes of a /20 or /22? Probably not he said.
  • An attendee inquired if the leasing of IPv4 addresses, instead of the buying and selling of them, had any legal or economic benefit. Ben Edelman responded that there is some economic question behind the idea of leasing in particular; specifically, who has the residual claim to the asset at the end of the lease if it turns out the asset is worth more at the end of the lease than the beginning. He added that he hadn't given much thought as to how that would be setup. The attendee added perhaps that approach might be tried whether policy allows it or not. John Curran added that he wasn't sure the value that arrangement would add to the community would outweigh the complexity it introduces. In later discussion, the issue of leasing was raised again by another attendee who asked how that differed from the current process of subdelegations via SWIP. The attendee added that in the case of legacy address space holders, the issue of whether they could be considered assignments or allocations was not well understood. The attendee concluded by stating that when IPv4 address space is transferred, the ability or inability to reassign that to customers is going to be important and also has some bearing on the fee structure that we're going to charge the legacies for those blocks. John Curran responded that in his personal opinion, the address space held by legacy address holders is often very well utilized, but that it is important to recognize that we want to set up a system, whereby a legacy address block holder, who recognizes it doesn't need part of its space can put that into the system, so it can be used for someone who does need it for networking and connecting to the larger Internet. He added that there is no reason that once those blocks have been transferred to someone who needs them for connecting to the Internet that they should be treated any differently than any other address block. He concluded by saying that the community needs to establish a system where those with need can still acquire address space, and not just a market where blocks are transferred for purely motives of profit or market dynamics.
  • An attendee offered up for discussion the idea of ARIN acting as a middle-man in a potential IPv4 address market, allowing ARIN to continue to allocate IPv4 addresses based on need as they have and act as an intelligent buying pool effectively for the community. Ben Edelman responded that was an interesting suggestion, but that the community might be more comfortable with the transaction taking place between the buyer and seller, as otherwise it could be perceived as an even bigger change in ARIN's role.
  • Paul Vixie asked the panel if, while contemplating the idea of a transfer economy for IP addresses, are you thinking in terms of a stable long-term market where IPv4 survives as an stable economic force where we live in IPv4 and bypass IPv6, or are you looking at the IPv4 market as a transition tool for making the transition fromIPv4 to IPv6 survivable? Ben Edelman replied that he thought either hypothetical could possibly be the case, and that either way, he believed market-based transfers were good as they allowed the community to move down a path that didn't preclude either case, pending a decision by the community on which they'd prefer. John Curran added that in viewing what the Internet would look like in twenty years with full IPv6 deployment, the pressures we currently operate under in conserving address space will not exist and that with the exception of how to deal with routing, address blocks could be much more freely available. He concluded by saying that in regards to the continued use of IPv4 in that time frame, it simply wouldn't be feasible when taking into account the continued growth of the Internet, and that he was worried that a market just facilitates a circumstance where if you tried to keep IPv4 running via a market, we'll just get deaggregation and unbelievable routing growth.
  • Paul Vixie asked Tony Li about a comment he once made that characterized IPv6 as being too little, too soon and what did he think of it now? Tony responded that he felt it even less way too soon. He went on to say that, with one exception, he'd never seen a network technology actually get turned off, and IPv4 is going to continue, whether we like it or not. Tony concluded by stating that an IPv4 market is necessary to make sure that we have a nice smooth transition to a longer term, but that continued growth will need IPv6.
  • An attendee asked the panel what would happen if a market was instituted in one region, but not in another, and what effect would a market have on the digital divide that exists today or in the future? Ben Edelman responded that depending on how the RIRs handle it, developments could take several different tracks, and that if a situation existed where the rules differed across regions, market speculation and other undesired impacts could develop. John Curran added that, as the economist noted, with a market we can end up with some strange side effects and unpredictability and that is why he personally is against the idea of a market. He added that a modified transfer policy would still allow space to be moved from where it isn't being utilized to where it would be based on need, whereas in a market where all IPv4 addresses are considered property and subject to market manipulation in their value, the outcome will likely be undesirable.

