ARIN XIX Public Policy Meeting Day 2, Draft Transcript - 24 April 2007
"This transcript of the meeting may contain errors due to errors in transcription or in formatting it for posting. Therefore, the material is presented only to assist you, and is not an authoritative representation of discussion at the meeting."
MR. CURRAN: Good morning. If people come on in and settle down that would be good.
MR. PLZAK: Good morning.
SPEAKERS: Good morning.
MR. PLZAK: That's not too bad. I hope everyone had a good time last night. (Applause)
MR. PLZAK: Okay. Let's go through some announcements here. A reminder -- the help desks are open, and if you need to talk to either registration services or financial services, feel free to do so. We have Richard running the wheel of fortune again. So you could win a few of the things that are left. Actually we've been -- had two more winners yesterday afternoon. So if you haven't won and you haven't filled out a little survey slip, please do so. And so -- actually one entry per day, so you can actually -- if you won yesterday, you can try again today. So Andrew, I don't think there's another lava lamp, so you're stuck with just a one. Sorry, can't get a matching (off mike) so. Please remember to complete the surveys for both the v6 workshop. Jordi uses that information very much so to help his future endeavors, and also we would really appreciate your information in the ARIN XIX survey. And if you do so you get a raffle, one entry per person, a raffle ticket, and there'll be a winner of the video watch. So a discussion reminder: Follow the meeting courtesies and rules of discussion, the reference of program in your meeting folder -- do go through them. I want to thank everyone for their adherence to that yesterday. It helped the session move along, and we accomplished a great deal yesterday. And we have another full schedule today, and I'm sure the chair will appreciate you following those rules. So, a thanks to all of our sponsors. So thank everybody including yourself. (Applause)
MR. PLZAK: And I believe a -- one of our special participants -- Dorothy, are you here some place? Ah, do you want to come up and say a few words?
MS. BOLLMAN: Thank you. Good morning. Dr. Moreno cannot be here this morning because he has another meeting at the very same time, but he wanted me to send his regards and to welcome all of you to Puerto Rico. He also wanted me to announce that the Internet exchange of Puerto Rico is going to be up and running by May 13th. The exchange is sponsored by PCH as well as Gauss Research Laboratory, Inc., and the founding members are CaribeNet, PrepaNet, Ultracom, and the University of Puerto Rico, including the Rio Peidras Campus, and also the high performance computing facility. And if you'd like further information about the exchange, it's very easy to remember the website, it's ix.pr. Dr. Moreno, as I said, has another meeting and he might be here later on this morning. But again, thank you very much, and welcome to Puerto Rico. (Applause)
MR. PLZAK: Thank you, Dorothy. Okay. Next on the agenda is the NRO Activities Report. I do not have any PowerPoint slides for you on this. So this will be a brief update. We are continuing to work with ICANN on a definition of the relationship between the NRO and ICANN, and we have made some progress. And we are currently trying to work out towards the -- for an exchange of letters between us. Also, we'd like to report that we have been having some very cordial meetings with the general manager of the IANA, the latest one being this -- a breakfast meeting this morning. And so we feel we have a very, very well established -- cordial communications link established that way. Also recently at the last ICANN meeting which was in Lisbon several weeks ago, we made several presentations, one of which was the Government Advisory Committee and the other one was to the open public forum regarding the depletion, the consumption of IPv4 address space. Also that is going to be part of the panel discussion this afternoon that will be panel consisting of the Board. We are -- we have been working with the IANA to help modernize, if you will, the registry files that they have before V6. IANA is undertaking some formatting changes and so forth. I'm sure Leo will tell you about that shortly. And so we have been working with him on that, and that work is proceeding as well. So other than that, the NRO continues to be the coordinating body. And so the joint administration of the DNS, that of the address space that was handled through the early exchange -- early registration exchange process that we went through a few years ago, operate seamlessly. And as was reported by Paul Wilson, although it's not a NRO activity as such, the RIRs are working together jointly to work on the requirements if you will, for a certification of number resources. And so that really concludes my update as far as the NRO activities. Glad to entertain any questions you may have.
MR. PLZAK: And if there are none, I would like then to ask Leo to come up and present the IANA report, assuming that Leo is going to do it. Okay. David did delegate that to you or give you a reason to come to Puerto Rico. Okay. So Leo used to -- as many of you know, used to work at the RIPE NCC. He has now drunk the Kool-Aid and he is with David. So -- Leo.
MR. VEGODA: Good morning, everyone. I am Leo Vegoda, and I'm the IANA Numbers Liaison. How do I get to the next slide?
SPEAKER: (off mike).
MR. VEGODA: Oh, this one?
MR. VEGODA: Oh, okay, lovely. So -- yeah, that's my e-mail address if you have any questions and you want to follow up with e-mail. If you want to interrupt and ask a question midway through, then just come up to a microphone. So an overview of the presentation, a little bit about IANA, some work we're doing on improving the services that we offer, some statistics, and an opportunity for some questions. Oops. Okay, so what do we do apart from destroying the equipment? We manage about a thousand registries. I went through the assignments directory and looked at all the registers. It was about a thousand and I stopped counting. We manage them for the RIRs, the IETF, TLDs, and just members of the general public. One example of that is the private enterprise numbers registry. That's also sometimes called OIDs. It's something where anyone can go and get a pen. And we've got a wonderful snazzy new interface, a new good-looking website that has some extra automation in it to make it easier to use. And it sits on top of an improved backend so that it's generally easier to keep the information for your private enterprise number up-to-date. The website does look better, and we're trying to make our main website look better as well, and will have more on that in a few months. As Ray mentioned, I've been working with the RIRs on improving registry data, in particular the IPv4 registry. We've been going through some of the /8s in the IPv4 registry on a case-by-case basis. Basically people contact us and say, "Hey, can you please update my information?" We work with the RIRs. I've done quite a lot of work with Leslie. And you've probably seen if you look at the IPv4 registry that we must have made about four or five updates to it over the last few months. We're also looking on doing a proper forklift upgrade to the IPv4 registry. Basically really making it much, much more useful than it is today; working on that with the RIRs, hopefully we'll have something fairly soon. I expect we'll be consulting on any major changes at the registry. But yeah, we should have something exciting fairly soon -- well, as exciting as listing numbers on a screen gets. You'll also notice the public data network numbers registry there. That's actually 14 /8; 14 /8 is a class A or a former class A that was reserved for mapping X121 addresses to IP addresses so you could connect your X25 network to the Internet. The picture is an X25 DTE I believe. I must admit I never use one myself, but there were about a thousand addresses in this /8s that were registered in our registry. And I've been going through the registry contacting people saying, "Can I please update your contact information?" And on the whole, they've been saying, "Actually we don't run X25 anymore." (Laughter)
MR. VEGODA: Not everyone has said, "We don't run X25 anymore." A couple of people who've said, "Let me check, I'm not sure if we still run X25 anymore." But got a thousand addresses down to just over a hundred. It's possible that at some point we could be updating RFC 3330 to go and chuck this one back into the pool available for allocation to an RIR. We're not quite there yet, but it might happen. And of course, that would probably be another two or three weeks of IP before address space if we were manage -- able to do that. We're not just updating the look and feel and the quality of the data. We're actually trying to update our backend systems as I mentioned before. We're XML- ifying our data so that we can output it into formats like text or HTML or raw XML. And maybe programmatically we need -- you know, we need to be able to meet peoples' needs more easily. And so that's a good way of doing it. It will also help us with the improvements to our main website. Hopefully it will be easier to find the registry that you want once we've done all of this. We've been working with Mark Blanche from VHNE on this. And well, hopefully we'll have some significant improvements on that fairly soon. Also DNSSEC -- we are looking at signing ARPA and the delegations within ARPA. And obviously, if you want to do DNSSEC on the reverse tree, that makes things easier. That's something we're in discussion with the IAB on as to the requirements and how to go about doing this. We've got Richard Lamb working with us to help us with our systems so that we can go and manage the process properly. But -- yeah, that's something that should be coming along fairly soon. We've got other things that we've got -- we're thinking about, we do need your input as well. If you like what we're doing, then please tell us. If you want more of something, please tell us. If you want to do things in a different way, please tell us. But if we just sort of bumble along without your input, then that doesn't work very well. So we do need you to contact us either by e-mail or here at the meeting, and let us know what you think. So here are the statistics I promised. So we've got a growing demand for IPv4. The chunk on the right is only the first quarter of 2007, so presumably that could be a little bit bigger by the end of 2007. One thing that was worth noting is the RIRs have said that they're only going to request up to two /8s at a go, which should mean there's a little bit -- it's going to be easier to manage the -- sorry, to measure the demand. So you'll go and see chunks of two /8s going out there. And it won't be sort of nothing, nothing, nothing -- there's five. It will make it little bit easier for anyone who is counting down to do that countdown. This is a picture of the Aral Sea. The Aral Sea was a big pool of water and it got smaller. And the IPv4 was a big pool that is sort of like also leaking away. This here is my artist's impression of IPv4 diminishing. It's the free pool with below 50 /8s. And though this is only going down, it's worth noting that in the work that we've done with Leslie, we've actually had someone who contacted us and said, "Please update my contact information," and they chucked one back. So that's quite nice, you know, it's obviously potentially worth a lot of money. So it was nice of them to go and say, "Well, here's one that we don't need anymore and you can have it back." So it's not only going down. There are little small increases, but I've not done in such incredibly small detail. And that's basically it. So if anyone's got any questions --
MR. CURRAN: Microphones are open.
MR. VEGODA: Yes.
MR. DURAND: Alain Durand from Comcast. When you are doing an allocation to the registries you go and update the IPv4 table that is available on your website. How quickly is this updated?
MR. VEGODA: Well, if I -- once I've updated the table in our system, it's published within an hour. Is that what you're talking about?
MR. DURAND: Well, do you update this file immediately after you do an allocation to the registry, or does it take some time, weeks or months before it gets updated?
MR. VEGODA: Do you mean the amounts of time it takes for us to evaluate the RIR's request?
MR. DURAND: No. After you make the allocation to the RIR, do you go and update the IPv4 tables immediately? I mean, can I just look at this and have a day-to-day picture of how many /8s are left?
MR. VEGODA: I see what you mean. Part of the process of making an allocation to the RIR is updating the IPv4 registry. So about the time that we go and send the e-mail to the RIR saying, "Here is your new /8s, we've updated the IPv4 registry." So there might be a 20-minute period where the registry hasn't been updated. But really we go and renew it on the website every hour anyway. So you shouldn't have any problems with it being very out-of-date. Okay, looks like that's it. Thank you. (Applause)
MR. PLZAK: Thank you, Leo. We are now on to the first policy proposal discussion of the morning. And we are about 20, 30 minutes ahead of schedule or something like that; anyway, we're far ahead of schedule. So the first proposal we discuss this morning is policy proposal -- 2007-4: Changes to IPv6 policy - removal of "interim" consideration. The policy was introduced in January this year and designated a formal proposal in February. This is the first public policy meeting in which it is going to be discussed. The text has not been revised. The policy seeks to remove the sentence that refers to IPv6 policy as "interim" policy. The shepherds are Dan Alexander and Bill Darte. Counsel sees no -- as of April this year sees no liability risk. Staff has no comments, notes that this would delete Section 6.1.1 of the NRPM. From an implementation assessment the resource impact will be minimum. It could be done within 90 days after Board of Trustees ratification, and would require a change to guidelines and some staff training. Not a lot of discussion on this topic; 10 posts by 6 people; 2 for, 1 against in terms of threads. Comments -- "We need a lot more cleanup proposals like this." "At this point in IPv6' adoption almost ANY policy is counterproductive to its deployment." The text is in your meeting program and also is available at this website. And so if it's IPv6 it must be Jordi. So Jordi, would you please come up?
MR. PALET MARTINEZ: Okay. This should be very short. Basically it has been set already. We have this text in the introduction/overview of Section 6.1.1 of the resource policy manual. This policy is considered to be interim -- an interim policy. It could be reviewed in the future subject to greater experience in the administration of IPv6. So the actual text is suggesting that it is an interim policy, and I believe is no longer the case. And in any circumstances it is clear that any policy is always subjected to changes. However, the rest of the policies don't have such texts. It may be even sometimes considered as a negative message for some people towards IPv6 employment especially because the reading of that text like "there is not enough experience," -- some people -- that it's not really following the situation. And the deployment work done which IPv6 can think okay, then it's better for me to wait a little bit until it is much clear, right? I think this is not the real situation, but clearly for some people can be very negative. And that's it.
MR. CURRAN: Microphones are open. Name, affiliation, whether you're for or against the proposal, and your comments.
MR. LEWIS: Ed Lewis, NeuStar. Somewhat undecided, probably for it. But I do want to ask the question, does anybody know why that text was there to begin with? Is there anybody who -- in the proposal when this got written into the proposal?
MR. PALET MARTINEZ: David is -- reply.
MR. KESSENS: I'm David Kessens, Nokia Siemens Networks. As one of the original people who was involved with writing this text, we put it in because we actually realized this policy was highly imperfect at the time we wrote it. And we wanted to put this into -- put it as encouragement to revise this policy and make it more final policy that actually is probably more balanced after we gain the experience in the field of IPv6. I have to ask, like, do you think it should be removed? You know, well, as one of the authors I think actually we shouldn't waste any time in this meeting considering and discussing and wasting a lot of time on one little word in the policy. I think we can spend our time better. And I really don't --
MR. CURRAN: So are you saying you're against adoption of this?
MR. KESSENS: No, I don't care.
MR. CURRAN: You don't care? Thank you. (Laughter)
MR. PALET MARTINEZ: One question I think is it is clear that the reason why the interim thing was there that I think it's clear that the policy has been already reviewed other times as other policies have been in the process of being reviewed so --
MR. CURRAN: Far microphone.
MR. ITO: Kosuke Ito from Japan. I share the same feeling with David Kessens as an original editorial member of the -- this policy.
MR. CURRAN: Okay. Thank you. Any other comments? Closing the microphones. The microphones are closed. Okay. Thank you, Jordi.
MR. PALET MARTINEZ: Thank you. (Applause)
MR. CURRAN: From time to time we need a show of hands in order to guide the AC in their deliberations. At this time I will be calling for such show of hands on this policy proposal. I'm going to ask for everyone in favor and then everyone opposed. Everyone in favor of advancing policy proposal 2007-4, please raise your hand now. Thank you. Everyone opposed to advancing policy proposal 2007-4, raise your hand now if you're opposed. Thank you. We have a count. Number of people in the room -- this excludes ARIN staff -- 96. All those in favor of advancing 2007-4 -- 44; all those opposed -- 2. Thank you.
MR. PLZAK: Okay. Now we're making some extremely lightning time here. Policy proposal 2007-9 -- modernization of the ISP immediate need policy. Here it is in February of this year and designated a formal proposal in March of this year. This is the first meeting in which it will be discussed. Text has not been revised. This changes the immediate need policy by reducing the minimum allocation from a fixed /20 to ARIN's current minimum allocation as defined elsewhere in the NRPM, and sets the maximum allocation size at a /16. It also defines a 30-day utilization window of the IP address space. The shepherds are Rob Seastrom and Stacy Taylor. Counsel sees no liability risk. Did make a comment which is actually on the merit of the proposal: "We also believe that this policy is beneficial because it prevents the waste of resources." Staff has no comments other than to say that this modifies Section 188.8.131.52 of the NRPM. Minimum resource impact -- excuse me -- moderate; 90 days after Board of Trustees ratifications and guidelines changes and staff training. And so 11 posts on this by 6 people; 3 for, 2 against. Comments -- "the possibility that the allocation would need to exceed a /16 would be exceptional as well, but since it may exist, the limitation must be removed." And said, "How reasonable can their IP address planning actually be, if they need the IP address component of that to turn around on an extremely short schedule? Answer: it can't." Text is in the meeting folder, and also available at this URL. And RS, floor is yours.
MR. SEASTROM: Thank you. So my hope that this would go more smoothly than yesterday was not entirely in vain. So this is policy proposal 2007-9 which has to do with the immediate need policy. One of the comments received was about how effective can peoples' IP address allocation policy be if they need stuff turned around on an extremely short timeline, and the answer is they can't. And that sort of underlines one of the problems with the immediate need policy, which is that most people don't know what it is. And probably because most people have been fortunate enough to never have to use it. The name is misleading. It doesn't actually get you IP addresses faster than you would ordinarily. And it is actually a lot more work than getting IP addresses under the normal policy. On Sunday, Ray cited RFC 790, which was the first articulation of IP address policy, which said if you need an address go and talk to John. And moving forward from that, the policy, has always sort of, had at its core, that if you need it you can get it. You just have to show that you need it. Over time, the main way that people have shown that they need IP address is by their previous usage patterns. You go and you talk to the analysts, and you get the amount of address space that would be reasonably expected based on a normal growth curve, and the amount of addresses that you have used in the past. Unfortunately, that doesn't always work. There are recorded cases where organizations need address space, and they don't have a previous history with ARIN, and yet by doing a huge amount of documentation, they can show that they actually need this address space, and that they are not just hording or something like that. So I have done three of these; once on my own, and twice as a consultant. Here are some typical situations where some -- where an organization might use the immediate need process to get their -- to get IP address space. In my particular case, I was working for an ISP that had outsourced all of its infrastructure, we decided after doing cost analysis that we now had enough customers that we could in-source that space or that functionality and do it economically. So we have this huge step function. We went from zero to a minimum allocation. So my organization had no previous history with ARIN; we got the address space we needed. Similarly, if an established organization is launching new product or service, and it is big enough for a joint venture, they create a new organization, a new LLC, a new partnership, that has no previous IP address history use as well that creates a similar sort of problem. And perfect example of that is MSO kind of, situation where you are selling on the cable plant overlay, and you are really sort of bound to the architecture of the CMTSs and such. And it is very easy to find yourself in a situation where you can't put more than /29 or /30 on each box, which serves hundreds to thousands of customers. And obviously your partner organization is not going to be thrilled with rapid churn of swapping round addresses. So bottom line on immediate need request is that it is an enormous pain to do it, and that's why it is done so rarely. Pretty much, if you can justify the address space by any other means in the NRPM, you'll do it, because documentation stack for mine when printed out exceeded the centimeter thick. So the current text of 184.108.40.206, which is the Immediate Need Policy, says that if an ISP has immediate need for address space, and the need exists on the day of the request, I can't rule stuff out on the same day, so it is kind of, that creates a constraint that is not really reasonable. ARIN may issue a /20. One size fits all. If the organization, such as the new company, shows justification; however, these cases are an exception. So between this requirement for the space on the same day, and the /20, this seemed unreasonably constraining, particularly, with some of the partnerships in the new organizations that pop up from time to time. So the policy statement is modify NRPM 220.127.116.11 to read, "If an ISPS has an immediate need for address space, and can provide justification to show these address spaces will be utilized within 30 days of the request, ARIN may issue a block of address space not larger than /16 nor smaller than ARIN's customary minimum allocation to that organization." These cases are exceptional. So the rationale, yes, the current existing policy constrains us too much. It is a one-size fits all policy, which simply doesn't work for MSO wholesale in a large metropolitan area. And it forced ARIN to hand out chunks of address space actually larger than the ones required, because it constrained ARIN to only handing out /20. And if the amount of address space actually justified was smaller, it makes sense to hand out the amount of address space that's actually justified rather than a bigger chunk. Any questions? Any discussions?
MR. CURRAN: Microphones are open. Far right, yes.
MR. LOEVNER: Michael Loevner, of Verizon Internet Services. I just wanted to comment --
MR. CURRAN: Excuse me. Are you for or against the proposal?
MR. LOEVNER: I am against the proposal as written.
MR. CURRAN: Okay. Thank you.
MR. LOEVNER: I think it is an improvement to the policy, but I don't believe that there should be a limitation. If you are able to show that you are going to need more than a /16 within 30 days, you should be able to obtain more than that.
MR. CURRAN: Do you wish to respond?
MR. SEASTROM: My take on that, having discussed that extensively with several people was that throwing the gates wide open, probably would not gain a huge amount of traction. Because somebody would say, "Well, someone is going to come along and request the /6," and they will have the justification or a business plan harebrained or maybe they will say that, "Yes, actually they need a /6 worth of space in the next 'n' number of days. So, the question is how far to the left do we want to slide that net mask, and if we continue having difficulties with not enough space being available under immediate need policy, and Leslie can speak to about how often the immediate need policy actually gets invoked, which I don't think is actually that often. I would prefer to slide it gently to the left with successive policy proposals in response to actual demonstrated problems rather than just say that there is no limit. This -- I was hoping to provide additional confidence that there wouldn't be waste of address space by putting the /16 in there. It's -- it was kind of arbitrary where to put up, but it seemed like a reasonable compromised position.
MR. LOEVNER: Okay. I can understand that, but the -- I have had personal experience with the Immediate Need Policy and there are exceptional cases where more than a /16 is going to be necessary. And, I think, the Immediate Need Policy hasn't been enforced as written in the past by ARIN staff.
MR. CURRAN: So -- and the tradeoff of the other changes to the proposal. Overall, you're against it, because of the limitation of /16.
MR. LOEVNER: Well, right now, I don't think that because the proposal is -- because the current policy is poorly written, that it's not totally adhered to by the ARIN staff, and that if it were to be rewritten, and it seems to be a well written policy now, I think, that there should still be some leeway for ARIN if the immediate need exists to go over that /16 boundary.
MR. CURRAN: Understood, but as the policy proposal is written, the question is are we better off with it as NRPM is today or changed in this matter. You're saying, overall, the change is that --
MR. LOEVNER: The change is an improvement, yes.
MR. CURRAN: Okay, thank you. Okay. Center front mic, yes.
MS. NOBILE: Hello? Okay, I'm sorry. Yeah, I just wanted to give you some starts on that. We went back to 2004, to see how many times immediate need has been used, and we have approved 22 requests in the past three years and three months, whatever.
MR. SEASTROM: So that's one every month-and-a- half.
MS. NOBILE: About that. And out of the 22 requests, 3 have been larger than a /16.
MR. CURRAN: Okay. Good information, thank you. Microphone, center back.
MR. BICKNELL: Leo Bicknell, Harrah's Entertainment. I support the proposal. I also support removing the /16 notion. I, actually, think perhaps, and maybe Leslie or someone can comment on this, that what it should say is a /8, because if ARIN doesn't have space, they can give it out in 30 days, and that's what they get from IANA. So, I don't think they could give out a /5 in 30 days. So if there is going to be a limit, I think, that should be a bit -- actually supporting that limit.
MR. CURRAN: Leslie, would you like to comment on that?
SPEAKER: (Off mike) /8 IANA.
MS. NOBILE: This is not on. Okay. Yeah, we wouldn't be able to provide a /8 either -- we never have one available.
MS. NOBILE: Always tapped into. And --
SPEAKER: (Off mike)
MS. NOBILE: Yeah, we don't. (Laughter)
MR. CURRAN: Okay.
SPEAKER: Leo will show us that the process is quicker now. (Laughter)
MR. CURRAN: So, as written you support the policy proposal?