Implications of Global IPv4/v6 Routing Table Growth

Speaker: Jason Schiller

Introduction: PDF

Ray Plzak introduced Cathy Aronson who had arranged for both this and the following presentation to be given at the meeting. Cathy explained how both presentations were intended to provide information on the scaling issues in the routing system. Cathy introduced the first presenter, Jason Schiller, who gave a presentation on the implications of expected growth in the IPv4/IPv6 routing table. Jason began by acknowledging the efforts of those individuals who either contributed directly to his presentation or whose work he used in compiling his data. He then went on to provide information about provide independent space, aggregation and deaggregation, and IPv6 routing table growth, as well as how those issues inter-relate.

  • Marketing literature specs currently available hardware just behind the 7 year projection
  • Capacity is likely worse when considering IPv6 and firewall filters
  • Vendors are optimistic that they can keep up but lack firm implementation details
  • Growth is unbounded and could get worse
  • Vendors unable to adapt to sudden changes in routing growth
  • Need to investigate architectural solutions
Discussion Overview [Transcript]
  • An attendee asked to what extent do your numbers and your projections implicitly assume we continue using the same kind of technologies we are? Jason replied that his estimates were created using the growth as it is today, and the technology as its deployed today as the model. He added that certainly a different architecture is needed, whether it occurs in the hardware or the protocol itself.
  • An attendee inquired how many of the routes we've seen deaggregated for traffic engineering purposes could we recoup if the right BGP community lists were used. Jason, after asking the attendee to clarify what he meant by community lists, restated the question as why can't we use communities to leak more specific routes but limit the scope of where those more specific routes leaked to. In response, Jason stated that right now the diameter of the Internet is very shallow at something like 3.5 ASes, and in order to make this useful you have to leak it to your upstream, and you have to leak it at least one hop beyond that -- the end effect being that you end up leaking your route to the entire default free zone.

Routing/Addressing Problem: Solution Space

Speaker: John Scudder

Presentation: PDF

Cathy Aronson then introduced John Scudder, who gave a presentation on the possible solutions to the problem of table growth from a router vendor perspective.

  • Current architecture will be with us for a while, meaning upgrade cycles
  • Continued planning and management of routing table growth rate required
  • Locator/ID research is promising, but many open questions remain
  • There will be a Routing Research Group meeting at the Vancouver IETF
Discussion Overview [Transcript]
  • A few attendees asked for several clarifications on points made in the presentation regarding network mapping and the specifics involved, as well as a definition of "provider independent."
  • An attendee, noting that the presentation mentioned a risk if there is a sharp uptick in the growth rate of routing table entries, asked John what was meant by "sharp" and how it will develop given the exhaustion of the IPv4 free pool of addresses? John Scudder replied that currently the hardware can handle perhaps a million entries, and that in a few years, it could be up to 10 million, so it's more a matter of exceeding what can currently be handled rather than a specific number. He added that in response to the second part of the question, he felt the issue was addressed as well as it could be by the earlier discussion panel. He felt that we don't have any other choice in the near-term than to build bigger routers and deploy them, because of the time necessary to get them developed and in the field.
  • An attendee noted that while at the high-end, hardware may be able to handle the increased load, it wasn't clear if those capabilities were going to make it into other product lines, and problems may result out at the edges. Jason Schiller added that there is definitely an economic piece, as having to pay more for hardware that doesn't result in any direct economic benefit can be difficult, as opposed to hardware that will allow a company to carry more bandwidth from which they can derive revenue.
  • A remote participant submitted a question for Jason Schiller regarding his earlier presentation, asking if the routing table size had changed in the past year. Jason responded that there seemed to be a fairly stable growth curve with only minimal discrepancies.

2007-16: IPv4 Soft Landing

Speaker: David Conrad, Proposal Author

Introduction: PDF
Presentation: PDF

Ray Plzak presented an introduction to the proposal. Highlights included:

  • Advisory Council shepherds: Bill Darte and Paul Andersen
  • Introduced on PPML on 11 May 2007, designated as a formal policy on 28 August 2007. This will be its first discussion at an ARIN meeting. No revisions have been made.
  • Presentation of Staff Impact Analysis, Legal Review, Staff Comments, and overview of PPML discussion activity. [see presentation and transcript links above for details]

David Conrad, as the author of the proposal, continued with the presentation of the proposal. David presented the text of the proposal and then the rationale.

Discussion Overview [Transcript]
  • Mark Kosters, ARIN Chief Technology Officer, brought up the question raised by ARIN staff about whether the intent of this proposal was to include this type of authentication within other forms of communication, specifically ARIN elections. Mark asked Bill Woodcock, as one of the other authors of this proposal to respond. Bill replied that he was neither for or against that particular idea and that the proposal did not include authentication methods for ARIN elections in its intent.
  • An attendee spoke in opposition to the proposal, citing requirements of 100 percent utilization to be unrealistic. David Conrad replied that he originally had set it lower, but discovered from ARIN staff that current practice was to require 100 percent utilization for previously allocated blocks. He also noted that one of the suggestions he received was to peg the utilization as whatever was stated in the appropriate section of the NRPM. The attendee also noted other concerns if this policy were adopted; it would be more difficult for ISPs to obtain IP address space and if not adopted by other RIRs, ISPs in the ARIN region could be at a disadvantage David noted that there is difficulty in the timing of getting this passed in all regions, and the importance of not delaying implementation dependent on the actions of other regions. He concluded by saying that in regards to the issue of end-user versus ISP, he would be happy to do similar a proposal on the context of end-users. Later in the discussion, another attendee raised similar concerns, stating that he did not believe there should be the same benchmark for every size of allocation, and would prefer it allow the size of the block in each phase to have a different utilization percentage.
  • An attendee, speaking in favor of the proposal, asked if David could speak to the compatibility of this proposal and Proposal 2007-18, saying that he thought he didn't see a direct conflict and that they would work well together. David replied that the only issue he was aware of was the issue of thresholds, because of the reservation of 5 or 10 /8s for the RIRs, depending on the proposal. Another attendee spoke out in favor of the proposal and also identified the lack of conflict between this proposal and the others mentioned.
  • An attendee asked if any modeling of the impact of the proposal had been done, specifically to show whether this gains us 18 months or 6 years. Geoff Huston replied that while it would be difficult to model, it wasn't impossible, and that it would take about four weeks to assemble the data. The attendee asked a follow-up question about the use of percentages, as opposed to an HD-ratio. David replied that as an HD-ratio wasn't currently used in IPv4, he hadn't considered it as his goal was to change as little as possible of the current policy.
  • An attendee speaking against the proposal stated that in order to support it, he believed there needed to be a requirement on the requesting organization to document a plan to move to IPv6. He added that the utilization thresholds might be set too high for organizations with network architectures in place that couldn't be changed. David Conrad requested a show of hands on how the attendees felt about either including or not including utilization requirements in the proposal. John Curran managed the straw poll, and the author noted the results.
  • An attendee asked if having specific thresholds in the proposal wouldn't encourage a run on address space, as opposed to attempting to limit it to some percentage of the available address pool. David replied that he didn't believe this would be the case, and that he had chosen numeric thresholds in the amount of the free pool to make time the variable, as opposed to the amount of address space. He added that ARIN staff would continue to evaluate requests as they have been, and so this would protect the community from a run on the address space.
  • An attendee spoke in favor of the proposal, stating that it encouraged responsible behavior in the ARIN region, and that while he would like the proposal to be accepted as policy in all the RIR regions, he did not believe others regions not accepting the proposal would be an acceptable reason for it not to be carried out in the ARIN region.
  • An attendee stated he was in favor of the concept behind the proposal, and was not concerned by how much this extended the life of IPv4, but was concerned that it should allow for all networks to continue functioning during the transition to IPv6.
  • An attendee spoke against the proposal, but said he agreed with the spirit of it, stating that his concerns were the balance between extra work for the ARIN staff and the benefit received to the community, as well as the possibility of encouraging more organizations to submit false data to justify address space.
  • An attendee spoke against the proposal, stating that it overstepped the community's authority in its requirements for documentation of use of RFC 1918 address space and proof of IPv6 migration plans. David replied that if you need IPv4 address space when there is essentially no address space left, then you're going to need to prove that you're doing something that justifies the community allocating address space to you.