MR. BICKNELL: Absolutely, and I would support in removing the limit.
MR. CURRAN: Okay. Other speakers at the mic over here.
MR. RICH: Yurie Rich, Command Information. I neither support nor disagree with the policy, I don't have enough information. My question is to the staff. Is the level of care given to assessing the requests any less diligent than you would for a normal request?
MR. SEASTROM: Having submitted three of these requests myself, I can state that the level of care seems to be a whole heck of a lot higher.
MR. CURRAN: Okay. Leslie?
MS. NOBILE: (off mike) We give it our all, no matter what you are asking for. The level of effort required on the part of the requester is greater, therefore the level of assessment is greater on the staff. There is a lot more documentation required.
MR. SEASTROM: Okay. I'll go with that assessment.
MR. CURRAN: Okay. Front center microphone.
MS. SCHILLER: Me, okay. So I'm --
SPEAKER: Your name?
MS. SCHILLER: Oh, sorry. Heather Schiller, of Verizon Business. And I'm not sure if I am for it or against it, because I still have questions. So, my question is, organizations that are applying under immediate need don't fit under the use of a /20 already. Is that -- it is not really that they need the space right away, it is that they don't already have a /20 with their space?
MR. SEASTROM: A better name for it might be that we have no history and yet we need IP address space and we can justify it through extraordinary means and documentation policy. (Laughter)
MR. SEASTROM: It won't fit without word wrapping, so --
MS. SCHILLER: Okay. I guess what I'm trying to say is, it seems like there is a gap. There is three types of folks that will be coming for space. There is the folks who already have a /20's worth of space and qualify under the standard policy. And it also -- under the use of a /20, it says, "If the organization has less, then they don't meet the minimum requirement of having a /20," and then -- but the immediate need policy, kind of says, "But if you can demonstrate that you need it, then you get it." It's sort of like --
MR. SEASTROM: I can see a situation where an organization that already had space from ARIN would be applying under the Immediate Need Policy. Depending on, precisely, what they are planning to do with it, multiple speed networks might be a better plan --
MS. SCHILLER: Right.
MR. SEASTROM: -- if they can justify it that way. Or the normal ramp-up process, and just coming back a week later and saying, "Hey, we're out," then getting more might be a reasonable process too.
MS. SCHILLER: I guess what I'm saying is if you take the immediate need part and the current policy for use of /20 like, it pretty much says, if you are an ISP and you can show that you'll use it, or that you have it. You'll get it. Am I wrong or --
MR. SEASTROM: I'm not sure I understand the question.
MS. SCHILLER: So if you take the immediate need part, which says if you have no prior history, but you can show that you will use this /20 in the next whatever time period, you know, immediately. And/or you can qualify under that or you can qualify because you already hold a /20. Like together you should --
MR. SEASTROM: Well, if you have a /20, and you need a /18, you may be in trouble.
MS. SCHILLER: Okay.
MR. SEASTROM: If you are trying to do it based on previous use, it might have -- how long you've had this /20, how long it took you to run out of it --
MS. SCHILLER: So it's really for when you are requesting something that's not double of what you already have. Or something of larger --
MR. SEASTROM: Yeah, and any -- if there is a step function in your growth curve, the Immediate Need Policy comes into play.
MS. SCHILLER: Oh, yes, thanks.
MR. CURRAN: Leslie, you want to --
MR. SEASTROM: Is that a viable assessment, Leslie?
MR. CURRAN: Yeah, you want to add on to that?
MS. NOBILE: So, a lot of ISPs are companies who have space will come to ARIN under the immediate need. But what they are coming to us for is, they have got a -- they have a new product or new service, they are opening a completely new organization ID. Therefore the new org ID has nothing; the new product, the service, whatever, has no space; so they have to justify this from scratch. We can't look at their existing history with their other accounts. So that's generally what we will see now.
MR. SEASTROM: Yeah, Joe's ISP Services does a joint venture with American Tower Company that does radio towers to do a WISP; new LLC, new org ID, no assets.
MR. CURRAN: So given the clarification, Heather, are you in favor or opposed to the proposal?
MS. SCHILLER: I'm for it. I'm in favor of it. I don't think it --
MR. CURRAN: You're in favor of it. Okay. Thank you. Center rear microphone. I'm going to be closing the microphone shortly. If you are in favor or opposed, please find a microphone.
MR. ALEXANDER: Dan Alexander, ARIN AC. I support this policy. It seems to fill the need that we need to address. My only suggestion would be that I kind of, noticed a similar point yesterday with 2006-7. We seem to be writing policy to take care of the inadequacies of another policy rather than revising the one that we need. You know, while I think this should move forward, it seems like this is really trying to address a gap in the initial allocation policy where the Immediate Need Policy isn't even needed to begin with. So it's just food for thought that we at the AC needed a --
MR. SEASTROM: Well, the reason that the Immediate Need Policy exists is a separate thing to the initial allocation policy. It's the initial allocation policy that makes an assumption that you're ramping up your initial allocation from stuff that has been slipped to you.
MR. CURRAN: Okay, Bill?
MR. WOODCOCK: I would just like to second what Dan said that the best policy is always policy that results in a net reduction on the length of the total policy document.
MR. CURRAN: So, I also will say that that's a statement that is obviously correct, because a clearer policy is easy to administer and implement, easier for people to change. The answer to that is policy proposals to simplify are more than welcome. You can submit one at any time. I look forward to seeing it. Okay, microphones remain open. Microphones are open. Any further discussion? Seeing no further discussion, microphones are closed. Thank you, RS. (Applause)
SPEAKER: So John may I ask you four questions. (Discussion off the record)
MR. CURRAN: One moment please. (Pause)
MR. CURRAN: Okay. All right. It comes time to -- from time to time seek a show of hands in order to provide guidance to the AC in their deliberations. At this time, we are going to seek a single show of hands for the purposes of whether or not their support to advance Policy Proposal 2007-9. So I'm going to ask for everyone in favor and then everyone opposed. All those in favor of advancing Policy Proposal 2007-9, please raise your hand. Thank you. All those opposed to advancing Policy Proposal 2007-9, please raise your hand. If you are opposed to this policy proposal, please raise your hand. Thank you. Do we have a count. Okay. Number of people in the room excluding ARIN's staff, 96; all those in favor of advancing 2007-9, 38; all those opposed, zero. Thank you.
MR. PLZAK: Okay. We are going to move on, so just pretend you had a break. You'll have another one coming to you in a little while. It doesn't take much to make you feel better alright. (Laughter)
MR. PLZAK: So anyway, we are now at proposal 2007-10. End Site Immediate Need Policy. This was introduced in February this year, and designated a formal proposal in March. This meeting is the first meeting it's going to be discussed at, and text has not been revised. This says that if 2007-9 is adopted, then 2007- 10, establishes the Immediate Need Policy for end users. The minimum assignment would be the ARIN's minimum assignment as to find elsewhere in the NRPM, and the maximum would be a /16. Also defines a 30-day utilization window of the IP address space. If 2007-9 is not adopted, then 2007-10 require replicating the text from the existing ISP and Immediate Need Policy into the end user section of the NRPM. Shepherds are RS and Stacy. Counsel sees no risk, also it says, there is a comment. We also believe this policy is beneficial, because it treats categories of applicants equally. And that's April of this year. Staff comments -- Text should 'have allocation' changed to 'assignment' because this refers to an end-user assignment, and notes that this will require new section in the NRPM 4.3.6. Assessment is that minimum, 90 days after the Board of Trustees ratification, requirements for implementation would be guidelines changes and staff training. Eight posts by five people; two for, two against. Comments are: "I entirely support." "I do not support this particular modification of the policy, because it limits the amount of the address space that can be obtained under this policy to a /16." "Shouldn't we be doing things to deliberately discourage folks from acquiring large quantities of IPv4 space?" Proposal text is in your meeting handout and is also available at this URL. And I will now turn this over to RS.
MR. SEASTROM: Well, again, so 2007-10, basically is a proposal to replicate whatever ends whatever Immediate Need Policy we institute for ISPs; exactly, for end sites. So the statement is to create a new section in the NRPM, which mirrors the intent of 18.104.22.168. The problem is, we have 2007-9 in place, so this gets a little bit confusing. But basically, when 2007-9 is finished, this proposal is to copy whatever is in 22.214.171.124 after 2007-9 is dealt with to 4.3.6. So first possible outcome, we ratify 2007-9. We have seen all this verbiage before. It is exactly the same as the previously discussed change to -- in 2007-9, with the exception it says of an end user. If 2007-9 is not ratified, it copies over what's already in the Immediate Need Policy, except the changes, the verbiage to read end user. So, how do we get here? Well, the problem is that the current policy is self-contradictory. It says both, you can apply for space as an end site, and that only ISPs can apply for space as an end site under the Immediate Need Policy. So, we need to make this consistent, and in the event that the consensus of the community is that we do not want it to be consistent, then you can expect to find me at the next meeting with another proposal to get rid of the inconsistency by striking text from elsewhere in the NRPM. Questions and discussion?
MR. CURRAN: A point of question to Leslie wherever she might -- okay. Given that the current policy proposal states -- I'm sorry, the current NRPM states that it applies and doesn't apply to end users, it would be fair for ARIN staff to proceed in either manner. I'm curious what the practice is with respect to the current policy and end users.
MS. NOBILE: Okay. We've -- because it is conflicting, and we would like it to be fair to everyone, we apply it to end users as well as ISPs.
MR. CURRAN: Okay. Thank you. Okay. Thank you. Go ahead, Owen.
MR. DeLONG: I think, given the multi-home end user policy that it is probably rare this would come into play for end users. Leslie is nodding in agreement, so I'm assuming that means that I am right. And she is the one that would know for sure. So I would actually oppose this -- Oh, Owen DeLong, Jupiter Networks -- I would actually oppose this policy and suggest that we should instead seek to, believe it or not, remove end users from the immediate need criteria. Because I think that absent multi-homing, it doesn't really make sense to be giving end users PI space, and that with multi-homing the end user criteria are actually much easier to qualify for under the multihoming policy than under this one.
MR. CURRAN: Understood. It is true that this policy proposal clarifies things to have the end users included in the policy. And clearly if this is defeated, we still have an ambiguity that someone needs to consider whether or not we strike them from with a separate policy proposal. Far microphone.
MR. LOEVNER: Michael Loevner, Verizon Internet Services. I feel that -- I oppose the policy. And I feel the same way as Owen also. But I think one of the other things is that when I read the end user existing policy it seems to already cover any immediate need that an end user would have if they -- there is not really a limitation on what they can get as their initial address space. And if there were a new organization created, I don't see the limitation for the end user either, so I think that they are pretty much covered currently and there is not an ambiguity from how I read it.
MR. CURRAN: Okay. Center mic, back mic.
MR. BICKNELL: Leo Bicknell, Harrah's Entertainment. I had not considered the problem that the previous two speakers have mentioned. And so, I want to -- kind of put Leslie on the spot again, if she could describe one of the situations that has come up since she said they have applied the policy the same way. I assume it has actually happened, and I am kind of, curious who would apply this to.
MR. CURRAN: Okay to Leslie. A case where this policy was applied with respect to end users; can you characterize such a condition?
MS. NOBILE: You know actually, I was just conferring with my staff on that, and we actually can't bring to mind a situation where any particular end user -- it is so rare, I just -- we just can't even remember.
MR. CURRAN: Okay. We don't have an example off hand.
MR. SEASTROM: I'm surprised to hear that question coming from Leo, because I would think that your corporate masters would be on the list of organizations that might need more than the maximum amount of addresses. They are typically offered for multi-homing purposes as far as globally unique addresses come.
MR. BICKNELL: Unlikely.
MR. CURRAN: Right. Microphones are open. Microphones remain open, as a point -- yes, all right. Yes.
MR. BRADNER: Just a question -- was the wording change that staff asked for, made?
MR. SEASTROM: Yes.
MR. CURRAN: As just a little point of order item. So, the parliamentary procedures we use to run this organization suggest that the comments be made on the policy proposals and addressed to the meeting or the Chair, which is a convenient proxy for the meeting. It's good not to directly address other speakers or reference other speakers if at all possible. It's just a polite mechanism. So I ask that when you are talking about the debates of the -- the merits of the proposal, you limit yourself to the merits of the proposal and don't comment on other particular people or motivations. Microphones remain open. Yes.
MS. SCHILLER: Heather Schiller, of Verizon Business. I was just looking at the NRPM and I'm not sure that there is a need for this policy, because in the ISP policy, it does ask you to show prior use of your -- or use of the address space that you currently have in order to justify your new request. And this just says if you need a slash -- like, there is no requirement to show previous use of address space. So how, I mean, and it also says like show 25 percent use in, you know, three months, so if you have a more immediate use when you already qualify?
MR. CURRAN: RS do you want to respond on that? You are saying there is not a need because this is already covered?
MS. SCHILLER: It looks that way.
MR. SEASTROM: I am prepared to be found wrong on this, but I would ask that you check out the amount of address space that you can get under multi-homing. I believe it is severely limited.
MS. SCHILLER: It says a minimum of a /22, and if it is a /20 or larger, it comes out of a specially reserved range.
MR. SEASTROM: Okay, so you can in fact get larger than /22.
MS. SCHILLER: It looks that way.
MR. SEASTROM: Okay.
MR. CURRAN: So it's at -- you're suggesting that the policy proposal isn't needed, because your reading is such that there is already that flexibility.
MS. SCHILLER: It looks that way.
MR. CURRAN: Okay. As such you are against this policy proposal?
MS. SCHILLER: Well, it didn't -- what you're saying don't make it bigger. Don't make it longer.
MR. CURRAN: Are you for, or against it, or you do not know?
MS. SCHILLER: I just don't see a need for it. And I would leave that to ARIN staff to kind of, verify for us whether or not it is actually needed or -- it sounds like if the current policy is sufficient, then just leave it be.
MR. CURRAN: Okay. Leslie?
MS. NOBILE: All right. We just went back and checked some files, some old files and it looks like 21 end user organizations have immediate need associated with their records.
MR. CURRAN: Uh-huh.
MS. NOBILE: So at some point -- and this is where -- this is history, this is prior to 2004. So it looks like we have -- used it for end users.
MR. CURRAN: Okay. Good to know. Recognized that we look at these policies, and they change over time. So there could have easily been a circumstance where this was the only mechanism, someone would have to do policy archaeology, which isn't my specialty. Owen?
MR. DeLONG: To answer Leslie's comment first, and then the comment that I was going to make. Owen DeLong, Jupiter Networks. Yeah, 21 end user organizations have immediate need associated with their records. The current policy section 4.3.2 minimum assignment, single connection minimum block is a /20 -- you can get anything larger than a /20, that you can justify -- 126.96.36.199 multihomed connection for end users who demonstrate an intent to announce their requested space in a multihomed fashion, the minimum block of IP space assigned is a /22. If assignments smaller than a /22 are needed, multihomed end users should contact their upstream providers. When prefixes are assigned which are longer than /20, they will be from a block reserved for that purpose. It does not have any upward limit on the amount of space you can get.
MR. CURRAN: Okay. So you'll assert this policy proposal is unnecessary?
MR. DeLONG: I maintain that assertion, yes.
MR. CURRAN: So as such you are against the policy proposal.
MR. DeLONG: Yes, I am still against the proposal.
MR. CURRAN: Okay. Thank you. Next, center forward mic.
MR. WILLIAMSON: David Williamson, Tellme Networks. For curious reasons, I am a big fan of PI space. I'm also against this policy. I agree that it is probably completely necessary, and I'm really frightened by an immediate need policy for end users that doesn't have the word multihoming in it. I think if without multi-homing you're just asking for an (off mike) walkup say, yeah, I have got a DSL line, and I have got, you know, a /18 behind it. Do I need an address? It seems unreasonable. So, I'm in fact against this policy as written. I'm in favor of clean up of policy language, so I'm looking forward to the next iteration that fixes the confusion.
MR. CURRAN: Okay.
MR. SEASTROM: Are you looking forward to coauthoring the next iteration?
MR. WILLIAMSON: Sorry, what?
MR. SEASTROM: Are you looking forward to coauthoring the next iteration?
MR. WILLIAMSON: I'll give you my card.
MR. SEASTROM: Thanks.
MR. CURRAN: Okay, microphones are going to be closed shortly. Back rear microphone.
MR. BICKNELL: Leo Bicknell, Harrah's Entertainment. Owen also didn't get to the next section. For those without computers, where it has a utilization rate, and specifies 25 percent immediate utilization, and 50 percent utilization within one year. And I would wonder since we have specific rates for end users listed, if the Immediate Need Policy wouldn't actually kind of create a conflict of those written, since, staff seems to apply a more rigorous documentation and other standard to it.
MR. CURRAN: So as a result of that perceived conflict, you're against the policy proposal?
MR. BICKNELL: I believe based on what I have seen here and heard that, yes, I'm against it.
MR. CURRAN: Okay. Thank you. Okay. I'm going to be closing the microphones shortly. Approach the mics if you would like to speak in favor or against this policy proposal. I'm closing the mics really, shortly. The microphones are closed. Thank you, RS, for your presentation. (Applause)
MR. CURRAN: At this time again it's time to help the AC with their deliberations by seeking a show of hands regarding this policy proposal. We are going to ask for a show of hands in favor and a show of hands against advancing this policy proposal. I would like to make sure if everyone is ready. My counters are ready. All those in favor of advancing Policy Proposal 2007-10, please raise your hand. Raise your hand nice and high. All those in favor, please raise your hand. Okay. All those opposed to advancing Policy Proposal 2007-10, please raise your hand nice and high. High, high, don't give up now. You're almost going to be counted. The eyes are going to pass over you. Okay, thank you. We have a count. People in the room, not counting staff, 96. All those in favor of advancing 2007- 10, zero. All those opposed 18. Thank you very much.
MR. PLZAK: Okay. Let's move on and look at Proposal 2007-5. And this is Changes to IPv6 policy -- removal of "multiple /48" justification. This was proposed in January this year, and designated a formal proposal in February. This is the first meeting at which it will be discussed. No revisions. Removes the current policy that ISPs are required to obtain approval from ARIN when assigning an additional /48 to a single site. Shepherds are Dan and Bill. Counsel sees no liability risk. Staff says, "ARIN has not yet seen or reviewed any requests for additional or subsequent assignments to the same end site." "The existing policy text is confusing and might indicate to some that even initial assignments of more than a /48 would need to be reviewed by staff. However, we have not done this and have allowed larger than /48 reassignments to be processed." "In addition, the policy text contains no criteria for ARIN to use in their assessment of assignments larger than /48." Note that the -- this would require a removal of section 188.8.131.52 from the NRPM. From the standpoint of implementing this, impact would be minimum. It could be done within 90 days. It would require some changes to our registration software, a change in the template, some guidelines changes, and some staff training. From the standpoint of the PPML discussions, been 8 posts by 5 people; 1 for, 2 against. Comments, "This I am for -- I think that ARIN policies ought to care solely about the justification to get more space or retain space, and not how assigned/allocated space is redelegated." "Absent any actual policy on what justifies an assignment greater than /48, we should continue to require ARIN oversight of these assignments." Policy text is in your meeting program, and is also available at this URL. And so I would ask Jordi to come up and make his presentation.
MR. PALET MARTINEZ: Okay. This is the text that we have now in Section 184.108.40.206 of the Number Resource Policy Manual. When a single end site requires an additional /48 address block, it must request the assignment with documentation or materials that justify the request. Requests for multiple or additional /48s will be processed and reviewed, evaluation of justification, at the RIR/NIR level. There is no experience at the present time with the assignment of multiple /48s to the same end site. Having the RIR review, all such assignments is intended to be a temporary measure until some experience has been gained and some common policies can be developed. In addition, additional work at defining policies in that space will likely be carried out in the near future. So the current text requires that the LIR is justifying to the RIR or NIR when assigning multiple /48s to a single end site. It seems that the reason for this requirement is the lack of experience, which seems to me unreasonable, after a few years this policy has been implemented, even if may not have been the specific cases which used this policy section. There is one additional consideration here is, what happens if instead of multiple /48s we are talking about multiple /56 or whatever prefixes. It seems useless, now that there is already deployment experience, to require a justification from the LIR to ARIN for assigning multiple /48s or any other prefix length. It, in my opinion, is up to the LIR to require actually this justification to their own customers and decide according to that. The LIR will be already responsible to justify to ARIN the usage of any allocated blocks when requesting for more, and this will already impact implicitly in their justification for this kind of, new assignments. With this policy change, ARIN and the LIR staff will save probably resources in a justification, which again, it seems to me unnecessary, and it should be completely on the hands of the LIR. In addition to this, the staff assessment already confirmed that there has not been yet any request for this, however, for them the existing policy text is confusing and contains no criteria to be followed. My reading for this is that this is again, completely useless for them -- the existing text that we have. And that's it.
MR. CURRAN: Okay. Microphones are open. We have lots of microphones. No reason to bunch up.
MR. HOWARD: I'm Lee Howard from Stanley Associates and also the ARIN Board. I'm -- and maybe I haven't understood the complete justification. But it looks to me like, if you completely strike section 220.127.116.11, then there is no policy under which an LIR or ISP can allocate to you -- can assign a second or future address block to somebody who already has one.
MR. PALET MARTINEZ: I don't think that will be the case because you already indicated in the -- in a previous change of the policy that the length of the prefix that the LIR is assigning to the customers is up to him. So that -- it's -- yeah.
MR. HOWARD: The first time there is nothing saying that once you've used a /48 or a /56 that you can then give somebody another /48 or -- and I haven't seen -- I think, as I recall, the guidelines say you can assign -- well, the following guidelines may be useful; /64, /56, /48. Nothing says anywhere that a /47 or a /44 would be a reasonable, or a good, or even a bad assignments to an end user.
MR. CURRAN: Lee, let me ask. Your reading of this is such that, if the policy was changed as suggested, there would be potentially no way to make an assignment -- an --
MR. HOWARD: Exactly.
MR. CURRAN: -- an additional assignment once that organization comes back.
MR. HOWARD: Exactly, that is my concern.
MR. CURRAN: Okay. As a result of that, are you for, or against, or have no opinion on the proposal?
MR. HOWARD: I abstain by policy.
MR. CURRAN: Okay.
MR. PALET MARTINEZ: The point I think it -- I will need to review the resource policy manual to verify what you're saying. Maybe you're right. But in that case, I think we will be in front of two possible cases; assigning multiple prefixes, or a different size of the prefix -- a longer one, or -- actually the other way around. A shorter one than the one that was assigned at the first time.
MR. WOODCOCK: Somebody --
MR. MARTINEZ: I think there may be even a situation with two different cases.
MR. WOODCOCK: Somebody has got to say it. Nobody could ever use a /48. (Laughter)
MR. CURRAN: Okay. We have people at the mics. We'll take the far microphone, right side.
MR. ITO: Kosuke Ito. I was an original member of the editorial teams, and please read the intention of this policy. And the background of the set-up of this policy, or this particular section is that the basic policy -- principle of this policy is based on the minimized overhead, so that the current v4 is still having second opinion on type of the policy it's implemented that's trying to be -- avoid those kind of things in a v6 policy. However, that -- more than /48 assignments -- I should rephrase it. Then LIR is entrusted by the RIRs to -- LIRs -- that which size of the assignment size can be decided. So up to /48. Then more than that, RIRs and other people -- at that time our editorial teams doesn't know what's the real requirement for the -- more than /48, or multiple /48. We didn't even know what's the actual requirement is, so that -- this policy is for the feedback mechanism that what's the -- actual market requirement is -- necessary is in there. So this is the part of the mechanism, not the -- second opinion, that kind of a stuff. It's -- a scheme. The RIRs -- well, once the LIRs is entrusted it to be decided the assignment size, then RIRs never know what's going on in the assignment procedure. So this is one of the feedback policy. So that's the only background of this --
MR. CURRAN: Right. Let me ask a question -- one quick question for clarification. It is true that if the wholesale deletion of the section doesn't actually get the intent of the policy, that the AC could relatively easily fix the wording to get the intent, as described. My question is working on the intent of the policy, are you for or against it?
MR. ITO: Intent -- if the -- keeping the intention, then I'm for it.
MR. CURRAN: Okay.
MR. ITO: But just removing this policy is not quite my -- not -- I'm not for it because this is -- but this policy is until we have a certain confidence over the -- what the assignment size is supposed to be, then that kind of accumulation needs to be done so far. Maybe then, five years accumulation of the -- our experience is good enough to remove this one, then it's okay.
MR. CURRAN: So you're against it even now due to the lack of experience that we have?
MR. ITO: I believe so.
MR. CURRAN: Okay, thank you.
MR. PALET MARTINEZ: Kosuke, one more question. The way the text is right now in the policy, it seems to me it's not clear, if it's referring to the lack of experience from a technical point of view or from the Registration Services' point of view. From the technical perspective, my belief is that we have enough experience. I mean, it's the same to assign multiple /48 that multiple /56 or whatever. That's clear, right? Do you agree on that?
MR. ITO: Yeah, if you -- technically necessary that much of the assignment size, then you just justify it. That's all.
MR. PALET MARTINEZ: Exactly. But from the registry perspective, what the staff is saying is, "We don't actually have a criteria to allow or deny this assignment." So it's a bit contradictory. I think it's clear that --
MR. ITO: Although, this policy doesn't say trying to allow or trying to refuse the assignment. So this is just a feedback mechanism. So RIR doesn't have any criteria, and LIRs also doesn't have any criteria. So this is how much that -- market needs are there for that more than /48 assignments. That's the only --
MR. CURRAN: Okay. We need to move onto other questions.
MR. CURRAN: Thank you Jordi. Front, center mic.
MR. LEWIS: Ed Lewis, NeuStar. I'm against removing the section because it says, a temporary measure until some experience has been gained, I think -- from here we just heard now. The experience hasn't been gained yet, so I don't think that this should not be removed until that gaining factor is at least met.
MR. CURRAN: Okay, thank you. Second, back microphone, center.
MR. DUL: Andrew Dul, Perkins Coie. I do not support this policy. I think it's trying to fix a problem that doesn't exist. I think we still need to collect this information. If the criteria for larger reassignments is not clear, then we should fix that problem, rather than striking the whole section that requires review.
MR. CURRAN: Thank you. Far right microphone.
MR. RICH: Yurie Rich, Command Information. I don't support removing this principally because there is similar wording in the v4 allocation for providing justification for additional space for end users. And so I don't think there should be any difference in the level of stewardship on the v4 space and the v6 space. They should be treated equally. Having said that though, I think that 18.104.22.168 conflicts with language that's available on 22.214.171.124 where it clearly states that RIRs/LIRs are not concerned about which address size an LIR/ISP actually assigns. So there is some conflicting language in this section that probably needs to be cleaned up.
MR. CURRAN: Okay. But this policy proposal isn't the way you would, say, suggest cleaning it up. Okay, thank you. Center, front mic.
MR. DeLONG: Owen DeLong, Jittr Networks, I oppose this policy proposal. I think that -- to put this in context, because I know for me, and I think for a lot of other people, the v6 prefix link numbers are kind of, confusing still. A /48 is roughly equivalent in terms of available subnets to a v4 /8. So what we're really talking about here is a policy that requires the RIR to essentially be informed when an ISP hands a single customer something equivalent to larger than a /8. I don't think that is an unreasonable requirement to preserve. I understand that the staff may be a little bit uncomfortable with the lack of guidance on what they're supposed to do with that information. Unfortunately, none of us really know what they're supposed to do with it yet because we don't have the operational experience. So for the time being, we have to kind of, trust in their judgment about what the right thing to do is, as they are presented with the situation because we can't, as a community react fast enough, and we have to place that trust in the staff. I apologize for any difficulties that may cause, but it's what we need you to do for the time being.
MR. CURRAN: Okay. Thank you. Center rear mic.
MR. SCHILLER: Jason Schiller, Verizon Business, UUNET. I have a question of clarification. Jordi mentioned something to the effect of the LIR is responsible for justifying the usage of the blocks to ARIN. And I'm just trying to understand how this works, because I'm reading in Section 2 -- 6.2.7, utilization. And it says, throughout this document, the term utilization refers to the allocation of /56s to end sites, not the number of addresses assigned within the individual /56s in those end sites. So it seems to me, utilization is judged on the number of /56s that have been assigned versus the number of possible /56s that you have in your allocation. If that's the case, then would it also be true if I decide to assign a /40 to an end site that I could consider all 65,535 /56s in that /40 completely utilized?
MR. PALET MARTINEZ: Yeah, that's my reading also. That --
MR. CURRAN: I believe that's the accurate reading.
MR. PALET MARTINEZ: Yes.
MR. CURRAN: Assignment of addresses to end sites creates utilization in the form of the number of /56s so assigned.
MR. SCHILLER: So if I decided to assign /40s instead of /48s, I could just get 4 times as much space from ARIN? Oh, I'm sorry. Yes, you're right.
MR. CURRAN: This policy says right now, you have to collect the bit of documentation when doing those /40s.
SPEAKER: But if it's removed?
MR. CURRAN: If this policy is removed, then you don't need the documentation when you do those /40s.
MR. SCHILLER: Yeah. With -- and in that case, I'm very much concerned that ISPs can just start assigning large chunks of numbers and getting huge address space.
MR. CURRAN: So do you think that this is a good policy to change, or are you for changing it through this policy proposal or not?
MR. SCHILLER: I am against the policy --
MR. CURRAN: Okay.
MR. SCHILLER: -- as it is written. I think that there are problems where end sites may require more than a /48. And I think that that's probably from the definition of end site. The problem, as I see it from a routing guy, to me, an end site is a stub network that has a connection point. And the way end site is defined, it's defined as an end-user who has a business relationship with the service provider. So the problem is if I have a customer that has a hundred discreet locations that all exist as end sites, I want to treat them each individually as end sites. But if the billing contact is a single person with a single address, I think the NRPM treats it as one end site. I think that might be a potential problem for requiring multiple assignments.
MR. CURRAN: So that's orthogonal to this proposal, but that would make something that could be changed, and another one that would be more than welcome to be suggested.
MR. SCHILLER: Yes. I agree.
MR. CURRAN: Thank you.
MR. SCHILLER: Thank you.
MR. CURRAN: Okay. Front microphone.
MR. POUNSETT: I'm Matt Pounsett. I'm CIRA and in the ARIN Advisory Council. I oppose this policy as -- or proposal, as written. I think 126.96.36.199 probably needs some editing, but I think removal is the wrong thing to do. Probably because of what Lee mentioned -- I mean, there are -- I'm sure there are cases -- probably edge cases. But there are cases, I think, where an end site will need more than a /48 and we need a way to assign that. And -- but on top of that, I agree, as well, that we still haven't reached that gate of having enough operational experience to decide what to do with this.
MR. CURRAN: Okay, thank you. Leslie, did you have a point of information?
MS. NOBILE: I do. I just wanted to give some numbers. I just checked the database, and we've only had 116 reassignments made of IPv6 space so far. However, 57 of them are larger than /48s. They all came in as the initial reassignment. We don't check that, as I stated previously. So there are many large reassignments being made. And just to clarify, I mentioned in the policy experience report that we use section 188.8.131.52 of the policy. That's the guidelines for /64s, /56s, and /48s. So when we're assessing an initial allocation, and we're looking at the -- the plan to assign 200 /48s, that's when we target in on what size do you plan on reassigning to your customers. We don't just ask for the plan. We don't just ask if there is a plan. We ask to see what the plan is, what are the customers like, what kind of assignment sizes are you giving, and then we compare it to the guidelines, and then we say, why are you assigning /40s, and they will tell us. And, you know, it sounds reasonable, et cetera. So that's where we do the assessment of that.
MR. CURRAN: Thank you. Okay. I'm sorry. The first, far right microphone.
MR. ITO: Yes, Kosuke, again. Can you please clarify that -- v6 policy allow the sub-allocation downstream high speed, so multiple /48 assignment -- and needs to be distinguished by that sub-allocation with like, a /40 or so. So then sub-allocation doesn't count as a utilized space. But as assignment of more than /48 is considered as an utilized space. So this is the -- we should clarify this.
MR. CURRAN: Understood. RS, do you want to comment?
MR. SEASTROM: Robert Seastrom, ARIN Advisory Council. I am opposed to this modification of the NRPM. It's my gut feeling that even if the approval of the -- of blocks that were slightly larger than a /48 were pretty much pro forma, that there is an important function here in archiving what the justifications are for that larger piece of address space over time, which will prove valuable to us in grafting policy in the future.
MR. CURRAN: Okay. Center rear mic.
MR. BICKNELL: Leo Bicknell, Harrah's Entertainment. I'm against the policy, as written. I actually think part of the -- interesting part of this discussion is there is 3 completely different subtopics here. The first one I'm completely for -- striking the note, because that shouldn't be policy. So I think that second bullet point really should just go from the manual. We don't need a note about our experience in the manual. Regarding the actual policy though, it's quite clear that there are two different cases that this policy could apply to. One is a block with a prefix length shorter than /48 to a customer. And the other is a customer that has a block of some arbitrary size, and for some reason wants a second block completely independent of some arbitrary size. And I think those are two very interesting different cases in the v6 world because there is still some intention to try and have -- each end site have one block. And so I think at least the authors -- and perhaps, some of them can comment -- would prefer that an end site have a single /56 as opposed to having 37 different /48s. And that -- there is an intention to drive the policy in that direction. And this one, kind of, confuses those two issues.
MR. CURRAN: Okay. So you're against the policy, but you ask that the author potentially comment on the --
MR. BICKNELL: Well, I think this section of the NRPM clearly needs to be rewritten to be much clear, and to separate those two different situations.
MR. CURRAN: Okay.
MR. BICKNELL: This policy doesn't do that. So it doesn't help us.
MR. CURRAN: Okay. So you're against this as -- and -- but you think that work is needed here.
MR. BICKNELL: Correct.
MR. CURRAN: It would be a great thing to submit. Okay.
MR. PALET MARTINEZ: I think that the input I got clearly is telling me -- it's not delighting. It is working on something different, and probably actually two sections. So I'm going to do so.
MR. CURRAN: We have lots of people who have spoken against the policy. I'm going to close the microphone shortly. If you are for the policy proposal, please definitely find a mic. I'm going to be closing the microphone shortly -- very shortly. RSN, real soon now. Microphones are closed. Thank you, Jordi for the presentation. (Applause)
MR. CURRAN: At this time, I would like to assist the Advisory Council in their work by getting a show of hands about advancing this policy proposal. So the policy proposal on the table is Policy Proposal 2007- 5. We're going to ask for show of hands for everyone who is in favor, and then everyone opposed. I see my people are ready. All those in favor of advancing Policy Proposal 2007-5, please raise your hand. Oh, raise your hand high. It's hard to distinguish between scratching and hair-tossing. Okay, thank you. All those opposed to advancing Policy Proposal 2007-5, raise your hand. Nice and high. Thank you. Okay, we have a count. Number of people in the room and not counting ARIN staff, 96. All those in favor of advancing 2007-5, zero. All those opposed 31. Thank you.
MR. PLZAK: Okay, we are now at the break. So a reminder, Richard will be waiting for you in the Cyber Cafe to spin the Wheel of Fortune, and actually the meeting will resume not at 10:50, but at 11:05.
MR. CURRAN: Okay. Thank you. (Recess)
MR. PLZAK: Okay, welcome back. I'm going to make a few agenda changes. We've already made a couple of -- a couple already and applied during the first session this morning. But before I do that -- before the break, I reminded you that -- actually it was opening sessions this morning that if you wanted to, you could answer a different question today and put your name in the hat. And Richard might draw it out of the box and you could spin the Wheel of Fortune and win another price. And I laughingly said well, Andrew you could win another prize, guess what? Andrew won another prize. (Applause)
MR. PLZAK: But as I said this morning it was not another lava lamp because it didn't exist. So now he has a beach towel I believe, it is to go with his lava lamp. And the other winner this morning was Vicky Cox, and she also got a beach towel. So lightning does strike twice in the same place. Okay, the agenda changes. For the rest of the morning up through lunch, we will begin with the presentation by Louis Lee of the NRO NC report. That will then be followed by the LACNIC report by Raul. Then we'll have a discussion of policy proposal 2007-7 by Alex. And we will conclude before lunch with an inspirational presentation by Axel of the RIPE NCC report. So with that I would then ask Louis to step up to the stage. And Louis does not have stage fright.
MR. LEE: Thanks Ray. Hi, I'm Louis Lee I'm here to give the NRO NC update. The function of the NRO Number Council is to serve as the ASO Address Council, so this means that everything we do we do as the ASO AC. And who are we? Well, Sandy is our ASO AC representative on the ICANN nominating committee again this year. The Nom Com selects several people to serve on various ICANN bodies, including three seats on the ICANN Board of Directors. Martin Hannigan has been working very hard in leading our nomination committee and interview committee to appoint a seat on the ICANN Board, more on that later. And me, I am serving this year as the co-chair of the ASO Address Council. Now how to service as an AC member? There are two ways to become a member of the ASO AC. First you can be elected through an open nomination process and elected by all interested parties residing in the ARIN region. This means that you do not have to be a member of ARIN to vote in the selection. Second, you can be appointed, this is also done through an open nomination process and a candidate is selected by the ARIN Board of Trustees to serve. And what is it that we do? We appoint two seats to the ICANN Board. They are currently held by Raimundo Beca and Dave Wodelet. We objectively process global policy proposals. We are currently refining our procedures to process these global policy proposals, and at this point, I would like to make a distinction between a global policy and globally coordinated policy. A global policy is one that defines how we wish IANA to operate with RIRs for a number of resources. And a globally coordinated policy is one that is implemented via coordination by all the RIRs without the involvement of IANA. Another thing we do is that we advise the ICANN Board on issues related to number resources as well as RIR related issues such as the recognition of a new RIR. And finally we certified that a global policy proposal has followed the PDP defined within each region. Now we conduct monthly meetings in our work, and for each meeting we've established quorum so as to make sure that the viewpoints from around each region are adequately represented. We then review our meeting agenda to make sure that we cover all the topics we need to cover and so we take notes -- take votes to determine the actions of the ASO AC tapes. We demonstrate our openness by publishing the minutes of our meetings. We also do so by including the management and staff from ICANN, IANA, and the RIRs in these meetings. We do not hide procedures, they are published on our website at aso.icann.org. We have a few new faces in our ASO AC meetings these days, since our last update, in the RIPE region, Dave Wilson was elected as Sabine's term ended last year. Sabine if you are listening, thank you very much for your work. And in the AFRINIC region, Ali Badiel was appointed to complete Sylvia's term. Thank you for your participation Sylvia. And joining our meetings from IANA is a new Numbers Liaison, Leo Vegoda. And what's an update at ARIN meeting without some stats. This graph shows how the attendance by the ASO AC members and our observers to the meetings over the last six months. The take-away message of this slide is that the member representatives from the ARIN region get excellent marks for attendance. Our regular mode of communication is over e- mail, our mailing list managed by the Secretariat. We have lists that are for private use, lists that are publicly viewable, and lists for our announcements. We also have special purpose lists made up for our committees to operate in small groups and make things more efficient that way. And we also get our work done through our semi annual face-to-face meetings. In December we met during the ICANN meeting in Sao Paulo. We have just also met here in San Juan yesterday, during this ARIN meetings. And of course we have our monthly meetings which we conduct over a teleconference bridge. In another set of stats, our mailing list activity increased in March mainly due to the work we are doing on the ICANN Board appointment. So what have we done for you lately? We don't have any global policies pending at the moment, nor have we been actively updating our processes, but we are continuing to make some refinements to the ICANN Board selection procedures and making improvements to our website. We have been working very hard on the appointment of the ICANN Board, and seat nine is up for reappointment. Now remember that global v6 policy from last year, we continue to monitor the implementation of that policy, and make sure that the /12 allocations to the RIRs by the IANA were done. We began our ICANN Board Appointment process in early February, which was a bit later than we have wanted. In 2006, there were some things that went very well and some things that went not so well. And so we apply what we learned then into this year's process. The nomination committee and the interview committee both were led by Martin, greatly streamlined our process this year so far. The AC looked at -- the ASO AC looked at the list of the nominations and determined the eligibility of each candidate. The interview committee then conducted face-to- face interviews with the candidates from whom we would like more information. Then the list of eligible candidates were posted in our website for public comments, and the public comment period just closed a few days ago. We are spending this week reviewing the comments we have received, both publicly and privately about each candidate and discussing our thoughts amongst ourselves. And for the first time, we have our online voting system to work out our voting for the ICANN candidate. And at its conclusion we will announce our selection on May 8th to begin the term in June. We've had some challenges in identifying exactly the procedural steps we were to take in our board selection. A couple of that with the actual review of our list of candidates and I ended up sharing the longest ASO meeting ever held, almost 3 hours. In our selection process, we identified eight out of eleven nominees as being eligible, four of which we invited to interview at the ICANN meeting just recently held in Lisbon. These eight candidates that we deem eligible, were listed in our website for public comments. We were a bit surprised that the number of responses to our call for comments, there were a lot more than we've seen in the past. Here are the eight candidates we have for appointment to the ICANN Board. Some of whom you may know, and I won't read all their name out for you here, you can see them. Over the next six months, we will be attending the ICANN meeting here in Puerto Rico and completing our appointment to the ICANN Board. We have a few administrative tasks -- sorry, we have a few administrative projects to work on, to address the challenges we've been facing in our primary duties. We continue to watch out for global policy proposals and refine our own procedures so that we are even better prepared to process the next proposal. And let's see. We have an NRO NC election coming up in October. Sandy George's seat on the NRO NC is up for election, please watch the ARIN announcement mailing list if you're interested in serving or if you wish to nominate someone. In summary, we maintain a high level of participation with our peers from every region, ICANN, and the public. We continue to find ways to make our processes more efficient. We are fulfilling our duties and sticking to our defined mission, and despite the challenges we face in our procedures, we will appoint an ICANN Board member to ICANN on time. And the feedback we've received indicate that we are meeting the expectations of our community. Special thanks go to Gerard Ross and his staff at APNIC. It's not easy doing the secretariat functions for the ASO AC, we hope that the online voting will solve some of the problems we've experienced during very important elections. Many thanks also go to the ARIN staff and especially to Therese Colosi. The ARIN staff performed the secretarial function in 2006 wonderfully, and they still provide assistance as requested. Thank you to the NRO for the advice and the support in our mission, and thanks to all the participants on the public policy mailing list. Thank you.
MR. CURRAN: Microphones are open, any questions?
MR. PLZAK: Thank you, Louis. (Applause)
MR. PLZAK: Next up is Raúl Echeberría, CEO of LACNIC, and I see he has finished preparing his reports, so he can now give it.
MR. ECHEBERRIA: Good morning everybody. Thank you, Ray. As I'm a very well organized person, so I have finished in my presentation few minutes ago. But I hope that everybody will enjoy this excellent work that I have done. So as usual I will make -- I will give the LACNIC report, it's a pleasure to be here after a few months, I have not seen you since St Louis. I just congratulate ARIN because the organization of this meeting is very, very good. And so -- also the contents of the discussion are very interesting. And the most interesting thing is the foosball tournament, and I support ARIN strongly to continue organize this foosball tournament, especially continue making the arrangements for -- sure, we can talk later. Later we'll meet again.
MR. CURRAN: So you can get in more practice.
MR. ECHEBERRIA: Okay. LACNIC is growing -- continued growing since its creation. Yesterday night, I had a very interesting talk with some of you, especially with Lee Howard, is he here. We were talking about the growing of the -- of our organization. That's if -- you cannot realize that LACNIC was created with -- we have about 130 members. In fact, when we agreed with ARIN in the transition plan, we had less than 100 members. And now we have 434 members. We are -- now as we will have an important increasement, because due to the arrangement that -- the agreements that we have signed in with Braslia and RIR is we're starting to recognize the membership rights of Brazilian ISPs. So we will have this new 135 members that are not included in this graph. So the total number of our members in the next few days will become about 570, for this very good news. This distribution of the membership per country is not so important. But due to the fact that we are here in the Caribbean region, it is important to remark that there's about 13 percent of our members come from the Caribbean region, especially from the Netherlands Antilles where we have eleven members. And we have about three, four members from each of the other Caribbean countries, and in total so that we have 52 Caribbean members. The staff is also growing in order to give some good services to our community, and as currently we have say fifteen full-time employees, and beside that we have about six part-time people as we continue having our outsourcing agreement with Brazilian registry. And so we have also some services that we would receive from their junior department. As in 2005, as I have already spoken about this issue, that we bought an old house in a very good point in Montevideo for building a new -- our new facilities. So they look at us in this way, 2005, because we made a very good -- very interesting investment. This is after having worked for nine months we had the opening of the new facilities in 2006 -- in December 2006. As in fact we move to the newer facilities in December 2006, the day after the big opening. Thus it was a very interesting meeting as -- I was a bit afraid because if something happened it could happen at that moment in Montevideo, probably it would have a very big impact in the stability of the Internet, because we had our collection of very important people related to Internet in the same place at the same moment. I realized that it was very late, so as I -- the only thing that I could do was pray. But as we have the more than 250 persons in this celebrations and more than 100 from around the world, many friends from the RIR's community, staff members, CEO's, and board members of our colleagues organizations. We had also many visitors from ICANN, David Conrad the director of IANA, with Michael Schumer from ICF, people from ITU governments from the region, members, association of ISPs from the region and also very important authorities -- local authorities. It was a really nice party. And so as I said the day of the openings as you can -- let me go back -- remember how was the house in 2005, and in 2006 it looked in this way, that's very nice. So that's -- the celebration was a very strange, because of very uncommon things happened during that party. And so let me know raise your hand if you have seen Ray before wearing a tie and suit. I have many other strange pictures. I have also picture from Axel Pawlik, Raimundo Beca wearing tie and suit too. But not everything is nice. So we are growing and we have new facilities, but there are many things that are contributing with our stress in the working environment, so let me show you my office now.