RIR Update - APNIC

Speaker: Srinivas Chendi

Presentation: PDF

Srinivas "Sunny" Chendi presented a report on the activities of the APNIC community and staff.

  • Consensus was reached on the policy proposal [prop-049] IANA policy for allocation of ASN blocks to RIRs, and was not reached on [prop-051] Global policy for the allocation of the remaining IPv4 address space. Two other policies have been put out for additional discussion, and four others have not yet been presented at an APNIC meeting.
  • Results from most recent survey were presented, as well as analysis of previous surveys including an examination of what items had been acted on.
  • Top three resource priorities were:
    • Technical research and development activities
    • Streamline resource requests and allocation process
    • Increase accessibility of APNIC meetings and policy processes
  • Next meeting will be APNIC 25, to be held in Taipei, Taiwan from 25 – 29 February 2008.

There were no questions or comments from the floor.

2007-18: Global Policy for the Allocation of the Remaining IPv4 Address Space

Speaker: Roque Gagliano, Proposal Co-author

Introduction: PDF
Presentation: PDF

2007-23: End Policy for IANA IPv4 allocations to RIRs

Speaker: Izumi Okutani, Proposal Co-author

Introduction: PDF
Presentation: PDF

Ray Plzak announced that, at the request of the authors and as these two proposals were so similar, they would be presented individually and then discussed jointly. Ray then began by presenting an introduction to the first proposal, 2007-18. Highlights included:

  • Advisory Council shepherds: Matt Pounsett Bill Darte
  • Introduced on PPML on 20 July 2007, designated as a formal policy on 28 August 2007. This will be its first discussion at an ARIN meeting. No revisions have been made.
  • Presentation of Staff Impact Analysis, Legal Review, Staff Comments, and overview of PPML discussion activity. [see presentation and transcript links above for details]

Ray then continued with an introduction to the second proposal, 2007-23. Highlights included:

  • Advisory Council shepherds: Dan Alexander and Rob Seastrom
  • Introduced on PPML on 17 August 2007, designated as a formal policy on 28 August 2007. This will be its first discussion at an ARIN meeting. No revisions have been made.
  • Presentation of Staff Impact Analysis, Legal Review, Staff Comments, and overview of PPML discussion activity. [see presentation and transcript links above for details]
Discussion Overview [Transcript]
  • An attendee spoke in favor of 2007-23, noting that regarding the concern for RIRs that use space more slowly being shut out, this would happen with or without this policy and this proposal provides for an even distribution. Upon questioning by the Chair, the attendee also noted that values for "n" of between 2 and 5 could possibly encourage more hoarding of the address space at the ISP/LIR level. Roque Gagliano added that this was also one of the reasons that it was important that this be a global policy.
  • An attendee voiced opposition to the idea of setting "n" as an integer, but spoke in favor of it being a fractional value based on the actual usage rate at the time the decision is made to do this. Roque replied that such a policy would made implementation much more complex and would open the door to more complexities, and in the end might delay anything from ever being implemented.
  • An attendee, speaking in favor of 2007-23, noted that while he supported a value of 1 for "n," he could perhaps go as high as two, but would object to any value higher than that for "n." Another attendee speaking in favor of 2007-23 cited similar reasons, but said that to go as high as a value of 2 for "n" would be dependent on the same policy being passed in all the other RIR regions.
  • An attendee stated that some data cleanup would also be required for the last set of /8s handed out, as they would include blocks that could not easily be used. Leo Vegoda noted that there were likely five or six /8s that could be problematic, and likely others where the problems aren't even known. Leo concluded by asking what could actually be done to "clean up" these blocks, as it would involve forcing a number of organizations to renumber. Later in the discussion, Izumi Okutani added that the same problem would exist, regardless of the method of distribution of these /8s. Another attendee also added that perhaps the solution was for these problematic /8s to be assigned to the region that caused them to be a problem.
  • Ray Plzak, speaking in clarification of the "gentleman's agreement" between the RIRs, explained that the agreement was simply that the RIRs agreed that regardless of the size of the allocation that they were entitled to, that they would receive no more than two /8s per request. He added that this doesn't preclude the RIR from coming back very shortly later and asking for more.
  • An attendee, stating that these proposals seemed to be reasonable, inquired of ARIN Counsel about the differences between reserving 5 /8s and 25 /8s, noting Counsels stated objection to 5 /8s being reserved for each RIR in the legal assessment. Steve Ryan, ARIN Counsel, responded that the reason 1 /8 per RIR was better was that there has to be a way to allocate IPv4 address blocks near the end which is not inconsistent with all of the prior methods, and that 1 /8 per RIR was not inconsistent while 5 /8s were.
  • An attendee asked if there was any significant difference between the two proposals other than the value of "n." Izumi responded that there was no other difference in implementation, but noted that the motivations for the proposals were different. Roque replied that his perception was slightly different in that 2007-23 included a provision requiring the RIR to post an official projection on IANA Exhaustion Date, which his policy, 2007-18, did not. He also noted that in 2007-18, the space was reserved for each RIR at the beginning of the process to ensure there are enough /8s at the end to distribute equally.
  • An attendee who was undecided asked the authors to discuss the points raised about fairness of the distribution. Izumi and Roque both replied that their motivation was not so much about fairness, but about increasing the predictability of events towards the end of the life of freely available IPv4 address space. The attendee responded that the reserved /8s would only give a small window of continuing IPv4 allocations in the more active RIRs and that he did not understand how a few months would enhance the predictability of events. Roque replied that knowing there would be a set amount of space at the end would allow each RIR region to develop its own policies on how that last bit of space would be allocated, and that would be up to each community to decide. An attendee pointed out that such a discussion had already occurred in the APNIC region.
  • Ricardo Patara, from LACNIC, made a short statement in favor of the proposals on behalf of the LACNIC Board and community.
  • An attendee spoke in favor of the proposals, but did not have an opinion on the value of "n," noting that any value decided on could have problems, but if we don't decide on one, we'll be in worse shape.
  • An attendee asked about the idea of each RIR receiving at least one /8 and possibly more at the same time as a final allocation, noting that the timing of such an allocation would have to be based on the run rate of the slowest user of space by moving back the clock to when their last /8 would be used up and that each RIR is going to need at least one /8 from that point forward, and some may need more than an entire /8 or perhaps a fraction of a /8. In clarification, John Curran stated that the trigger event with both proposals was based on a certain point of the available space, not on the run rates.
  • An attendee spoke in favor of the proposals, citing that regardless of the value of "n," there would be no real difference to the active RIRs, while it could be a major benefit to the growing RIRs that would give their communities years to make the necessary transition to IPv6.
End of Discussion:

John Curran took a straw poll to determine the sense of the room. After the tally was conducted, John stated that this information would be provided to the Advisory Council for use in its deliberations.

Closing Announcements and Meeting Adjournment

Speaker: Ray Plzak, ARIN President and CEO


Ray Plzak reminded attendees of the hours that the ARIN Help Desks and Cyber Café were open. He encouraged all attendees to complete the meeting survey available on ARIN's website. He reminded the audience about the ARIN social that evening, and that Day 2 of the ARIN Public Policy Meeting would begin at 9:00 AM. Ray concluded by extending thanks to the network connectivity and Cyber Café sponsor One Connect IP.