MR. CURRAN: Is there Board openings? (Applause)
MR. ECHEBERRIA: The interesting thing is that there you can get the wireless network from the beach, and so I cannot do it, so I have realized that it would not be a good example, but -- okay let's come back to work. So during 2006 we have approved in the -- people have approved some policies in -- according to our PDP. I would not speak about each of them, and also Richard have made a very good summary in the beginning of the meeting. So beside the policy that we have adopted during 2006, some important resolutions have been taken in our annual meeting, one of them is about the reduction of fees for non-profit organizations of actually 50 percent. Also the annual assembly after considering the things that were happening during the information society summit, is they decided to share -- to mandate the LACNIC staff for doing some actions in order to increment their relationship with all the stakeholders in the region. We have worked very much in introducing planning during 2006, as the scope of that exercise is 2009, we have new definitions about our vision, mission, objectives and also the projects. These are big challenge for the organization because it implies a very change of mind. We have to change dramatically the way which we work. And it is also -- we defined some adjustment in the organizations. We have -- we are adapting softwares, accounting softwares, for example for separating everything from the -- what is related with strategic planning and operative expenses, for examples. And it also means that we have to prepare annually, different fashion of budgets for being considered by our board. So we are now in the process of working on those items. Our new vision, I think this is very interesting, because you can see that's we see ourselves as a much more than an organization that administered IP addresses. We see ourselves as -- as an important player in the construction of collaborative efforts toward the development and stability of Internet in our region. Also I will not bore you as about the mission, but you can find information in our website. We continue involving in other projects with a big focus in development. We continue with the +Raices project, which means plus roots, as we have installed on their agreement with ISC three anycast root servers in Chile, Argentina, Venezuela, two more coming are coming very soon, one in Panama and another one in Ecuador. We have also received through ISC support from Cisco in the donating some of the equipments that are involved within those projects. We are currently looking for partners in the Caribbean region. We continue with a second phase of FRIDA program. FRIDA program is a result of a partnership among LACNIC, IDRC from Canada, and Internet society. It is a program that consists of small grants for funding research projects regarding an ICTs. The total amount of this phase is $650,000, and the next call for projects will be issued exactly this week, probably tomorrow or the day after tomorrow. We are involved also in other actions like supporting and organizing different meetings around the regions. We have sponsored an Internet -- an Internet day in Argentina last week, also we have sponsored and participated in the Caribbean Ministerial meeting in Anguilla, organized by the Caribbean Telecommunication Union. It is an activity in which we participated whether with ARIN, and also with other important Internet organizations. We are also working in a traffic measure projects. It is a very important issue for the region, because the interconnection cost and everything related with the interconnection has been defined as one of the most sensible, and important topics for the regional community. We are actually working in a same project with PCH, and we expect to have some announcement regarding this project and during LACNIC -- our next meeting. In LACNIC X, our next meeting, we would have some policy discussion as usual. We have now in this -- in the agenda 7, policies for being discussion. This is the information that I received from the policy chair today. So the deadline for submitting proposals is today, so probably some other proposals will come soon there are some very interesting proposals -- I should ask to you to check this everything -- all the information regarding those policy proposals that are being under discussion. In LACNIC X, as usual we would have many activities like tutorials regarding network monitoring and IPv6 implementations. We will have a very interesting and important meeting of Regional Interconnection Forum in which we will have a debate also about what does interconnection cost -- or interconnection products means for all this stakeholders. As we would host the fifth Latin-American IPv6 Forum as the LACNIC annual members meeting, and there is also, as usual our open policy forum, as other organizations like the Latin American Association of CCTLDs and Internet Society will host regional meetings also together with us. So you are all welcome to our meeting. The meeting will be held in Venezuela in Margarita Island, in order to continue with this new fashion of RIR meetings thus in very nice islands. So I welcome to Venezuela, Margarita Islands in the -- from 21st to 25th of May, thank you, muchas gracias. (Applause)
MR. CURRAN: Thank you. Any questions? Thank you very much.
MR. PLZAK: Thanks Raul. That picture of Raul's office did not do it justice. If you look at the front of the building, if you recall on the second floor, there was a row of glass, Raul's office actually projects out over, so he has glass on one wall, glass on a second wall, glass on a third wall. He doesn't have a corner office he has a front.
SPEAKER: This big window office.
MR. PLZAK: Yes. And it is very comfortable.
MR. CURRAN: ARIN staff should know there are no openings, okay, don't even think of it.
MR. PLZAK: But one other thing I would like to point out, I can remember the first LACNIC offices. It was like a 10 x 10 room at the most. And so in a very short period of time, they have grown because this is their third, actually fourth facility that they've been in, so congratulations. Next step is policy proposal 2007-7: Creation of Policy for Subsequent end-User IP Requests/Assignments. Here it is, introduced on the PPML in February of this year; it was designated a formal proposal in March. This is the first meeting at which it will be discussed, and no revisions. It establishes the criteria for a subsequent end-user IPv4 address space requests. The shepherds are Alex and Suzanne. Legal assessment from the standpoint; no liability risk. Another meritorious comment by the Counsel is that, "We also believe that it is useful to be specific in a manner articulated by the proposed policy." Our staff has no comments other than the note that this would cause a change to the NRPM by adding a new section 4.3.6. Minimum is the resource requirement and requiring only changes to guidelines in some staff training. 10 posts on the PPML by 6 people; 4 for, 0 against. Comments; "I entirely support all of this proposal, and policy text is absent, meaning it is left to the staff's discretion. Is the current arrangement unacceptable? If so, why?" Proposal is in your packet as well as this URL, and so Alex, if you could come up.
MR. RUBENSTEIN: Thank you. I'm here to discuss Policy Proposal 2007-7, which is Creation of Policy for Subsequent End-User IP Requests and Allocations. What is it? You could read it online, as Ray said. It's creating a new section 4.3.6. As I said it is a new section, not a change of any existing policy and it will address the assignments of additional IPv4 resources to end-user. The proposed text as it stands -- "Additional Assignments: In order to justify an additional assignment, end-users must have efficiently utilized at least 80 percent of all previous assignments, and must provide ARIN with utilization details. The prefix size for an additional assignment is determined by applying policies 4.3.2 and 4.3.3." Why? Today there is no written policy to address an end-user coming back to ARIN for additional IP space. If an end-user does apply for more IP space, the ARIN staff applies what they perceive to be "efficient utilization" criteria. For instance, the end-user must have utilized at least 80 percent of the last assignment, and may provide ARIN with utilization, I'm sorry, must provide ARIN with utilization details. While this is practice, there is no policy today. This policy is part of a general NRPM cleanup initiative, and it will politicize what is already being done. Is it safe? On April 13th, staff assessment post stated that they had no comments, and that the General Counsel believes that it is good. Resource impact is minimum, overall PPML postings showed support because it clarifies a policy that will follow current practices. I saw two significant posts on this policy, one was somebody asking is there something that needs to actually be done. I responded saying that the ARIN staff has asked for guidance on this topic, and I didn't get anything further from him. And one person said I entirely support this proposal, the author should be thanked, you are welcome. Two e-mails in 53 days, some sort of record. And that's it, any questions.
MR. CURRAN: Microphones are open for questions.
MR. RUBENSTEIN: As I suspected.
MR. CURRAN: Microphones remain open. Yes, Leo.
MR. BICKNELL: Leo Bicknell, Harrah's Entertainment. I support the policy as written. I think, perhaps it might be better to change the last sentence instead of having both sections to just say as in section 4.3, so that if other policies are added or removed in the future regarding end-user assignments that all still applies, because I think that's your intent. And I think possibly that's something we as the AC could just do, if others are interested in that -- other than that perfect.
MR. RUBENSTEIN: Okay.
MR. CURRAN: Okay, thank you. Far microphone left side.
MR. DESTERDICK: Mark Desterdick, Verizon. And I just hope I am not opening a can of worms here, when --
MR. CURRAN: I'm sorry, sir.
MR. DESTERDICK: Okay.
MR. CURRAN: Are you for or against the policy proposal?
MR. DESTERDICK: I'm against it as written because I don't feel that it's explicit enough in terms of additional -- efficiently utilized at least 80 percent of all previous assignments. Would that include legacy assignments and space that was not assigned out of the ARIN -- out of ARIN. And I guess, that's where I'm looking for additional clarification, because even having this policy if it's because -- ARIN staff is looking for guidance on how to handle it, I think we need to address if it's going to -- if the utilization of legacy space is going to be included or not, is, you know --
MR. CURRAN: What's -- what is your recommendation on the matter?
MR. DESTERDICK: My recommendation is that legacy space be excluded from any utilization calculations, and that only -- only space that is know to ARIN, and that has been allocated by ARIN be included in any consideration of utilization, thank you.
MR. CURRAN: Can I ask Alex as a point of information, what is your intent with the policy proposal with respect to legacy space?
MR. RUBENSTEIN: I haven't actually thought about it, but I think that it implies it would not be included, because it specifically states, must efficiently utilize all previous -- I'm sorry, would be included because it specifically states all previous assignments.
MR. CURRAN: I'm -- I am about to go for what's the practice, Leslie. Could you describe what ARIN's practice is in interpreting all previous assignments?
MS. NOBILE: Yeah, we actually do look at legacy space. So if somebody has a, you know, /8, we say, what are you doing with it? Why are you coming back to us for additional space, tell us what you are doing with what you have.
MR. CURRAN: Right. And so if this policy were adopted, is it correct to assume that you would interpret all previous assignments to include legacy space, okay unless it was otherwise changed, okay. So for clarification, for purposes of assessing the policy proposal 2007-7 recognize that ARIN staff will interpret this to mean all previous assignments including legacy. Yes, Owen.
MR. DeLONG: Owen DeLong, Jittr Networks. I support the policy as written, I would also support Leo's amendment, I think it is an improvement, and I would absolutely oppose any policy that attempted to not include legacy space in any such consideration.
MR. CURRAN: Okay, microphones are open. People in favor or opposed to the policy proposal, who would like to speak your mind, please find a microphone. RS?
MR. SEASTROM: Rob Seastrom, ARIN AC. This seems to codify a reasonable policy without unduly constraining staff's latitude and therefore I support it.
MR. CURRAN: Okay. Microphones remain open. Front center mic.
MS. SCHILLER: Heather Schiller, Verizon Business. I am in favor of the policy, but with regard to the comments about legacy space, and the legal lawyer is still here, right.
MR. CURRAN: Uh-huh.
MS. SCHILLER: So he can correct me, but folks who have legacy space aren't held to the NRPM because they never assign the RSA, is that correct?
MR. CURRAN: So I am going to actually unfortunately, I'm going to rule that orthogonal to the question I can't, but we have an open microphone session coming later, and it will be a great topic.
MS. SCHILLER: But I guess what I am saying is if people are going to make a decision about whether or not they are in favor or against this policy based on that -- that question -- clearing up that question might help or change people's opinions on how they would vote.
MR. CURRAN: Correct. And so as I said, for purposes of interpreting this policy, ARIN staff has indicated that they do currently assess legacy space, and would continue to assess legacy space in interpreting this, if so adopted.
MS. SCHILLER: Okay.
MR. CURRAN: Microphones remain open. We have people in favor or opposed, that's two very big circles of people, please come up to the microphones and speak about the proposal, if you have anything new to add. I am closing the microphone shortly, any last comments? Microphones are closed, thank you. Thank you, Alex. (Applause)
MR. CURRAN: At this time, in order to facilitate the ARIN AC in doing their function, I am going to ask for a show of hands about advancing this policy proposal, I am going to ask for everyone in favor, and then everyone opposed. I see we are ready for such a show. All those in favor of advancing Policy Proposal 2007-7, please raise your hand. Nice and high. Thank you. All those opposed to advancing 2007-7, please raise your hand. Thank you. We have a count. Number of people in the room excluding ARIN staff, 101. All those in favor of advancing 2007-7, 43, all those opposed, 1. Thank you.
MR. PLZAK: Okay, next topic is the RIPE NCC report presented by Axel Pawlik. He will show you the tremendous view from his office -- we are looking at one of the many canals in Amsterdam. Axel?
MR. PAWLIK: Well, my office has glass at least, which was not the case when I was there to sit in Raul's office and where do I point?
MR. CURRAN: Just anywhere.
MR. PAWLIK: Anywhere, okay point anywhere. So yeah, quick update high level from the RIPE NCC what have we been up to over the last half year or so. Let me see, exciting changes, we have reorganized the office, and reorganized ourselves. We have new departments, and old departments have been taken apart and reassembled, very exciting. About a year ago we set, and thought about how we could increase efficiency of what we are doing, and how we are doing this. Basically we have been looking at goals in departments. We have found several -- slightly potentially conflicting goals in one and the same department, so we split them a bit. It's sounds bigger than it -- than it really is, and we don't grow although, we have new key staff here on the slide. We have what you would call a new senior management -- part of the senior management team. Andrew de la Haye as the Chief Operating Officer these days. We have started a customer services department headed by Gregg Stephenson. The idea here is to take all those annoying customer calls out of the other departments, so that they can focus on doing the work instead of interacting with you know -- guys like you, that is really so annoying and distracting. We already have a new registration services manager, Flor Paredes, she started a few weeks ago, this month only, on the 1st of April, it doesn't mean anything though. She is finding her way and getting to know the team and the like. Then we have Mark Dranse, who was looking after information services. Those things like the test traffic measurement stuff, the DNS1 and similar projects, which in addition to the registry services that we do. And we also are very hopeful that maybe today we can actually conclude the hiring process of an external relations officer, which shows that that we are taking this part of our activities very, very seriously, so we hope she will be with us soon. Okay, that means over the last year or so we have been navel-gazing quite a bit, you know, organizing ourselves. Now, that focus is shifting outwards again. We have quite a number of strategic ideas for some of our services that we are vetting in turn through our board, and of course, we will be bringing out to the community, so watch the space. Number resource certification. As you know, the RIPE NCC is activity plan driven, we write an activity plan every year that is then vetted through the community, and the board approves that by the end of the year usually. We were looking at initially what's APNIC was doing in this field. We said, well this looks very reasonable, we should be doing something like this likely as well. However, we had second thoughts and we are thinking with anything like that would actually be adopted by the community, so we are asking them. We have set up a taskforce with five or six representatives from our members in there, basically the idea was to install a prototype, have them look at it, have them feel how it might impact their business processes, whether it would be very expensive, or awkward to work with this thing. We have had a first meeting, we will have a second meeting in two weeks' time during the upcoming RIPE meeting, we will have reports of this meeting and the next one with the idea that by the end of this year we have some community shared view on what we should be doing with this. As said, we are taking outreach quite seriously with the newly being hired external relations person. You know that we do regional meetings. We did a membership survey, quite some time ago, where we heard the idea of doing something more for our members in the outlying regions, thinking there of Russia, and of the Middle East. We have been going there for a few years now. Basically, not to hold RIPE meetings there, but to bring them some information, bring them updates of what is happening in case they -- they can't travel, and those things have been proved extremely popular. So I am not really saying this here, but we might actually take a RIPE meeting down to the Middle East next year. Membership survey, yes, as I said, we have done that one survey this is based on, I think it must be four years ago. We've done another one, basically, a smaller one to verify that we are on the right course. And we are thinking of starting another one later this year, after we have reorganized the company, after some of the changes should have impacted our members, we want to see how we are doing there, so that will come up. And I have someone here who called the MENOG 1 that is the Middle Eastern Network Operators Group. That is something that we have been posing to the Middle Eastern people more than two years ago, they came together it's great that you are here, and can't -- can't you run something like this more often, and we said, no we can't because you are the operators in the region. You should do that yourself. Of course we would be willing and -- to help them and you know, to organize a little bit. And it took them quite a while, but they have done so, so we did what we pledged to do, we supported it organizationally. I have been known to somewhat cautious about the RIPE NCC involvement in this, I don't want it to be seen as a RIPE NCC activity because it is not, but we helped them they were happy. They had over 80 attendees from 15 countries, I didn't know there were that many countries in the region really, but it was great, apparently, I wasn't there, but Paul said they had great fun, but he would say that right, I guess. So we are planning a second one, or they are planning a second one, we are planning to support the second one. Adjacent to the regional meeting that we'll be doing there in Oman, quite likely in November this year, that will be fun. Government and enhanced corporation, that is always fun. We do round tables for governments and regulators. They have been quite popular; I must say the numbers of people who come, are not extremely high. Last time we had about 45 people, I think roughly in the room, about two thirds of them would qualify as the intended audience for this. It was actually quite good this time because just the week before we had a DDOS attack on the root servers as you know, and then in the APNIC region, the IPv4 countdown proposal came up, so they were quite interested in these things, it was great. Now seeing that not that many of them are coming to any one -- on any one day, I said we -- let's -- let's do another one. We'll have one in September -- on the 17th of September, the idea is to give an update of what happened over the last half year, but also given another date for the others, you know, hundreds and thousands of others who should be coming there as well, so maybe another 50. We also have heard from that same group that these rounds tables are just great. That they love to come to Amsterdam and have them. However, they would also like some -- some closer contact with the community because those roundtables of course are exclusively for governments and regulators. So we said, yeah, you can go and meet the community. Of course you are welcome at RIPE meetings in any case, but we will reserve the Monday afternoon of the RIPE meeting for you, for the interaction between governments or regulators and the community. And we find it very hard to get talks to that, that is strange. Of course the same people who said they would like this said we can't speak as government's openly in public places. We don't feel comfortable with that, we don't want to do this, so we don't know what to do with this. If you have any ideas let us know. We'll continue to reserve that slot because it's there, and you know, we want to give them the opportunity to do this. I don't know ITU study group 2 I went there basically talking about, ENAM and what we do for them and update there, but also of course I was asked whether the IPv6 deployment wouldn't win the routing tables totally, and whether that there is something that we've taken in to account, so basically it's liaising with those guys regularly. As said external relations group, there is a group now which is being developed. We want to do as one of the first things -- things that we've wanted to do for three years already, and never had the time or the -- or the people to actually do that, look at -- at who our stakeholders are ISPs -- other industries there -- is there anything that we should do for them specifically, things like that. And of course we want to increase our PR activities towards the governments to just more, and maybe slightly different things. And of course together with our colleagues in the RIRs and the NRO we -- we will go to (off mike) and do our utmost to defend what we are here to do, our -- with our processes. Now, let me see, yes. Now, I believe this is the ARIN meeting. I've been led to believe that it would be lots of sun and lots of beach, we've seen some beach, not that much sun. So this is a case of -- of (off mike) expectations I must say. So of course you're all welcome to the Tallinn meeting, and if you've seen the slide before, probably, it will not be covered. Probably we won't have snow, and I've heard that although the winter goes a little bit late up there, it should be sort of spring like, I am not promising anything, if it's nice you'll be happy, I hope, otherwise I hope to see you all there. Thank you very much, if you have any questions, let me know.
MR. CURRAN: Any questions for Axel?
MR. PAWLIK: Thank you.
MR. CURRAN: Thank you. (Applause)
MR. PLZAK: Just a reminder on the weather, as I pointed out during my opening remarks on Monday, Susan wanted to make sure that everyone stayed in their rooms, so she ordered this all up for us. Okay, it is now time for a lunch break. So let's first of all thank everyone for their support. (Applause)
MR. PLZAK: A reminder, this room will not be secured, so take your valuables with you. Help desks are still open. Please complete the meeting surveys, and the IPv6 workshop survey. Earn that raffle ticket for that watch. See you back here at 1:30 p.m. Thanks. (Whereupon, at 11:57 a.m., a luncheon recess was taken.)
MR. PLZAK: Okay. Good afternoon. And we are ready to start this afternoon's activities and festivities. We were very efficient this morning, and got a lot done. And so we have one last really important topic to go through with a couple related policy proposals, or could be construed as related policy proposals. First up is a panel discussion relative -- discussing the IPv4 consumption and the panel consists of the ARIN Board and we are all up here. And to set the stage I've got a few slides to go through. First of all, this is the trend through 2006. Leslie gave you the stats the other day through March of 2007. But this is sufficient for our discussion here in that you see that there is an increasing demand for IPv4 space, and here is what is the allocations to the ISPs by the various RIRs. And with a slight dip that occurred in the 2002 timeframe, you can see there's a general steady trend growth of demand up, which is a good thing. It means the Internet is expanding and doing all those nice and even wonderful things. But with demand is always the other piece of the economic puzzle, which is supply. This is the devolution if you will, of the central pool as administered by IANA up through March of this year, and you can see a steady decrease in supply. So when demand increases and supply decreases, that tends to cause some problems in that if you take a projection -- and this one that's done by -- was done by Geoff Houston you see two curves there, the red line that's done a steady downhill line is the IANA pool, the green line across the bottom is relatively stable if you will. That's the RIRs maintaining a relatively level supply based upon being able to draw from the IANA. But at some point you see that the red line, it goes down to nothing, and then shortly thereafter you see the green line sort of do the same thing. And what that means is on the projections based upon past consumption by the RIRs and the ISPs, that somewhere around 2011 the central pool is expected to be depleted. And somewhere 2012, 2013 the RIR pools are expected to be depleted. And so this leads to the question of what could happen? Well, number of things. Well, expected that IPv4 demand could continue, and there's no reason to expect that it would not -- could continue. You can expect -- some people would expect the use of NATs to increase. Some have strongly suggested that you would see a secondary or grey or black -- whatever color market you want to call it, actually become very prevalent. And lastly, you can see the frantic deployment of IPv6. The term "frantic" is used because general indications is that IPv6 is not being deployed as rapidly as those that would love to see it deployed are expected -- are experiencing. So when you're going to do a frantic deployment situation, things obviously become costly. And so with IPv4 space being depleted, and actually going back to something that Scott mentioned I believe it was the other day, you know, was that these things were looked at in the '90s, and the IPv6 was sort of designed to take the place of IPv4 because it was recognized at some point of time that this would happen. So the real question is that if IPv6 is to replace IPv4, why not IPv6? Well, there are some tactical issues. Lea alluded to some of that the other day and the influence of the RIR policy process -- sounds maybe pushing the technical solutions along a little bit, and the other one really is economic incentives. There has been a number of things happened over the last several years in the RIR policy process to attempt to use the distribution management policies to spur the adoption of IPv6, and it's been done both through policies and also through fees. And if one was to note that for example in the ARIN region in some shape or form, fees for IPv6 address space have been waived since the year 2000. And it's been waived pretty much so in the other regions in one shape or form as well. And the policy since the globally coordinated policy was adopted in 2003, basically has become more and more liberalized. And in fact one could consider that every one of the policy proposals that's happened in the v6 arena over the last several years have been designed to make it easier to get IPv6. And so from one perspective I've heard it said that, you know, if you look at the liberalization of the RIR policies and the fact that fees have been waived, in a sense the RIRs are attempting to give it away and are not succeeding at that either. So the real question is why not IPv6? And so the real thing is what can be done to take care of those technical and economic incentives? So the stage is set. The IPv4 demand is increasing and the IPv4 supply is decreasing and IPv6 is not there. And these are the things that the panel would take up this afternoon, and I'll turn that over to John.
MR. CURRAN: Thank you. Okay. So the Board is not without notice of the changes that are happening with respect to the IPv4 address pool. Clearly, some of us have been working on it for a decade or more. But what we are trying to do is we're trying to figure out what steps should we be thinking about for ARIN as an organization, so that we can be true to our mission of stewardship but access to addresses so that the Internet can continue to grow and expand. Those are contradictory goals in some occasions, and it means that as we move forward I expect that as the pressure between the two of them open access to Internet numbering resources and responsible stewardship gets more and more difficult, we're going to see a lot of innovative policies and suggestions. We actually have two policy proposals right now on the docket, which we'll talk about after this panel, but there's also been a few other suggestions that people have made about what ARIN should do to provide appropriate stewardship, apply appropriate back pressure, apply appropriate incentives for v6. I guess I'm going to ask each Board member now whether or not they feel as though there is more that can be done, or is there -- are some of these goals such as incenting v6 or applying back pressure actually meaningful? Any volunteers from the Board to go first? Scott?
MR. BRADNER: Well, let's see now. Long time ago working in v6 space, Mr. Curran wrote an RFC that he referred to the other day, which basically said if there isn't a reason to deploy, there isn't a reason to deploy, and that's not going to be successful. That's the gist of the RFC. I think that's still the case, though we don't -- we're not in a position in my opinion, to actually be pushing v6 quite yet. We're at the position today that we were at before Microsoft recognized there was an Internet. There was a time when Microsoft wasn't including v4 stack as part of Windows. And it made for good lunches and dinners for FTP software and Wollongong software and renting out the Aerospace Museum in Washington. It was a nice time.
SPEAKER: (Off mike)
MR. BRADNER: I didn't quite -- to get that.
MR. CURRAN: Intercon was one of those companies --
MR. BRADNER: Intercon too, yes. Well, there were a number of them. It was a good time had by all for a short period of time until -- well, actually not that short a period of time. Well, Uncle Bill finally woke up and yes, there is v6 and Vista, yes there is v6 and XP, but you got to tweak it a lot and turn it on and all kinds of stuff. And yes, this stuff in Vista -- well, Vista is still as its name implies, out there on the horizon someplace. And so I don't know that we're in a position to say, well, let's do more to push v6, because it isn't there; it isn't there on the common desktop. Sure it's there on the desktops of many of the desktops that are up here, the laptops that are up here. But what's the advantage of incenting people to deploy something they can't use? And most of the industrial world is in a condition where they can't use it. So I don't see anything on the incentive to deploy v6 side giving away addresses, giving away the labor of allocating addresses let's say, not giving away addresses. We've tried that, doesn't seem to make much difference. Now, so the other side is the putting the breaks on the v4 side. Well, we tried that a long time ago. ARIN and many of the other RIRs that existed at the time were very reluctant to give out large chunks of address space saying, "Well, if you're a single home, why don't you just use a NAT?" And we did that for a while. It didn't seem to make much difference, and today I think it is not going to make much difference either. NATs are going to be growing as the difficulty of -- the perceived difficulty of getting v4 address space goes. And of course it'll be NATs with v6 too, that's a law of preservation of NATs. So I don't see much light at the end of whatever tunnel you happen to be in.
MR. CURRAN: So in terms of incentives we've gone about as far as we can go, and it doesn't seem to be very effective. And in terms of trying to disincent v4 usage, it's not something that looks to be fruitful either.
MR. BRADNER: We've danced both those dances, don't seem to be -- don't remember any successes.
MR. CURRAN: Okay. Any Board member want to go next? Lee.
MR. HOWARD: I got no choice.
MR. CURRAN: Oh, you got no choice? I'll choose someone.
MR. HOWARD: So one of the things that has -- that often comes up in the debate in various parts of the conversation that we have, is specifically how we can use financial incentives and in our case, fees. One of the things that we as a Board have consistently tried to do is only set fees such as to recover the expenses that ARIN has. And we've overachieved that goal in some cases. We've very carefully tried not to use these to incent specific behaviors to more than a minor degree. And that's almost a theological point of view again that we don't -- and largely because we don't want to put small organizations at a disadvantage. So that would include a small organization that wants to deploy v6 or a small organization that is trying to deploy IPv4. We want to make it possible for them to use either set of address space without financial -- without creating a financial disadvantage for them, because we have tried to encourage the development of the Internet in our region. I will also say that -- well, let me say that I don't speak for the entire Board or the finance committee when I say any of that. I can only speak for myself, but I do -- and I will -- but I will make one more -- I think I owe the organization a commitment that -- or a clarification on the current set of waivers. As John pointed out -- I'm sorry, as Ray pointed out, we've been trying to give this stuff away for free or nearly free for as long as we've been giving this stuff away. And there haven't been many takers. Right now, the few waivers are said to expire at the end of this year. And what's not explicitly clear I think in the current fee schedule is what the fees will look like after that, whether as an IPv4 holder fees continue to be waived or whether you have to pay both. I think the -- well, so would -- the finance committee and the Board will get you a clarification on that. But again, I think we're trying to use financial incentives from ARIN's perspective very gently in terms of directing behavior because of the different levels of financial resources available. And I don't think we've seen -- we have seen many different proposals, and obviously the finance committee is going to be discussing those in the very near future.
MR. CURRAN: Okay. Any Board members want to comment? Sure. Right at the end.
MR. PLZAK: Ah yes. Specifically, the issue of incentives. And there are several different major categories of stakeholders here. Obviously, this community is one set of stakeholders, but there are other stakeholders. They get characterized in the Internet governance debate as being private parties, the public sector, and so on. There is certainly -- if you look at the public sector, that's the code word for governments. There's certainly things that governments could do. From an economic perspective in -- to someone that has already entrenched value in IPv4 use for -- you know, you've got that equipment; you've got that network, what's your return on investment to convert IPv6? Because that is a strong economic incentive. Anyone that's gone through any renumbering, substantial renumbering, that's what you're looking at to do perhaps; replacing equipment sets -- those are new costs. There are different things over there, and so the question becomes is there some encouragement that different people could give to their governments to do? I think someone the other day mentioned tax breaks. I don't know if that would work with giving these tax incentives to people to use v6. Certainly, the developing portions of the ARIN region, and they exist in every single country in the ARIN region to include the United States. There are other things that would be incentives such as loans and grants and so forth. From the private sector code word businesses, again it's adopting it and using it. Governments adopting and using IPv6 is another thing. Strong incentive for a government to take is that if you want to do business with me, you have to use IPv6. I think there's some push is starting in that direction with regards to what is going on with for example the U.S. Government. So Lee said yes, there is some financial stuff that we can do but it's very minimal, and it only affects a small base. There's a large number of other things that could be done. And so while we basically are concerned with the management of the distribution allocation assignment if you will, of address base, there's not a lot else that we can really do about it. And so perhaps we need to think of ways, and maybe the community could help the Board maybe formulate a letter or something that would go to ICANN or somebody that -- to encourage other things to occur.
MR. CURRAN: Okay. We've had two Board members, yourself and Scott, basically say this whole transition not much we can really do here at ARIN. We can do some education and outreach, but not much more. But we also saw uncover the fact that it's a economic issue, there's a cost involved in leaving the safety and knowledge of v4 and moving into the brave new world of v6. Doesn't ARIN contribute to those costs, organizations one way or another? Does anyone want to comment? I have Bill and Bill -- which one do you want to go first?
MR. MANNING: I'll let Bill go first.
MR. CURRAN: Okay.
MR. MANNING: And so since I'm letting Bill go first -- (Laughter)
MR. MANNING: Go ahead.
MR. WOODCOCK: My thoughts on all this have changed a bit over the last say, six or eight months. And I had a really good conversation with Alain Durand yesterday, which kind of put a cap on that change of thought. A year ago, two years ago, five years ago, eight years ago I thought, well, v6 is obvious, it's necessary, it's a simple solution to a very big problem. It's not a small solution, but it's the minimum necessary solution, and it's sufficient. And when people need it, they'll adopt it. And the idea of trying to encourage people to adopt a solution at a time when there was no economic reason to do so, that is to incur a new cost prior to it being needful, is in an industry that grows as quickly as ours, a waste of money, right? The net present value of any onetime investment is, you know, high, right? The sooner you take that investment, the more you could have gotten done with that money in the meantime in promoting growth if you didn't have to take that investment yet, right? A thousand dollars in hand today can be spent to do something today. If you don't need to do it until next year, that $1,000 will have grown to $2,000 by then, and you'll only need to spend half as much, right? And so my thought was, well, so we've got v6 addresses, there's no problem getting them. You'll all pick them up when you need them, and that's good enough. And so, you know, I looked at the efforts that particularly, you know, the European Commission have been promoting, of sort of advertising v6, right? At v6 marketing campaign some thought, you know, this is really pointless. This is selling us something before we need it. And I don't change my mind about any of that. I think that lowering the price, lowering the hassle of getting it, you know, giving people bigger and bigger blocks that they're still not using, none of that really changes. That's all supply side. And what we don't have is a ramp up in demand. What we have is a looming cliff in demand that we are approaching the bottom of, right, whereas we'd really like it to be a long gentle hillside that we could stroll up. But that's just not the way it's playing up, right? We're coming up on this cliff very, very quickly. And what projections like Geoff Houston's don't show you is that the rate at which we're pedaling on our bicycle towards the base of the cliff is increasing, and it's going to increase faster and faster as more and more people realize what's coming, right? So it's not 2012, you know, it's a couple of years from now at most. I know how long it takes people to make decisions, I know how long it takes people to roll out new platforms, you know, that's no time at all. That's coming right up. And so I think what I think now as opposed to what I think -- thought a couple of years ago is that yes, people will get this stuff when they need it if they're smart. And if they were smart they would have started on this seriously a year or two ago, because if they're just starting now, you're going to get stuck. You're going to get to that cliff, and you're not going to be ready to climb it. And I don't really know what to do about that, you know. I do know that I as an individual demand v6 services from my transit providers and they give them to me, and that's good, but you know, not enough people are to create a real market demand. And ARIN can't create a broad-based market demand for something from the membership. So I'm a little perplexed -- can't legislate people into being smarter than they are. And I think people are not being smart enough right now.
MR. CURRAN: Okay. Bill number 2.
MR. MANNING: I'm going to owe (off mike) 10 bucks in about five years, because that's when v6 is going to take off, all right? That'll be a decade, little more than a decade late. I see ARIN as an enabler. I do not think that it is appropriate for ARIN to anticipate or economically incentivize people to adopt technologies. I believe that ARIN is appropriately constituted to as Bill said, to give people what they want when they want them. And ARIN as an organization should be prepared to do that. But as soon as we start getting into encouraging or discouraging adoption of particular technologies, I think that we step a little bit outside of our core mission, which is wise stewardship. And so to that end, I'm a little dismayed about a lot of the discussion. I recognize the policies and procedures. The policies are suggested and come through an open process by members and others in the ARIN community, and it reflects their perceived needs. But there is some -- I mean we really should have some pushback that says that numbers are numbers, and you should get the kind of numbers you need when they're available -- if they're available. And from an education and outreach perspective, I'm going to have to parse your analogy about the impending cliff that we're approaching the bottom of. I'm not quite sure I understand that. But if there is a lack of particular resource because there isn't any more available, we should let people know about that. If there is some guidance about IPv6 being functionally equivalent for all intents and purposes, maybe we should tell people that, and help people understand what those equivalences are instead of them coming up and saying, "No, no, no, I really have to have, you know, IP version 2. I can't use IP version 4. I've got to have version 2," or "I got to have version 4. I can't use v6." Well, why not? Right? So there is some education component there. So I'm a little -- again, I'm hesitant about encouraging or discouraging particular technologies, and ARIN as an organization should really act as an enabler.
MR. CURRAN: Mr. Vixie, I know that we've heard the discussion with cliff, and actually I've heard you refer to the hard landing versus the soft landing. And I'm having a funny -- I have a suspicion we're talking about a similar circumstance. Would you like to comment?
MR. VIXIE: So stewardship over Internet Protocol Resources -- check the mission statement behind me on the wall. "Internet Protocol Resources" means among other things, address space. It is the job of ARIN to make sure that our members have enough address space to fulfill their chosen missions. But that is not the only Internet protocol resource that our members need. It happens to be the only one that ARIN is in charge of handing out, but it is not the only one you need. What we need as members of the Internet community is enough space in our router to hold all the entries. And so 10 years ago, 15 years ago, when people started talking about classless addressing versus classful, and we started trying to decide whether filtering division ought to be in, whether we should be handing /19s out as the minimum allocation and so on, what we found is that everything we did try to conserve, routing table slots simply made people who wanted to use those routing table slots for traffic engineering to fib about their needs and waste address space.
SPEAKER: Could you use another term?
SPEAKER: Fib isn't the word.
MR. MANNING: "Fib" is the wrong term to use.
MR. VIXIE: Well, I'm using a double entendre on purpose --
MR. MANNING: Oh, well, okay.
MR. VIXIE: Okay. So correspondingly, things that people want to do to conserve address space ends up burning through routing table slots. You know, gee, if you carve things up into /28s then, you know, by God, your utilization percentage will go up which are asking every other network operator both in our region and in the world to, you know, potentially consume hundreds of entries to represent your network when they could represent -- do with just a few. So Bill was talking a moment ago about enabling, and I agree I don't want ARIN to become a prescriptive organization. I want to enable what it says here, the advancement of the Internet. And I know that some of the things ARIN does -- education and outreach, function as enablers for that kind of advancement. I know that the thing Randy Bush is working on, the router PKI initiative, is going to help with that. As we head toward what Bill number 1 called the "wall that we're pedaling toward" or what John referred to as the "soft landing versus the hard landing," you know, when address space gets short, gets into really short supply, and when Mr. IANA over there hasn't got anymore to give us, we can expect massive de-aggregation of the addresses people are able to get. We can expect massive piracy, and if we can somehow enable a secure PGP industry to sort of prevent that kind of thing from happening, that will soften the landing. So I would like to enable a softer landing than the one we're headed for, because Bill number 1 is quite right. We are accelerating toward the wall, and as more and more people understand the nature of the coming scarcity, we're going to see more hoarding, more hoarding of address space, more hoarding of routing table slots. Lot of people are going to want to say, "I got here first, I got mine, you can go pound sand." I would like to enable a softer landing than what we're going to get if we simply continue as we're doing. And it's very clear that presenting IPv6 and saying, "Here it is, come and get it" is not enough. And nothing that we've done to sort of set up fee structures has changed anything about peoples' business costs for the transition. So I don't have a silver bullet to yank out and show you all, but I can tell you ARIN will not have achieved its goal of stewardship if we get the crash landing that we're headed for right now.
MR. CURRAN: Okay. You mention the phrase "hoarding." And I guess if there is a -- if the difficulty, effort, complexity of applying for address space remains constant, but the desire to get additional address blocks increases dramatically with the same amount of pressure or effort to get a block, and meant people much more anxious to do so, it's anticipated that we'll have more people applying, or people applying more frequently. Should the pressure, should the effort required to get an address block remain constant until we get to know there are no more left, or should it change over time? Who would like to in the Board -- I got -- I have one, two, three -- who wants to speak on that matter?
MR. BRADNER: I was going to speak actually continuing on this previous topic a little bit. And then I can talk to this.
MR. CURRAN: Okay. Why don't we have the people who asked, which is Ray and Scott speaking about the topic at hand, and then we'll move on to hoarding? Go ahead, Scott.
MR. BRADNER: We're dancing around where we talked about yesterday, which is not the soft landing versus hard landing; it's the panic versus non-panic, the feeling of scarcity, whether there's scarcity or not. So Geoff's -- the pretty picture we saw of Geoff -- from Geoff Houston is all well and good, but it ignores those people involved, ignores those people involved that are starting -- that will start to panic that they can't get the resources they want. And without keeping that in mind, I think we're looking at the wrong picture. We're not looking at something that's 2012 or something like that. Paul mentioned that was closer, that was closer than that, a lot closer than that, based on what the people who are in the business think their ability to get resources are. And then on to your question of should that ability be infringed upon by hardening the process, I don't think we need to harden the process by changing our rules. I think the process will be harder for those that are panicking and trying to get resources they don't -- they can't justify, simply because the ARIN staff will continue to enforce the same set of rules which will make it harder when they -- when the folks that are coming in and exaggerating their requirements or making up new requirements. I mean I don't -- I think that we don't have to do anything because we're already doing it.
MR. CURRAN: Okay. Ray.
MR. PLZAK: Slightly different thing here, and it actually goes back again to some of the things that we discussed yesterday, but --
SPEAKER: Ray, could you get a little --
MR. PLZAK: Yeah for one, was to look at the figure in terms of /8s of v4 space that have been allocated either by the RIRs or by the -- as legacy address space and compare that total number with the number of /8s or v4 space that are visible, i.e., perceived to be routed on the Internet, you would find a rather nice difference. And there's some that would point that, "Hey, that rather nice difference is stuff that people are sitting on or don't know they have or are waiting to do something with." And so that in and of itself could be viewed as being a part of the fuel for this secondary black gray whatever color market you have.
MR. CURRAN: Right. I don't want to rewind to the legacy panels. I would like to keep us focused on this topic not including the legacy space, though obviously there is a lot of interaction depending on the direction we take. Anyone want to comment on Scott or Ray? Okay, go ahead, Bill.
MR. WOODCOCK: So following on what Ray was saying, I think we all have some at least implicit belief that as things get scarce and certainly after we are out, there'll be some degree of monetization, right? I think many of us are aware that there are companies which are bought and sold which have access to address space, and that this is often treated as a part of the value of what's being bought and sold, unfortunate as that may be under our current policy and way of looking at things. And I think that that trend will probably continue as things get tighter. I think that that's inevitable. And I think that that will tend to squeeze out a lot of the space that Ray mentions. I think we are seeing this already. We are seeing space which was previously unannounced becoming announced. I think that is taking up part of the slack in the increasing rate, right? I think that part of the reason that we're not seeing the rate of RIR to LIR allocations increase as much as they might, is because we're also simultaneously seeing an increase in the rate of announcement of space which was previously unannounced but already allocated. So there was some argument yesterday about whether there should be more vigorous reclamation efforts on the part of the RIRs. And ultimately I think that's irrelevant, because ultimately the reclamation will happen one way or other. It doesn't need to be the RIRs sitting in the middle of it, right? The RIRs could, you know, have policy applying to that that, you know, they can implement if the address space we're going through them, and they could certainly facilitate aggregation which would be a wonderful thing. But again, something Alain pointed out, if the RIRs get nothing left; there'll be no way for them to even do that, right? If there's no pool sitting still in the RIRs, there'll be no way for the RIRs to facilitate aggregation of a bunch of little blocks coming in and putting one big block out. So I think we need to not run all the way out to the end. I think that's not the only concern, right? The other concern is every RIR has research allocation policy, one which, you know, my organization has taken advantage of, and one which many other organizations have taken advantage of. And you know, you could think about whether that policy will go away as we run out. I think, you know, we need to try and figure out whether our priority is to maintain the RIRs control over the way in which organizations get blocks or maintain the stewardship goals, because I think those will start to diverge as things get toward the end.
MR. CURRAN: Okay. Mr. Vixie, would you like to respond?
MR. VIXIE: I wanted to answer your question about the hardening of policies. So if we're not on that yet, you can come back to me.
MR. CURRAN: Okay. You can go ahead.
MR. VIXIE: So we want to transition, because we pretty much think that v6 is what's going to save us after we make either a hard or soft landing in v4. And various people have agreed and disagreed about whether v6 would ever come, or whether it was a one-year, or five-year, or fifty-year timeframe. But you inspired an interesting line of thought with your question about hardening of policies, you know, because I know that one way we could harden policies would be to require a Big 8 accounting firm to go do an audit of an applicant to make sure they really have the net worth they're describing on the paperwork because, you know, the ARIN staff is the best in the business because there's only so much you can do with the pile of paperwork that's been submitted to you. But I was thinking what if we had some kind of a policy requiring the new network builds be dual stack.
MR. CURRAN: You're -- new -- you're saying a policy that says new network builds must be dual stack. Now, since we don't license network builds these -- unless I --
MR. VIXIE: I'm sorry --
MR. MANNING: You were sleeping John.
MR. CURRAN: Unless the charter changed, or I wasn't looking.
MR. VIXIE: I'm sorry. Let me reformulate that.
MR. CURRAN: I was thinking you were saying maybe that if someone wants to build a v4 network and looks for addresses for it that we tell them you have to take v6 addresses?
SPEAKER: Not just take them but use them, right?
MR. VIXIE: Not just take them but use them. I mean, we tried to validate, you know, we ask for evidence about their utilization percentage. We ask for their build out plans. Could we start asking for their v6 build out plans to help justify v4 address blocks because it is irresponsible for any member of the Internet community whether in ARIN's region or elsewhere to do a new v4 only build of significant size in 2007.
MR. CURRAN: Okay. I got Lee, Bill, Bill, and Scott who all want to comment. Let's go with Lee.
MR. HOWARD: So should we also be requesting their router configurations in address zone files?
MR. VIXIE: Only if we have reason to suspect fraud, which I think we will have reason to suspect as we get closer to the hard landing that I am predicting.
MR. HOWARD: That would be a yes then. Okay.
MR. CURRAN: Bill.
MR. MANNING: Okay. So I'm the guy that was after Scott and Ray.
MR. CURRAN: Right. So speak and then speak.
MR. MANNING: Okay. Then I'll speak and then speak.
MR. CURRAN: One little point of order. In case I haven't said this yet. The microphones are actually open. And people who would like to speak -- they really are open. (Laughter)
MR. CURRAN: People who would like to skewer their favorite board member on -- with a question on topic should find a microphone at some point.
SPEAKER: Wow, I guess, you hadn't said that --
MR. CURRAN: Yeah, continue.
MR. MANNING: So, I'm going to speak first and then you get to skewer these other guys. I guess there are two things. One of them is that unlike consumables like oil, IP addresses are a close system. They can be recycled and reused relatively easily. One of the discussions in the legacy space is how do we reclaim these things when there was no mechanism when they are handed out. And I think we've got that in hand for the v6 space certainly and for most of the v4 space. But we don't seem to have reasonably good track records about large numbers of people taking advantage of this recycling system which ought to be reasonable. Now, let me touch on the other bit that I think is interesting with regards to coding if I can jump into that cesspool. Leslie, what is the minimum allocation size that we hand out?
MR. NOBILE: For what?
MR. MANNING: In V4 space today.
MR. MANNING: A /20.
SPEAKER: Unless you're multi-homing.
MR. MANNING: Some number, right?
MR. NOBILE: The minimum allocation size is /20 unless you qualify under the multi-home policy which is 21, 22. So it's various.
MR. MANNING: Right. Somewhere in there.
MR. NOBILE: Yeah.
MR. MANNING: And so I'm going to say that I qualify under multi-homing policy because I've got ASNs and I've got a couple of upstream ISPs. And I'm actually only really going to -- I'm really only going to use five of those addresses. So does that mean that ARIN is forcing me to hoard the rest of that /20, whatever it is?
MR. CURRAN: It's a good question.
MR. MANNING: In some sense we encourage hoarding by having allocation policies which favor global routing.
MR. CURRAN: Okay. I agree. Bill, you also wanted to comment on, I believe, Paul Vixie's licensing networks comment.
MR. WOODCOCK: Yeah, well, I don't think it's an issue of licensing networks. I think that right now we have in the research request process we have the requirement that you explain whether your request involves a renumbering out of some other space and a plan for how you're going to do that. I think that's probably the mechanism which could be extended, right? That -- a possibility would be a policy proposal which said that new V4 requests include an explanation of whether there would be renumbering from v4 to v6, and if so how? And if not, why not? That's a possibility.
MR. CURRAN: Mr. Bradner, you wanted to comment as well. And then we'll turn to the mics.
MR. BRADNER: On the question of asking something along the line of have you stopped meeting your V6 wife -- (Laughter)
MR. BRADNER: -- then when are you going to show us your plans for deploying V6 in order to get v4? I like Steve Ryan. I don't want to keep paying him too much money though -- speaks to me of some significant anti- trust issues that I too wouldn't want to be on the same side of the universe of.
MR. CURRAN: Okay. I'm going to open it up to the microphones. Front and center mic. Jordi.
MR. PALET MARTINEZ: Jordi Palet.
MR. CURRAN: You have to tell me whether you like the question directed to a board member or -- one for me to choose.
MR. PALET MARTINEZ: Well, actually, most of them are somehow comments and questions regarding different things that have been commented. I don't remember to whom each one. So I am --
MR. CURRAN: Try to keep it to one or two quickly. So other people can speak.
MR. PALET MARTINEZ: Okay. So I go for a couple of them now, and then come back. The first question is I really like the idea of this possible policy, I guess, about asking, if you come to IPv4 space, you need to get also IPv6. I'm not sure if we need to actually check if that's being used or not, maybe not immediately, but in certain period of time. And I think it can be something very, very similar to what we have for the 4 byte ASN. I think that could give us some hint about what to do. Then another point is. The people is usually forgetting that moving to IPv6 it may have a cost if you want to do it overnight. But typical if you plan ahead it's not that much costly. And actually the cost is already happening. I have made a presentation. I am working in a paper which is the cost of not deploying IPv6. And, especially, since Vista had been released I am already monitoring an increase on traffic, on IPv6 traffic, which clearly shows that things are happening, and that cost of not deploying IPv6 is already measurable in networks of ISPs. I think it's important to tell the people. Hey, pay attention because maybe you are paying more transit cost or whatever. Not using transition mechanism, I am not talking about just native IPv6 deployment.
MR. CURRAN: Okay. Interesting comments. Does any board member wish to respond about his comments, or regarding a don't ask, don't tell policy for checking up on v6 multi-homing?
MR. CURRAN: Okay. So we got Bill, Lee, Scott. Bill.
MR. MANNING: Giving people things they don't want or feel that they don't need is recipe for waste.
MR. HOWARD: In response to Jordi's assertion that the cost is negligible if done. If applied overtime. I'm very curious about what the time frame is over which it must be spread in order to become negligible, and if we have that much.
MR. CURRAN: Okay. Scott.
MR. BRADNER: I'm looking very forward, very much to your paper on the cost of not deploying. I think that could be a very helpful thing, and I hope that gets so good distribution because that's exactly the sort of insertion of facts into the discussion in a discussion where a fact is sometimes scarce will be helpful. Thanks.
MR. CURRAN: Okay. Rear microphone, yes.
MR. DARTE: Yeah, I'm Bill Darte. I'm with the ARIN AC, represent Washington University in St. Louis. The -- yeah, it seems to me that the whole issue of addresses in the future are the haves and have nots. The people that have them don't need IPv6 because there is no incentive to get them because they already got them. The people that don't have them are the ones who might be motivated to get IPv6 because that might be the way they get connectivity unless they have a black market to go to, to get things. And then I think that's disincented because the routing table issue growing. I think the -- they could be driven to v6. And if they are, and want to go that way then that's going to put pressure on providers to offer those services and so there is going to be a pressure to move in that direction. And to the extent that they're deploying services that other people would like to get to and maybe it causes people with v4 only to want to go that way too. And then the other thing is the market itself. If there were a way of precluding that market or diminishing that Market from becoming robust I think that that would be something that that ARIN might contribute to.
MR. CURRAN: Could you clarify, precluding that market, you're speaking about a black market, or a market for address space?
MR. DARTE: Yeah, market for address space. So the question becomes is it within the purview of ARIN itself to say, no we don't think you can do that? And the first person that does it will assume, or if that's not the appropriate pursuit then perhaps, you know, it's a legislative thing. You know, you might go to a legislator -- legislature and say, lot of bad things will happen if we allow the emergence of a market like that when we think this is a public good or something like that.
MR. CURRAN: ARIN would have to do lot of legislatures. Any board comments? Yes, Scott.
MR. BRADNER: Just one, Bill. Your talking about people being driven to IPv6 reminds me of the movie Driving Miss Daisy, where Ms. Daisy is walking down the street, and the cars sort of passing along behind her. That's the kind of driven I've been seeing.
MR. CURRAN: Okay. Far microphone.
MR. GODAS: Hi, name's Joe Godas. I'm Cablevision --
MR. CURRAN: I'm sorry speak --
SPEAKER: You got to like the microphone --
MR. GODAS: Joe Godas from Cablevision, northeast cable operator. I will disclaim this by saying I've been revved by my colleagues. The topic I'm going to bring up is the cerebral equivalent of rearranging deck chairs on the Hindenburg. But I'm going to forge forward. Basically, the implication is a far-reaching financial impact for us as an industry. So that's why I have to really put it out there for the panel and see what you think. The issue is, we have video services that are on set-top boxes. And we have RFC 1918 to help us with these kinds of issues, right. But, you know, that space is rapidly -- being depleted, and the natural thing to do is go to ARIN for help, and say, hey, you know, help us with this address problem which you're able to accommodate nicely through the guidelines of how addresses are allocated. The real question here is, if the intention is not to route the space and, you know, can't the request that are going to start coming in more repetitively from other operators can't they be satisfied in the same ilk of perhaps this is the trade of thinking side, allocating the same space under the auspices of not routing it and saying hey, we need it back in five years. In other words buy the industry time to solve its layer 7 issues of what's going on. You know, the acceleration of IPv4 depletion isn't necessarily happening because the Internet is expanding. Maybe it's happening because it's leaking. And, you know, we're just putting nails on the coffin a little sooner than maybe perhaps it needs to be.
MR. CURRAN: Can I just make sure I understand. You are wondering about the Board's views on a policy or activity to address might be the depletion of RFC 1918 space used by cable operators and others.
MR. GODAS: Yeah. And what's happening is that another RFC 1918 is happening via happenstance, and not just once, but perhaps several times via, you know, the same policies that are there are actually, you know, help users connect to each other. Its destructive compliance to an ARIN policy, but it's not helping anybody.
MR. CURRAN: Okay. Board member comments on, potentially address depletion issue in the non-global uniqueness space. Yes, Bill.
MR. MANNING: Okay. So several years ago when there was a discussion about 1918 space I had another bet with an individual that say that, what we ought to do given the number of (off mike) relationships on a global scale we've got to invert RFC 1918 space and make 1918 space the globally routed address space for BGP peering sessions and give everybody else all of the rest of the address space. All right. That means you effectively get all of the IPv4 space for doing your set-top boxed. The problem you're going to run into, eventually, even if you adopt. Even if we were to magically adopt that and everyone was to agree to that is that you're going to eventually run out of v4 space period, and you're going to have to do something else eventually.
MR. GODAS: Absolutely. I'm talking about deferring the moment of truth for a little while long because the people we have to talk to solve the application problems I don't think it'll be ready in three years we need more time than that. What we're doing is --
MR. WOODCOCK: You don't have three years and deferring it a little bit further is few more months.
MR. MANNING: You might have a few more months. I'm actually thinking that ARIN isn't really sort of a drug pusher, and I'm not sure that we should encourage addiction.
SPEAKER: So what's --
MR. GODAS: Everyone is going to suffer though. But I understand that's why I disclaim the rearranging deck chairs thing, you know, it will --
MR. CURRAN: It's an interesting idea. I think that the idea has been heard, and there is a question about whether that idea is an idea for this forum to solve potentially, or another forum. Somewhere there is a guy who used to be up here but now those are IANA stuff. And he is probably one of the guys to talk to first because they might. You might -- the global RFC 1918 space is -- it's something that's going to take multiple coordination among multiple bodies, including the IETF and the IANA. So I would look to the back for a guy with a smile on his face. But we definitely heard the request. Front center mic.
MR. ECHEBERRIA: Thank you. Actually, I have 17 comments.
MR. CURRAN: 17 comments. Perhaps you and Jordi can get together and pair them up. Let hear the first two. (Laughter)
MR. ECHEBERRIA: I will try to concentrate in those I will leave some that I'm sure that other people will address.
MR. CURRAN: What's you're name, affiliation?
MR. ECHEBERRIA: I was in front of -- okay. My name is Raúl Echeberría. I'm the CEO of LACNIC.
MR. CURRAN: For the people out there in audio land you need to say who you are.
MR. ECHEBERRIA: Yeah, okay, thank you. Sorry. I think that it is interesting to realize that if we consider all the -- an allocate IPv4 space and the 40 percent of the already allocated IPv4 addresses are not being used. We have more than one third of the total IPv4 addresses. It's not that too bad a situation. If we compare with the oil market for example it's the same situation of the oil. So it is -- people think that we have use -- we have already used two-third of the available oil in the world. So -- I think that's -- we can do many things with IPv4 if we do the things -- the right things. I think that we should work in recovering this space as this is something that we are doing in every area. And I think that we have still too much work to do together with other like IANA. I think that's -- what is very important is to resolve the issue of the last /8. This is too much pressure at this moment. So this discussion is being held in many forums that is one proposal from (off mike) people. Another proposal has been submitted to today to the LACNIC policy list. I think that this is something that we have solved in a short period. So because there is a paradox today that -- it's that they ask more IPv4 address, and RIR allocate more IPv4 the RIR can get from an allocated pool. So probably we'll solve this problem in a short time with the RIR's community. The public forums will feel more relaxed in order to wait for example, more concern about the policies. The promotion of IPv6 is also a point. I think everybody has a role in this point, the government, private sector as Ray addressed it before. This idea will bring money in the point that the -- this kind of organizations cannot use waiver so another incentives for promoting a technology, but -- because we are talking about a very specific case. This is something that doesn't happen every year. It's a very specific case. And what we have to do in order to benefit the community is to promote the idea the whole community has to be ready for IPv6. It doesn't mean that we are promoting the idea that IPv6 should be used in a short time over IPv4. So what we are doing is something that is in benefit of the community. And if we use those tools in benefit of the community it is very good. Last point -- only three comments.
MR. CURRAN: Yes.
MR. ECHEBERRIA: We have to take care of developing countries as they -- this IPv4 consumption of course have negative impacts or could have potentially negative impacts. We have to take care that negative impacts are not reared in developing regions that (off mike).
MR. CURRAN: Okay. Thank you. Any Board members wish to comment? No, no, no, no. Okay. I'm going to move to the microphone in the back. Yes.
MR. DURAND: Alain Durand. First of all I would like to congratulate you and thank you for having this panel today. I think it's very important not have the head in the sand anymore and talk about the end gain on IPv4. Second -- a number of things are going to happen, but -- and you would like them or you would not like them but this monitorization is really going to happen in a big scale and this is going to be big numbers. We're not going to talk about dollars and cents, but several orders of magnitude more than that. So the question really in my mind is how to ease this transition and to mitigate the impact of this transition to the network, and what is going to happen to this community because if this community doesn't want to manage it this is going to be managed outside of this community. And in the end it could just hurt much more. I'm very sensitive to this notion that people are going to have to but wholesale blocks and resell them to whoever wants to but them at the highest price, maybe on eBay. So people are going to probably have smaller blocks and smaller blocks and larger number of (off mike) networks that are going to be advertised and paid by everybody. So how to do the garbage collection of those routes is very important topic. And my opinion is that ARIN should have a mission to ease this process. And if you could simply have -- you know, at least have a large enough address space where people can say, oh, I have small blocks -- and I would like to go and trade them for one unique block that will have tremendous impact on that (off mike) of the Internet at large. But this is only possible if there is enough block in reserve, okay, if we run out of blocks this will not be possible to be.
MR. CURRAN: Okay. So you're saying ARIN should be looking at exploring facilitating not only the market, but facilitating the ability to exchange and get beneficial use of all the address space even the small pieces that might be hidden somewhere with no place to go to aggregate them.
MR. DURAND: Actually, I'm much more worried about the small pieces than the big fragments. Somebody has a /12, want to advertise a /12 and a /13 that's fine. But if you have 100 out of 500 /22, /21 this is going to be a big issue.
MR. CURRAN: Understood. I am going to close the microphone soon. Any Board -- Ray, you want to respond to the last comment?
MR. HOWARD: Yes. That's very real issue. And I don't think we're putting our head in the sand when saying that maybe we are discussing monetization right now or whatever. There is two things that to bear in mind here. One is, we have taken a position that IP addresses are not property, and it certainly has an influence on how we would deal with any kind of a market situation. On the other hand, if you have a market that evolves and as, for example we move towards a certificate system where the attestation of a route object, IP address block to a particular organization becomes much more critical and we stand as the person that acts in a certain sense, now, like a title company. People could be buying useless address space on some kind of an open market. So there is a dichotomy right there as far as us being the ones that attest to the uniqueness of address space. So then that kind of begs the issue to -- does the community look at what they can do with the transfer process that somehow rather would maybe recognize the market? But then at the same time help distribution of the address space. So I just throw that out as food for thought.
MR. CURRAN: Okay. Thank you. I want to go to the front microphone.
MR. DeLONG: Owen DeLong, Jittr Networks. I will point out that ARIN recently won pretty resoundingly a lawsuit on the basis of the fact that IP addresses are not property. If we start allowing for, and subscribing to the theory or monitorization of addresses we have basically just contradicted ourselves on that theory. I think it's a really, really bad idea. Now --
MR. CURRAN: -- let me just -- I want to clarify. I'm going to direct your summary to Counsel on the basis on which we won because while there is some support for ARIN's position it may not be quite what you stated.
MR. DeLONG: Sure.
SPEAKER: The second seat. Okay. I think he woke up.
MR. CURRAN: Microphone one.
SPEAKER: Try again.
MR. RYAN: I think it would be more accurate to say that what the court ordered was that Mr. Kremen comply with our procedures. The court made no definitive finding with regard to the nature of the arguments that he made with regard to whether it was a property or a service, the court didn't reach that because the court ordered him to follow our procedure. So in that sense the court recognized ARIN as a legitimate holder of the role of dispensing the resources and ordered Kremen to comply with those procedures. But it isn't -- I apologize if I left the impression that it was as definitive as that. And the court never reached that ultimate issue because it didn't need to.
MR. DeLong: Okay. I stand corrected. In any case looking at this, the whole theory that ARIN has operated under so far is addresses aren't property. And I think we need to stick with that for a lot of reasons including just generally the good of the community. ARIN is only authoritative as a registry, and so far as the operators that actually implement routing recognize it as such. It has always been that way and no matter no ARIN does or says it will pretty much remain that way because ARIN doesn't control the routing tables. ARIN admits that as a matter of policy and it's sort of a good thing. It acts as a check and balance to a certain extent. If ARIN stops being the registry that the people who run routers recognize as authoritative life gets very interesting. And I don't necessarily look forward to that world. But basically there is very little we can do. Finally, last comment, handing out v6 addresses to people who may or may not want them just because we feel like handing out v6 addresses, not a good idea.
MR. CURRAN: Okay, any Board member comments?
MR. HOWARD: I do think that ARIN has slightly more charter of credibility than simply consensus with the operators' community. Certainly if we lost that then we would be in dire straits. But there is some origin to ARIN that is not simply driving that.
MR. CURRAN: Okay, back microphone, yes?
MR. WILSON: You have Paul Wilson from APNIC. I wanted to raise a possibility that I've heard about mentioned casually for years, and that's the possibility of using the Class-E space 240/4. It seems to me that if it's going to be done, it should be done soon. There is a -- sort of a small window of opportunity there, and it seems it could be used in a couple of ways. One is public unicast, which cannot be a big "ask" in terms of upgrading everything, but the other one is coming back to what was mentioned earlier, the use of that space for private use. And seems to me that if in the case of private operator with their own private network wanted to use this address space within their network then they got a much better chance of knowing what their infrastructure is and being able to upgrade that infrastructure and know what the cost in the plan is going to have to be to upgrade a private network. The one reason to really look at it seriously is because of this issue, depletion of 1918. We had discussions in the APNIC regions last year with, about an operator who wanted to deploy -- this large Telco who wanted to create a national VOIP network, and predicted they'd be after 3/8s within a couple of years and 8/8s within three or four years. They were quite prepared apparently, to use a private space for that. We've but of course couldn't, we've already got one operator, one VOIP network in the Asia-Pacific, which has got substantially more than a /7 for their VoIP network and that's presumably one which could also potentially use private space. So the -- and RFC to re- designate that Class-E space doesn't seem to be made of the rocket science, and upgrades by router vendors and PC manufactures are a very frequent thing these days, and the upgrade required to enable, use of that space is pretty trivial.
MR. CURRAN: Okay, so two question, everyone wants Class-E address space, Scott, do you want to comment?
MR. BRADNER: What I thought I heard you say was we should start using Class-E address space and like 127 and things like that.
MR. WILSON: Class-E is 240/4 is the 16/8ths at the top of the --
MR. BRADNER: Yeah, we looked at that as part of the v6 effort and we were told by vendors that it was pretty deep-buried into their codes and rather many places that this stuff was, this was stuff, that address space is poison and it would take a lot of work to dig out all the places that was --
MR. CURRAN: Question?
SPEAKER: We got two other comments for Bill.
MR. MANNING: Mirroring Scott's work there was a group of addresses long time ago called Martians which were the boundary blocks around the Class 4 address space, so any way so that particular set of blocks was also hard coded as was the E space in vendors code, and when we went to CIDR it took about three years and aggressive vendor buying to recode almost everything to get those Martians eliminated. We have substantially more equipment in the field now, I don't think the E space is going to be useful, I really don't. Even if we declare it to be useful the code won't support it.
MR. CURRAN: Maybe you want to comment, Paul?
MR. VIXIE: Yeah, I've a question for Scott, rhetorical question and then a rhetorical remark for Bill No. 2. So, is it more less effort to recode those devices than to deploy these six, and for Bill No. 2 I want to point out that since we are, at least Mr. Wilson is contemplating using the space for private use, it strikes me that it would be up to an ISP to order their own devices, which is a different problem than trying to reassign those Martians where you have the open community who must all recode their devices, it's a smaller footprint.
MR. MANNING: Recoding my LinkSys boxes only works on a couple of LinkSys Boxes.
MR. CURRAN: Okay, I want to move on briskly. Front microphone?
MR. MAEMURA: Okay, my name is Akinori Maemura, from JPNIC and we are one of the authors of the IPv4 countdown policy proposal. And first of all I thank you very much for the facilitation of this -- the discussion allying with our proposal. That's really great, thank you very much. And I have just one comment so -- through this discussion of the our specific proposal we -- there is a big discussion about the reclamation or, market mechanism for the second hand IPv4 addresses and the thanks to Geoff for and Tony's research about the IPv4 address lifetime. We have a very fine projection of the when IPv4 address, the IANA stock would be depreciated. Then while the -- what we are not so clear is how much the reclaimed IP address space will be available after the IANA stock has run out. So, that would be one question to the Board. If you have any idea for the estimate, how much space of the reclaimed IP address and what is the pace of the supply of such kind of secondhand IP address, I would like to have that from you.
MR. CURRAN: I want to make sure I understood. I believe you're asking about the once the IANA ability to handout is expired, how long the registry, thinks it can continue assigning address space?
MR. MAEMURA: Exactly.
MR. CURRAN: Okay.
MR. MAEMURA: How long and how much can, how much reused, reclaimed IP address will be available for the IPv4s.
MR. CURRAN: Right, would any Board Member like to comment?
MR. BRADNER: Well, wasn't that Geoff's picture that was shown, the blue and green line, the red line?
MR. MAEMURA: No, that is the just the new IP before this, but I mean that --
MR. CURRAN: It's not counting reclamation?
MR. MAEMURA: Yes, right.
MR. CURRAN: Yes, right?
MR. PLZAK: Excuse me, the -- you know, the fact is that ARIN would probably never run out of IPv4 address space, the point is that how IPv4 address space would they have that would be aggregable enough, to be useful. And so it would seem to me that as the garbage collection occurs, is that there could be some aggregation, and you could pull back certain amounts of address space but certainly what the RIRs are going to lose the ability to do, is to allocate address space against current policy minimum sizes. And so going back to an earlier comment that you made about hardening policies and stuff like that, there's also the conception that perhaps you changed what the minimum size is that you would allocate as well, because the realism is that we all have holes and stuff in the address space now. And it has occurred for a couple of different reasons but the fact is that it does exist. So it will always be there and then how much would be there to continue to do something in a meaningful manner is the other point.
MR. CURRAN: So the answer is there's -- they may not run out even after the IANA stops giving address blocks, there will always be some address space, it just may not be allocatable in any of our current policies or the polices need to change in order to be able to harvest those little blocks?
MR. MAEMURA: Right.
MR. CURRAN: Okay?
MR. MAEMURA: Okay, thanks.
MR. CURRAN: Far microphone?
MR. SCHILLER: Jason Schiller, Verizon Business/UUNET. The question for the Board is generically, we've talked about, after exhausting that there is going to be a market for resale of IP addresses. So, the question is, if there is going to be a black market anyway is it preferable to try and have a legal market for that? And, if you believe there should be a legal market for it what happens to IP stewardship? If somebody can afford to buy an IP address, can they simply buy it or should we still try to enforce the things that are important to stewardship such as, efficient use and things like that?
MR. CURRAN: Okay, so the Board, the question is posed, if the Board heads in the direction of facilitating, if this organization moves towards facilitating an efficient market, do markets exercise stewardship? (Laughter)
MR. CURRAN: I'm asking, only the Board members wish to comment?
MR. HOWARD: I think we don't have sufficient information yet. I think that we are still looking at the question, and there are major issues that have already been raised about whether we want to determine the monetization of IP addresses, whether there are trust issues that are problematic for us and, you know, whether it would work efficiently. So, I don't think we can answer that yet but, obviously the conversation is happening here.
MR. CURRAN: Scott?
MR. BRADNER: I don't think it is quite the right question, because it's not a question of simply, can we have an efficient l market in IP addresses? We can certainly do that, there's certainly many ways to do that. But, how can you have an efficient market in IP addresses and not have $17 million routers? We've got to; it's a two-factor situation, I agree with Lee that we do not have enough information on knowing how to solve that two- factored equation.
MR. CURRAN: Microphone over here, number one?
MR. SCHILLER: Yeah, just to follow up, is there specific information that the community can provide you to help you make this decision?
MR. BRADNER: I think time will tell.
MR. CURRAN: I hope we have enough time for time to tell, thank you.
MR. SCHILLER: Thank you.
MR. CURRAN: So, I'm going to go actually to a comment from our remote participants, there are remote participants out there, who have seen snippets of this show as its broadcast. Ray, could you read the comments from the participant?
MR. PLZAK: Yes, this is from Tony Hain, just as a parenthetical comment. The connectivity upstream from here is kind of choppy so he sent this message to PPML and what Tony says is, "In any case the discussion about buying time to prepare for IPv6 is just a stalling tactic that will only make the eventual effort that much more painful. It is absolutely understandable, that people want to have to more time to prepare. But to do that they should have started earlier. Making the current processes more painful for newcomers just so the lethargic can feel better, is not stewardship." (Laughter)
MR. CURRAN: Paraphrased -- (Laughter)
MR. CURRAN: -- those of you who moved quickly towards the lifeboats are in far better shape. Now, Tony's comment should be taken to indicate there is a concern that the efforts to prolong or extend IPv4 address space used for life, do tap in many cases the same resources that are needed to help facilitate IPv6 transition, and there is not an infinite supply of resources. If you have to do few things, its going to prolong the time it takes to get them both done. I am going to be closing the microphones shortly. So, I'd would like to have people move to mics if you want to ask the Board a question on this. Back microphone, Mr. Conrad?
MR. CONRAD: There are actually two comments. One, with regards to the 1918, additional 1918 space, there have been several efforts that I'm aware of to actually work on a draft proposed to the IETF to allocate additional 1918 space. As far as I know, none of them are currently moving forward. There is a, I think, the past experiences in attempting to move additional 1918 space, into the registries has been somewhat controversial.
MR. BRADNER: But it's been consistent.
MR. CONRAD: Yes, it has been very consistent and I think people got tired of hitting their heads against walls. I would make the observation that if you look at RFC 1918 space as address space that is used internally within a network that you don't have any assurances about it being routed outside of the network, there is this brand new address space that could be used called IPv6 for the same purpose, but that might be seen as a bit cynical. (Laughter)
MR. CONRAD: The other issues, with regards to what Maemura was raising with regards to, the amount of address space that would be remaining when IANA runs out of its pool, I believe, and folks might be able to correct me if I'm wrong, that the projections that Geoff has been doing have been based on historical allocation patterns. In particular, the size of the blocs that have been allocated were derived off of historical block allocations. The mechanism, the amount of address space, IANA allocates to their regional registries has recently changed as an operational consideration. So, there is a much more granular allocation. What this will mean is when IANA runs out of its free pool, the amount of free space that has been allocated to the RIRs will actually be significant less, significantly less than it would have been under the previous allocation regime. That doesn't obviously address reclamation efforts, it's just a -- I suspect that the gap that you see in Geoff's slides between the red line and blue line, will actually be much, much tighter than it is right now.
MR. CURRAN: Good point. Any Board Members want to comment?
MR. BRADNER: Sounds rational.
MR. CURRAN: Yeah. I am closing the microphones, get to the microphone now. Microphones are closed. Yes, front center?
MR. BICKNELL: Leo Bicknell, Harrah's Entertainment. I actually think I'm probably in a minority in this room, in my situation in that for those of you who don't know I used to work for a large Tier-1, global ISP, pick-your-marketing-buzzword of the day. And now I work for Harrah's Entertainment, the largest gambling company in the world or so they claim, I don't know how they come up with that number. And --
MR. CURRAN: Is there a difference between these industries? (Laughter)
MR. BICKNELL: Well, there are a couple of similarities and a couple of interesting differences. So, the similarity is and both of those types of businesses, my experience is the people who control the money will delay spending it until the absolute last possible moment and usually a few days after the last possible moment. (Laughter)
MR. BICKNELL: So, any notion of, you know, we can buy more time, just buys them more time to sit on their money and deploy it late. But, there is an interesting difference and what I really wanted to speak about up here, is I think there is a segment of the ARIN constituency that is basically not represented at these meetings, and that is business like Harris. Typically, the people who run them, operate them, build network for them do not come to this party. And I have heard lots of people stand up and good discussion on things like routing table slots, and aggregation in the Global Internet. The reality is for us. We have 25,000 desktops internally. They are all v4 numbered, they are proxied to the Internet, they have mail gateways to the Internet, and there are no Internet routes internally. In fact, many networks don't even have a default route internally. We are not deploying Vista until minimum 2009, because that's the fist opportunity to put it in the budget.
MR. CURRAN: Amen.
MR. BICKNELL: We will not be running v6, probably until 2015 whether the Internet is or not, because I am sure we will be able to buy a proxy to the Internet that will speak one on the inside, and one on the outside and a mail gateway and so forth. And the reason this is a problem, in my mind a gigantic problem, is my perception of companies like Juniper and Cisco and all these other wonderful people. If they sell one port to our upstream provider for us to connect and right now I am managing almost 3,000 routers and switches. So, guess where the money is for those companies? They basically subsidize ISPs with equipment. I understand that you may spend $9 million on your gigantic router; I used to do that in my previous job. Out of Cisco's total revenue it's almost unimportant. So, if you really want to bring along v6 you got to find a way to bring along companies like Harris internally to get Vista deployed or other IPv6 capable OSs to get their routers and switches upgraded. And right now, it's just cost. That's all they see. It costs money to deploy all that. There is zero benefit because we don't need it internally, because we proxy it to the outside. And that's why, as much as I hate to say it, the only solution I've seen to this problem is syntax; people don't stop smoking even though it raises your insurance premium, shortens your life, we've known for 50 years it's bad for you. The only way we've gotten it stopped, passed laws that you can't smoke; raised the tax on cigarettes to $9 a carton, so it does it. And whether it's ARIN raising fees, or whether it's the U.S. government stepping in and saying for every IPv4 you use it's going to cost you $90 in tax, that's what's going to get people to convert, that's what's going to show up on their radar as a budget item that gives them something to offset. And the sad thing is the tax is going to have to be higher than the cost of conversion. So if it's going to cost Harris $10 million to replace all that stuff, it better be a $20 million tax or its still not going to happen.
MR. CURRAN: Okay, any Board member comments on the syntax?
MR. MANNING: Wow. (Laughter)
MR. CURRAN: Okay, any others?
MR. WOODCOCK: I'm afraid, I just don't get that. We are not talking about forcing anybody to convert from v4 that they are doing today. It's a question of growth, right? If you need for some reason, more internal addresses, which are routable to the outside directly, not through a proxy, then yes you would need v6.
MR. BICKNELL: You are missing my point. You're correct that there is no reason for people like us to upgrade today, and my point is that the vendors that we're all talking about not developing code for, 99.999 percent of their revenue comes from us.
MR. WOODCOCK: Uh-huh.
MR. BICKNELL: So, why would they ever give the ISPs the v6 features they need, when it's.001 percent of their revenue?. They have the same bean counters. So, you have to bring along people like to me so there is a business case for the people who developed Vista and IOS and JUNOS to have a business case to give it to the ISPs.
MR. WOODCOCK: What v6 features do ISP need that they don't have available to them?
MR. BICKNELL: I think there is a very recent discussion I don't know if that was on PPML or NANOG. I think it was NANOG about the utter lack of any useful dynamic attribute assignment, i.e. DHCPv6 working properly in --
MR. WOODCOCK: How is that an ISP issue?
MR. BICKNELL: ISP is assigned addresses that way all the time. My home cable modem works via DHCP, and right now with IPv6 I can't get name server addresses I can't get domain names, I can't get e-mail servers it's useless. As I said on the mailing list I turn on my box if doesn't work then I am going to unplug v6, and plug back in to v4, because at home I want to surf the web and play a game and not be a routing engineer. So, until somebody writes that software and deploys it in every router and in every network v6 is broken and useless.
MR. CURRAN: Okay, would other Board members comments? Scott, no?
MR. BRADNER: Just to say the fact that it's not implemented and deployed doesn't mean somebody didn't write it down as to how to do it, because it has been written down how to do it, and then willfully or carefully ignored.
MR. CURRAN: Uh-huh. Okay, back microphone?
MR. ALEXANDER: Dan Alexander, Comcast and ARIN AC, my apologies because the wording isn't entirely worked out in my head but Owen raised a point and others have that, really sticks with me in that we've spent an extreme amount of effort, and time and resources around continually stating IPs in our property and, you know, this is a resource for every one. But in the same breath, we continue to have these conversations. Well, there is going to be a black market and people are going to start selling them on eBay. So, it almost seems like their classic behavior of, everyone works, you know, together and under a common set of rules until there is a fire, then everyone tramples each other for the door. It almost seems like we're doing that before anyone really lit the match. You know, is -- at what point at an RIR level do or has anyone had the conversation or decided, you know what, we should take a stand or we should try and put the policies in place to, you know, to prevent this open market. You know, it was, the point was raised earlier securing PGP you know, we've had lot's of conservations about securing registration, records, things like that. Have, I guess my question to the to the Panel would be have RIRs as a whole gotten together and had any discussions around, are we going to, you know, take a stand and try and continue what we've been doing, you know, the methodology we've taken this whole time or we just going to throw in the towel, and let the fire run out of control?
MR. CURRAN: I am going to direct your question to Board with a slight redirect. The question for the Board is, has ARIN basically folded, given up without asserting, potentially taking the position that there won't be a market, and that we need to prevent that from occurring as opposed to the going on with a quiet acceptance that it has to occur inevitably? With respect to inter-RIR conversations on the same topic there is some interesting issues about that particular item, and I'm going to set those aside for the moment, but recognize that coordinated registry action is a difficult topic in this area when you are talking about policies that may prevent the allocation of addresses to people. So, for the Board though, has ARIN just given up and accepted that there is going to be a market and should we still facilitate it or is it possible to continue our position and decide that it's worth standing for IP and Ray?
MR. PLZAK: Excuse me I don't think that we have addressed it; I think it's the real answer. We have taken a -- one position in regards to the property in respect of numbers. We have said that, yes we know that there is one out there, and yes we say that one could exist, but we haven't really taken a position one-way or other about what we would really do with it. I threw out some food for thought before, but that wasn't a position, that was just some things to spur conversation.
MR. CURRAN: Okay, front micro -- sorry, far microphone, left side.
MR. GODAS: Yeah, I just want to respond to some of the responses that came in. I want to reiterate this isn't a function of being lethargic or burying our heads in the sand, of where v6 is coming, where it needs to be. There is no question as an ISP we are going to provide v6 addresses when we need to, okay? Basically we are not going to pay or get pulled into black market of v4 addressing. That's not the point; I think people are kind of jumping on that. I just believe that the problem should be solved, for layer-3 needs to connect people to people and that these layer-7 enterprise-specific needs are accelerating an already difficult issue. So it can be looked at or not, but, you know, RFC 1918 dismissed or not it's being recreated. If someone asks for addresses and we are going to, a bunch of us are going to be doing it. Anybody doing the IP video endpoints and knows that they are not going to route it, they are recreating that RFC right under your noses and it's going to continue happening multiple instances. A /8 would definitely buy people 5 to 10 years of solving this very sticky layer-7 issue that's going on for their enterprises. If that's the case then keep re-issuing the /8 and say fine. I mean it doesn't dismiss us from solving it at layer-3 for our high-speed data customers, but this enterprise-specific issue is just a pleading address pool facet that needs to happen. That's it.
MR. CURRAN: Good comment. I will point out that we had someone else from the floor that pointed out that the composition of attendees of ARIN, and I'll go as far as to say the IETF may not have a lot of extensive experience with providers like yourself, or large enterprise users who are extensive users of that address space. And so, it may take finding a constituency or group of such folks in order to really show the level of interest that you are looking for. I have closed the mics -- front center, yes?
MR. PALET MARTINEZ: Okay, one thing, we I think tend to forget is even if we exhaust the IPv4 space the transition is still possible. I will not obviously suggest that part and I insist that the important thing is to go ahead and plan ahead. But, it is possible to use a situation where we have private IPv4 addressing the space and NAT and still run dual stack with IPv6. And in fact there are solutions, which I believe that once there is more and more IPv6 traffic in the network many operators will go for, instead of a dual stack situation just IPv6 core and access net cores and keep NAT and private addresses space in the edge and this solution is, called it software. That's one thing that we need to remember because probably it can change a lot, the picture of a catastrophe or whatever, right? That's one point. The second point is some how there have been already some comments about this before. I think it's clear that there will be black or grey or whatever you want to call, market, and I am not sure if the registries need to take active part on that market but what I am convinced is that if we don't do some work on that and try to see where -- how that market must behave, that will be a serious problem for this community. So, I really want to encourage everybody to start thinking on that and taking a position and looking at possible scenarios about how that need to work to -- to be, let's say an oversight and avoid intervention by regulators or governments unless, we want to reach that point. And, a correction to something that Bill said before. The European Commission didn't spend money in let's say, doing marketing about IPv6. In general, the European Commission and I guess, that many governments and public institutions do like I guess, here NSF is investing in developing new technologies and not just about Telecom; Medicine or whatever. So, just a clarification.
MR. CURRAN: Good to know. Any Board comments? No, no, no, no. Back microphone?
MR. BEUKER: Scott Beuker, Shaw Communications. I just wanted to say that I think that ARIN had a duty to spread the word more about the impending walls shortage. Sorry?
MR. CURRAN: Speak into the mic it might just be me getting deaf up here after a while, go ahead.
MR. BEUKER: Okay. So, I was just saying that I think ARIN has a duty to spread the word with regards to this better. There may have been efforts in the past but timing is everything and I think 5 years out -- even right now if you were to put out a press release, people would ignore it. But, there should be a plan going forward, you know, that reaches a peak and say, two or three years to make everybody aware; I mean, just your common Internet user that there is a problem and have them ask their ISPs, their businesses, what are they doing, you know, to prepare for it. The analogy for me is Y2K, I mean, eight years ago, everybody I knew, who knew that I knew a thing or two about computers was asking me left, right and center, what's going on? What's going to happen? And yet right now, nobody seems to care or know about IPv6. So, something's gone wrong. And in anticipation of the comment what that -- well, that's not ARIN's duty, to that I say that it is because how is ARIN going to continue to function as it does right now three years from now when a room like this will be standing-room only as people come out of the woodwork, in order to find out, you know, why was my last IP request turned down? I've got a business to grow and you guys are costing me millions and billions of dollars. So it's in ARIN's best interest to anticipate this, to spread the word and to solve the problem before it's a problem. Thank you.
MR. CURRAN: I'm looking forward to those meeting, yeah. (Laughter)
MR. CURRAN: Any Board comments? Okay, the microphones are closed. We have two final comments, one from a remote participant, Ray?
MR. PLZAK: Yes. This is one more comment from Tony Hain. He said, "I just heard Paul's comment about 240/8 and it sounded like Scott commented about implementations being difficult to fix. Even if the vendors implemented a change and shifted within 18 months, an aggressive window, there is a very, very, very large installed base of systems that can't/won't be upgraded to allow use of a block that was undefined at the time they were tested and shipped. By the time those work their way out of the network we will be long past the point where the 240/8 block might have been useful, Tony."
MR. CURRAN: Okay, any Board Members want to comment? Scott?
MR. BRADNER: That's correct and it's somewhat relevant. Now, the postulate was to use some of these addresses for special environment, special enterprise network, or something like that where it's not as likely that this is the case. I have systems at Harvard that have had the same operating system on them since the late '80s and its unlikely anybody is going to touch them because they were setup by a graduate student, who's graduated and gone to hell. (Laughter)
MR. BRADNER: And so, they're never going to get changed and so we could never introduce that kind of address base into our environment. But, new closed environment deploys and for example in a video-on-demand sort of environment, something like that or new environment you could do something. But the amount of address base there is so small, relative to the overall problem; I just don't think anybody would find it worthwhile.
MR. CURRAN: I'm going to step in and comment. I just want to say that even a world where v6 is being deployed, there'll be because of transition reasons, as Jordi alluded to the need for IPv4 unique addresses to communicate with the still unconverted IPv4 Internet. In that world, one functional unique address, even partially functional unique address could be very valuable. I -- I have a list of the people who said, it's not worth the effort to free up that block and I'm going to refer those folks who come to me -- to people on that list. So, keep that in mind.
MR. BRADNER: All I need is one address. Ray said we got a lot of individual addresses lying around.
MR. CURRAN: Right. Okay, final comment? Marla?
MS. AZINGER: Marla Azinger, Frontier Communications. I want to make a -- there's a misperception I believe, from listening to people the past couple of days, about legitimizing commerce in the black market and should ARIN do something to legitimize and maybe take the reins of making it a commerce site ARIN or any other revenue? But, the point being, people seem to believe that if we do something to legitimize it, the black market will go away. I do not believe a black market will ever go away, its life, it's who we are. So, my question is pointed at the Treasurer, do you believe that if we were to do something it really would get rid of a black market?
MR. HOWARD: I think you're asking me, is there a way to get rid of something that I don't know whether it really exists?
MS. AZINGER: I think we could say it exists, the posting on eBay, how many years ago was proof of that? And you know they're not posted on eBay anymore. (Laughter)
MS. AZINGER: I guarantee you there's probably people still selling IP addresses here and there, so black market exists and I just don't believe that anything we could do would get rid of it completely. So, I just --
MR. HOWARD: Well, okay so, stipulating a black market, the question is will some sort of marketization affect the black market to either -- well, to eliminate it --
MS. AZINGER: Correct.
MR. HOWARD: -- and I suppose the goal would probably again, with that stipulation not be elimination, but minimization, right? So, we want to say if we could setup -- if we were trying to setup a market for address space, so that there was some sort of control over it in ways that community decided as a whole, that the goal would be not to make this -- to make the black market go away, but simply to create as much of a -- of a market driven by community consensus as possible as opposed to having a thriving black market.
MS. AZINGER: So you believe that it would minimize, but not eliminate?
MR. HOWARD: I think that would be the goal.
MS. AZINGER: Okay.
MR. HOWARD: But -- and then to sort of restate this, the -- my response to Jason earlier that I don't think that we have enough information yet and we haven't done enough analyses yet to be able to say whether it would work.
MR. CURRAN: Mr. Woodcock?
MR. WOODCOCK: I think to get some of the value judgments out of it let's just not talk about a black market for a moment, but just say -- stipulate that a market exists in a commodity, right and look at historical examples of people trying to influence prices in commodity markets. What we've seen is that in order to do that you have to have both money and a supply of commodity and be willing to lose both, right? You have to have essentially, as much or more than anyone else has stockpiled and you have to be prepared to lose both in order to control the price and I just don't see that that's either feasible nor a useful goal for ARIN. We don't have an inherent need to control the price in a commodity market, its not ARIN's role, I don't think.
MR. HOWARD: So, it sounds like you're saying we could -- there's a possibility over there, there could be a possibility of a subsidized market sponsored by ARIN or I guess, to subsidize by ARIN and assuming they were still a stockpile that creating subsidy for that stockpile --
MR. WOODCOCK: Well, again backing off from IP addresses and ARIN, right? If there's a commodity market and you're trying to control it, you have to be prepared to dump into the commodity market or to buy at above- market price, right --
MR. CURRAN: Right.
MR. WOODCOCK: -- depending on what you're trying to do?
MR. CURRAN: I'm going to --
MR. WOODCOCK: This is a very expensive process and not one that I see any reason that we would want to engage in.
MR. CURRAN: I'm going to step in here and say, because the -- we're talking about a topic that hasn't had enough discussion and hasn't had enough Board thought on, there is no Board position and so right now what you're hearing is a lot of speculation. The reality is that we don't have a position per se other than the historic position of ARIN, which has been that transfer of address space is per our policies. And so, I know some folks have advocated other alternatives; we haven't ever had a proposal where we could sit down and look at that says this makes sense for the community.
MS. AZINGER: Okay. My main point was I just wanted to clarify that we cannot say a black market will totally go away ever and I've just kind of been hearing obviously that thought.
MR. CURRAN: I actually -- I haven't even seen proposals that suggests that's the goal. I have seen people advocate that maybe some side effect of marketing could solve other problems. But, again no one has put a proposal in front of us to look at. So, it's very hard to --
MS. AZINGER: No, I'm going off of listening to people talking --
MR. CURRAN: Sure.
MS. AZINGER: -- that's all it was. Thank you.
MR. CURRAN: Okay. This has been hopefully an enlightening discussion for everyone. We feel its timely at least, maybe even has close to past timely as we safely could be and I think it's a good to get the community thinking about what we face over the coming years. We actually do have two policy proposals that are forward-looking, with respect to how ARIN should deal with IPv4 and potential change in the availability of address space. Right now are we going to go into those or are we going to break? Break? Okay, I'm going to turn this over to Ray at this point and we'll handle those a little later.
MR. PLZAK: Actually, I thought we'd just call it a day and have a little respite, eh? No? (Laughter)
MR. PLZAK: Thank you very much. I guess, another way to look at this if I could pipe-in one last comment, it's about April 12th on the Internet. The significance of that date is that the Titanic sailed on the 10th and hit the iceberg on the 14th. So, anyway, Richard will be waiting for you at the Cybercaf to spin the wheel of fortune, Andrew is no longer eligible, having won earlier. Still some prizes out there. The meeting will resume at 3:50, we will take up those two policy proposals at that time, thank you. (Recess)
MR. PLZAK: Okay, last break is over with. The agenda, now, is for us to take up a Policy Proposal 2007-12. The IPv4 Countdown Policy Proposal. And quite frankly, many of the issues around this proposal were discussed earlier for about two hours. So we'll see how more energetic we can get on this topic. So it was introduced on the PPML on February of this year. And designated a formal proposal in March. This is the first meeting at which it will be discussed. Text has not been revised. And Stacey is arriving on stage. Basically, what it does, it halts the direct IPv4 allocations and assignments from ARIN two years after the day that the IANA pool is equal to the 30 /8 IPv4 address blocks. The shepherds are Bill, Alec, and Suzanne. Legal Assessment, and as of April 2007. "This proposal addresses an important policy issue that is worthy of extended debate and consideration, but it proposes to do so in a way that may inadvertently create profound legal issues that would dramatically increase ARIN's potential legal liabilities." "The policy proposes to set a hard date to terminate IPv4 allocations. Adoption and implementation of such a policy has a clear legal impact; it could, for example, be deemed a denial of service by ARIN, which is a utility provider of such services." Continued. "For example, if IPv4 is exhausted and unavailable from IANA, there is little or no legal risk to ARIN. However, there are dramatically increased risks to ARIN associated with refusing to provide IPv4 addresses if ARIN has such addresses available, and is refusing to issue any of them to anyone based on a well intentioned but absolute policy." "ARIN's Legal Counsel will need to carefully evaluate any policy which results from this activity. Based on my current understanding, and I am willing to constantly reconsider and do additional research, I am likely to recommend that ARIN not adopt such a policy in its current form because of the profound legal risks it creates. This is one of the few times my advice prior to consideration of a policy has been so direct." Staff Comments. It is unclear what the reference to, "There will be no change to the policy on A- date" is. What does this mean? On this day and in the future? What happens to customers who come in after T- date to request IPv4 space? Do we deny these requests? Do we recommend IPv6? Do we do nothing? How can we deny a legitimate request when v4 resources still exist? Author did not indicate placement could be put in as new section 4.9 of the NRPM Section 4.9. Also, that section would need a heading, perhaps, "Availability of IPv4 Address Space." Need to make clear this applies to both assignments and allocations. Change the text that says "RIRs must" to "ARIN must." And references such as, "set the date" to "ARIN will set the date." Change /8 references to the amounts written out, such as "10 /8" to "ten /8s." Just some grammar stuff. And next, there is some other text things. It says, "Who decides, what the projections, how much?" It talks about removing the words, "Allocations or assignments" to "critical infrastructure" after T-date should be defined by a separate policy. Prior to T-date, there may be an increase in requests, perhaps necessitating more staff. What happens to reserved space? If IPv4 is replaced by IPv6 then one could assume that organizations would return their IPv4 back to ARIN, and that space would be available for reallocation. However, this policy would effectively preclude ARIN from reusing this returned space. A comment regarding revenue. The proposal would hasten the loss of revenue by approximately 12 percent per year, 4 percent if the IPv6 fee waiver was lifted. And the indicated NRPM Change. The implementation to actually do it, minimal. Determining A and T dates assumes coordination among the RIRs. And basically, there will be a guidelines change and a staff training problem. 155 posts by 47 people. It's actually -- almost look like a denial of service attack on the PPML there for a while. Zero for, seven against in terms of major threads. Discussion topics. I won't go through them here. They were posted, and it's a large amount of material. The -- lot of other things that are mentioned here were previously discussed, in the previous couple of hours. But some interesting comments regarding NAT, IPv6, v4-v6 translation, 6 to 4 protocol 41, black markets, fragmentation, routing table, vendors, aggressive vetting, price controls, IANA's allocation size, trading v4 for v6, legacy space reappearance on the net, police functions, an x-rated panacea, switch versus transition, IPv4 'emergency' reserve, and outreach and education. "The Proposers' point of view is to anticipate the run-out period and avoid a mix-up situation as far as we can." "I don't believe a policy is needed for this, just better publicity of the fact that IPv4 addresses are running out soon." "More likely to cause a run on the bank right after adoption, speeding exhaustion, not holding it up." "I think we should switch to IPv6 after the T-date. However, I do not believe the T-date is as soon as people think because I believe that recommendation -- reclamation should be tried." "Perhaps the greatest benefit ARIN is providing here is giving us a forum to discuss the issue." So Chair will be challenged, I think, on this one. The proposal text is in your handout and is also at this URL. I would like to call one of the proposal authors, Toshi, to come forward and make his presentation.
MR. HOSAKA: Thank you Ray. Hello everyone. I'm Toshiyuki Hosaka from JPNIC. And also thank you for providing the good summary, and intense analysis for this policy proposal. Actually I made a summary for the PPML discussion to the Japanese community in Japanese, but it's become four pages. So that's not the summary. Anyway, this is the proposal 2007-12, IPv4 Countdown Policy Proposal. So this proposal is to respond in an orderly way to the upcoming exhaustion of the IPv4 address space. And also to ensure that LIRs can receive IPv4 address block until pre-determined date. And this proposal does not intend to artificially drive up IPv4 addresses. We do not have such intention. And this proposal is not for IPv6 promotion. So how much IPv4 address left? I have -- I do not have much about this slide since you have seen this picture in last session. But we have -- and less than 50 /8s left currently, and IANA free pool will be exhausted in June 2011, and RIR free pool will be exhausted in June 2012, according to the projection by Geoff Huston. So current problems. Current problems is that the final date of IPv4 allocation is ambiguous. So LIRs do not consider IPv4 address exhaustion as an imminent issue. So they will likely face confusions such as re- addressing their network or making subsequent request -- subsequent allocation request at the very last minute in within a limited timeframe. Last minute rush. Someone called this a run on the bank. I think this is a very good analogy. And LIRs will be forced to build networks with a big architectural change either with hierarchical NAT or with IPv6, or even with another solution if there is -- in a very short timeframe. This is a very difficult situation for the ISPs. So this proposal consists of four principles. Global Synchronization, and set and announce the date when the IPv4 allocation is terminated, not change the current address policy for the extension of IPv4 address lifetime, and have separate discussions on "recycle" issue. Let me explain each by each. Number one, global synchronization. All five RIRs should proceed at the same time for measures on IPv4 address exhaustion. For ensuring fairness across the regions, and to prevent confusion such as an attempt to receive allocations from an RIR outside their region. Second, set and announce the date when the IPv4 allocation is terminated to ensure all LIRs or ISPs can receive IPv4 allocation until such date. Also, this is to give time to LIRs or ISPs to prepare for the network re- configuration such as large-scale-NAT, or IPv6, or whatever. Number three, not change the current address policy for the extension of IPv4 address lifetime. Because making large changes in the current policy towards conservation is very difficult in reality, as someone says the demand for the IPv4 is still growing and they keep growing. Number four, separate discussion on "recycle" issue. I mean, reclaiming and reallocation by this word "recycle." Recovery of unused address space is very, very important and should be addressed of course, but should not be tied with this proposed policy, because if we can succeed to reclaim as much as -- for example, 10 /8s, we can survive only one year longer. So this is the details of the proposal. Announce the day in which the IANA pool becomes less than /8. And terminate new allocation or assignment from RIR -- I mean, ARIN in this case, on the day exactly two years after A-Date. This is the proposal, as written. But most important thing -- most important intention for us -- to -- is to set the flag date such as T-Date. So current proposal says, we will terminate the new allocation. But we may change this proposal to have not 'terminating,' but 'adopting' a very strict policy or some exception of policy. But anyway, this is the current proposal, which is to terminate new allocation assignment from RIR. So benefits. Benefits is that the final date of IPv4 allocation is clearly demonstrated among the community well in advance. I mean, two years. So LIRs, ISPs, and users can prepare for the exhaustion. They can make a request of subsequent allocation, or they can make a renumbering, or they can plan another business plan, or they can adopt IPv6. In ARIN, RIRs can make the last allocation -- final allocation and avoid causing feelings of unfairness among LIRs by doing this. We have submitted the same proposal to APNIC in this year. And this three points regarding -- for these three points, we have reached the consensus; Global synchronization, and not change current address policy for the extension of IPv4 address lifetime, and separate discussion on "recycle" issue. But to set and announce the date when the IPv4 allocation is terminated -- this -- regarding this point, we have not reached the consensus. So this proposal is sent back to the mailing list for further discussion. And as you know, the long threads -- there was a long thread in the PPML, and we have a lot of comments on this proposal. For example, this is the artificial measures to move IPv6, or we are making IPv4 exhaustion happen faster. Or we should not sent cut-off date while IANA or RIR still have IPv4 addresses to allocate. Should not reserve the address or should adopt more strict policy. And we have more comment. Reclaiming unused address is essential, which I agree with. And should consider IPv4 address trade market creation. That's it for now. Thank you.
MR. CURRAN: Okay. Questions. Microphones are open. I recognize -- that's Scott. First I'm going to grab him. Center, rear microphone, Mr. Bradner.
MR. BRADNER: Scott Bradner, speaking for Scott Bradner. I have some questions about this proposal. I can't speak for the ARIN Board, but I can speak for this ARIN Board member, that if our lawyer says, "Don't go near that with a 10-foot pole," I'm not going to go near it with a 10-foot pole. And the lawyer has said that. I will say that it would take a great deal of effort to craft a proposal, which would more likely bring out the lawyers than this one. Antitrust issues here are really pretty amazing. I can't imagine doing this. But I also think that doing step 1 would probably mean we would never have to do step 2. If we make an announcement saying two years we're not going to give out anymore addresses, within two years, we wouldn't have any addresses to give out. (Laughter)
MR. CURRAN: Okay. The question of whether or not this would create a run on address allocation. Do you have a comment, Leo?
MR. BICKNELL: Leo Bicknell, ARIN Advisory Council. I just wanted to say from the Advisory Council perspective, you know, we got this policy very close to the meeting. The Advisory Council decided to kick it out, and it went through the petition process. And obviously, the wording and language doesn't match what a PPM or a NRPM policy would normally match in a region because we didn't have time to work with the authors, and re-word it -- that kind of, thing. And just from my own perspective, I would like that not to cloud the discussion here. I think there is very important concepts we need to talk about, and this is particularly not a case to get pedantic about, you know, it's worded this way or it's formatted this way. If people wanted to do some fraction of this, or all of it or whatever, we could obviously, fix that either through the Advisory Council process or some other mechanism.
MR. CURRAN: I agree that we have a situation here where we have lots of major topics about the intent and the possibility of the proposal. This is not a wordsmithing session per se. Front center microphone.
MR. DURAND: Alain Durand. I oppose the proposal as written. That said -- as I mentioned earlier, I have some sympathy about the idea of reserving some address space to create some kind of, swap space that could be used as garbage collection when the blocks will be fragmented way too much. So I think this is an interesting goal. Now, that said, about two points. The first one is setting a A-Date and a T-Date is kind of, useless because we have all seen the rate of assignments from IANA to the registries increasing. So saying that it will be two years between /13 and /10, for example, is probably not going to happen. It may just be a few months. So there's really no point there. Only one date is enough. And second, I'm not sure if this is something that should be taken by the registries and by ARIN in this case. This is more of something that should be taken on by IANA to say that they will stop allocating things to the registries, so that registries get essentially out of the way of those legal issues.
MR. CURRAN: Good question. Okay, microphones are open. I'll take the far left one.
MR. POUNSETT: Matt Pounsett, CIRA and the ARIN Advisory Council. There is a fair number of problems, I think, with this. And I'm just going to address one, because I know everybody else will have lots to talk about. The -- I don't actually support this proposal, because of, what I think is, the most obvious problem -- not the -- necessarily the most serious, but the most obvious to me is that one of its main benefits comes from having a fixed T-Date, so that everybody knows well in advance of when this is going to be. However, there is -- there doesn't seem to be a way to deal with the probability that we have guessed wrong about when that date actually is going to happen. We can't -- the T-Date can't slide if the graph changes because then it's no longer a predictable date. But if we -- if our estimates are wrong, or if there is a run on addresses, and we run out well in advance of that, then we haven't really accomplished anything anyway.
MR. CURRAN: Okay, thank you. Would you like to respond?
MR. CURRAN: Okay, far right microphone.
MR. ITO: Yes. This is the position -- my name is Kosuke -- position of that -- one of the proposal. And I -- we understand that there is a many or lots of the problems involved in this proposal. However, the first goal is to kicking off the discussion. And also we would like to have -- certain consensus of that global synchronization is important. And then -- well, the background of the setting of the termination date and terminate the allocation is the reason why we sort of that way is -- that is the most simple to be understand. If we get into that the more strict allocation policy, then how we can decide -- that kind of, a policy. That's -- create another issue. That's why we thought of that way, but even the determination is a quite -- raised out, more antitrust sort of, stuff of the issue then. We are not sticking to that idea. That's our consideration of this policy.
MR. CURRAN: Okay, good to know. Rear microphone.
MR. LEWIS: Ed Lewis, NeuStar. One of the questions I had about the policy that I'm not sure -- I've talked to you about this policy off and on, but the term "Global Synchronization" -- that topic, is that referring to you want to see the same policy enacted by all of the RIRs? Or is it intended to become a global policy affecting IANA, because that would make me read the policy in different ways because the staff comments and the legal counsel comment was ARIN can't do this. But if this is supposed to be a global policy, it's got to be discussed here, and then go to IANA, then IANA is the one cutting off everything, not ARIN. And -- it makes me interpret the legal comment differently. But then we have the other -- the other part as though -- as if in order to get to IANA, it has to discuss no other regions. You know, we have a catch 22 here. We have to say yes, we like this policy. Let's say we just do.
MR. ITO: In the parts, our intention is a global -- globally coordinated policy, so that -- among the RIRs.
MR. CONRAD: (Off mike).
SPEAKER: -- much less attractive.
MR. CURRAN: Yeah, so -- I'm sorry. Dave, I didn't hear your comment. Why don't you step forward.
MR. CONRAD: Yeah, I was just -- my view -- my interpretation was that it was a globally coordinated policy as opposed to a global policy, i.e., it would not necessarily go through the process of the ASO and going up to ICANN's Board, and stuff. It would be -- the RIR is coordinating among themselves.
MR. CURRAN: Right. As written, it's a policy that seems to be introduced and asked for coordination among the RIRs, but it's a policy that is -- each individual RIR's to implement. That's not to say it couldn't be made into a global policy. And there are some interesting side effects of ICANN making the decision to turn off the spigot. So that's a very good point actually, well raised.
MR. LEWIS: And there is a -- okay, so I came back to the mic to say that if it's going to be seen as we want ARIN to adopt -- this I would say, ARIN should not adopt this for the reasons staff and legal said. That's -- so I'm against the policy, as an ARIN policy.
MR. CURRAN: Okay, understood. Okay, we've got front mic, center mic. Owen.
MR. DeLONG: Owen DeLong, Jittr Networks.
MR. RYAN: John, let me --
MR. CURRAN: Sure, go ahead. Steve Ryan, Counsel.
MR. RYAN: Steve Ryan, Counsel for ARIN. Let me give the audience a couple of thoughts just so you understand the legal policy issue, okay. If you wanted to have a global policy imposed that shut off the spigot, then the easiest way to have the global policy imposed is to have government do it, because government is not subject to the antitrust laws, we are. So whether it's David, the IANA, and ICANN, or LACNIC, RIPE, and ARIN, and AFRINIC, and the five RIRs acting as individual actors, we are subject to the sovereign governments of every nation in which we sit. So when you do the antitrust analysis, the antitrust analysis has to be done in Brussels. It has to be done in Washington. It has to be done in California. It has to be done in Uruguay, it has to -- you get the idea. So as you think about this, we sometimes treat government as a problem in this area. If you want to think about it, government could be the solution if you actually had agreement amongst the parties, and government imposed the solution. It's one way of thinking about it. I wanted to make sure that people understood that what we're saying is, there -- you can address meaningfully the exhaustion issue, as policy. What you can't do is, say, we are going to turn the spigot off because that brings you the antitrust problem.
MR. CURRAN: Good point. Okay, center front, Owen.
MR. DeLONG: Owen DeLong, Jittr Networks. I'll go back to my original comment. But in response to what Steve said, I just love the image that it creates in my head of the United States Government telling ICANN and the IANA when they have to stop giving addresses to APNIC and RIPE, and the wonderful international implications that will carry. (Laughter)
MR. CURRAN: You might want to go to other governments first.
MR. DeLONG: Yeah. Okay, I really, you know, respect our honorable colleagues from Japan for bringing this proposal to us, and you know, it took a lot of work and a lot of bravery on their part to stand up here in front of us and get slaughtered like this. Unfortunately -- (Laughter)
MR. DeLONG: Forgive my bad Japanese, but (in Japanese). This is not the right policy for us, and I don't think it's the right policy for the Internet in general. I think artificially turning off the spigot at some arbitrary point in time determined largely by Geoff Huston's prognostications is not a good idea. I think that we should have the ability to use reclaimed address space, and such going forward if there is such a thing. And I think that there are better places to focus our effort. Thank you.
MR. CURRAN: Okay, good point. We have back microphone.
MR. WOODCOCK: Bill Woodcock, speaking not as a member of the Board, but as a confused member of the public. So I really respect the idea that increased care and attention needs to be given to this matter, as we approach exhaustion. And I appreciate the idea that a set of practices may need to be put in place around this process. And I think a lot of very good thought has gone into that, and I appreciate the effort that you're making to begin this discussion. I think that the very specific issue that, for instance, Matt has brought up, which I think is the most fundamental one with this as proposed right now, is the issue that we have on the one hand a pool of resource and we have a rate of (off mike).
MR. ITO: -- (off mike) you see principles. So, as long as we can guarantee that's how much we can reclaim in that space, then we can rely on that part of the space, so that we have to, looking at the fact of the consuming rates of the IP address space versus the free ports, that's my comments, thank you.
MR. CURRAN: Got it. Microphone center, back?
MR. LOCH: Sure. It's Kevin Loch from Carpathia Hosting. I do not support this proposal, and I just wanted to say I support and expect that ARIN would continue to assign globally unique addresses as long as there are globally unique addresses to assign. And I think there will always be globally unique addresses to assign even though they maybe, after being rationed with large -- long waiting lists and under perhaps very different policies than we have today. But that job doesn't end just because we're going to encounter a train wreck in a very short period of time. And to expand on what Scott Bradner said about the addresses not -- a two-year timeframe being too long, I would suggest it's probably closer to two weeks, because that's about as long as it would take everyone to get their last-second applications processed, or maybe two months because of the backlog. So --
MR. CURRAN: Okay. Center, front mic.
MR. SCHILLER: Jason Schiller of Verizon Business UUNET. I wanted to try and look at the topics of this policy. The first one, the Fixed Date; I think there have been a number of comments that says the Fixed Date is basically not doable because we can't possibly project when the date of exhaustion is going to be, and as soon as we announce it, there's going to be a run on the bank and it's going to be much sooner than that, which would require us to continuously moving that Fixed Date in, which makes the point of having a Fixed Date moot, in the first place, not to mention the legal implications of having a Fixed Date. The second topic would be notification. There's some consideration that people aren't taking v4 exhaustion seriously, especially people in this country, they think that they have enough IP addresses for the foreseeable future or at least until they retire. And I think if you think notification is useful and will help with v6 adoption, then simply go through the consultation process and ask each of the RIRs or IANA or whomever you think is a good organization to do this to continuously publish when the exhaustion date is based on the best data that we have. And this can either be Geoff Huston's numbers, if you think those are good, it can be Tony Hain's numbers, if you think those are good. If people on the community think those numbers are bad, then please go out and collect some data and publish a similar number, and tell us why your number is better. I don't know that ARIN is even equipped to make a projection or if they would simply pick one particular projection over another. And then the third topic was synchronization; the fact that all RIRs shouldn't run out at the same time. If you think there's use in this, if you think that it's fair that they should all run out at the same time, why not simply make them run out at the same time? Whichever RIR runs out first, picks up the telephone and calls the other RIRs and says, stop making assignments and allocations, or for the ones that have space left over, they start reallocating it to the RIR that's run out until you all run out at the same time.
MR. CURRAN: Interesting comments. Any -- I see the response here. So I am going to be closing the microphone shortly on this policy proposal. Center rear microphone?
MR. DURAND: Alain Durand. I have a question but actually would be more toward David that sits behind me, which is essentially, could we think about this, the other way round and put a problem up-side-down and say -- because we all uncomfortable with setting up a date. So, but we -- somehow we have this idea that we want to have this reserved space. So, why don't we set up the reserve space now and say, identify a number of blocks being one block, five blocks, 10 blocks, 12 blocks, 20 blocks /8 blocks, I mean and say that was our reserves to be used in a post exhaustion world? So, question to David is it something that IANA could technically do or not?
MR. CONRAD: I think it's safe to say that it is entirely possible for IANA to edit a text file and put the word "Reserved," on a set of lines. What that actually means will be left to the readers as an exercise. Yeah, I'm pretty sure about that one, yeah.
MR. CURRAN: Okay.
MR. CONRAD: To answer that question, its yes.
MR. CURRAN: Yeah. Leo, you wanted to comment earlier?
MR. BICKNELL: I -- Leo Bicknell, Harrah's Entertainment; I actually have a little bit of a different concern and I'm actually thinking, after the train wreck or hard landing or whatever we want to call it, I'm going to assume that somehow we get to widespread IPv6 deployment, and that after that occurs, at least some good, kind-hearted souls will return their IPv4 resources voluntarily, because they feel like it's the right thing to do. And so at some point in the future we will actually have plenty of free IPv4 resources again, because the vast majority of people will be on these V6. At that point, it seems to be that there may still be a valid reason for people to run v4 networks and, in fact, to even come back at that point and get new v4 resources. And this policy, as it's written, doesn't allow that future to exist. It assumes that once we've run out of v4 addresses, it's dead, done, and will never do anything again.
MR. CURRAN: And for that reason, you would be against this policy?
MR. BICKNELL: Correct.
MR. CURRAN: Okay. We -- this -- I'm not going to go any further. Okay, I'm closing the microphones though, so please stand in front of a microphone or you will not be able to speak on this proposal. Far right microphone? No?
MR. ITO: Maybe, okay -- okay.
MR. CURRAN: Any for and against?
MR. ITO: I'm for, of course, and -- (Laughter) ME. ITO: -- I guess, my just a personal feeling is that announcing date, such as in a football game, is you know, too many warnings. It's quite relief for the future readiness till we get ready before that -- something happen. Then, maybe at that time we can consider the fact -- based on the fact, we can reconsider the policy as well. But, setting those kind of a date based on some sort of a -- how could say -- estimation is quite variable for us to get forward. Then we never know how many ISPs can deploy the IPv6 at a time. So, maybe we can -- we don't have to worry about the v4 exhaustion, but we never know the times. So, set of date is really important, that's why.
MR. CURRAN: Okay. I am going to close the microphones. Microphones are closed. David, you only got to question, and didn't get to speak.
MR. CONRAD: It's okay.
MR. CURRAN: It's okay? All right, we might hear from you again in the future on this topic.
MR. CONRAD: Undoubtedly. (Laughter)
MR. CURRAN: Center rear microphone?
MR. LEWIS: Right, Ed Lewis, NeuStar. I actually wanted to add on to something Jason said the last time he was on the microphone. Thinking about globally coordinating a date of exhaustion, it's going to be better, but when one registry runs out, they all effectively run out because there won't be -- no RIR has a long large inventory, that's capped by how much I get from IANA anyway. The other thing I want to bring up to is that once there is no more new v4 to give out, the registry still have to get keep track of what they've already given out and the v4 Internet still runs. So we're not talking about the end of the Internet as we know it; we're not talking about the end of the registry as we know of, we're just talking about the end of allocation of new space; that's one thing to keep in mind in setting this date. There's also going to be the registration of or transfers of space or recycling of space. So the T-date -- these determination dates don't make as much as sense to me now as individuals because that just means there's no new stuff out there. They'll still with the -- registries still have to function over time.
MR. CURRAN: Understood.
MR. LEWIS: And so I'm still against the policy.
MR. CURRAN: Center front microphone?
MR. SCHILLER: Jason Schiller, Verizon Business, UUNET. I just wanted to respond to a comment Leo made about the policy not changing. I suspect what the author's intent was -- is that once they announce when exhaustion is likely to happen, they didn't want subsequent changes in the policy that makes it easier for people to get address space, and therefore to drive exhaustion to happen sooner. So, I might suggest some sort of alternate text that says, there will be no changes in policy that are perceived to result in increasing the rate of assignments and allocations until exhaustion is reached, until T-date is reached.
MR. CURRAN: If that change was made, would you change from being opposed to in favor of this policy?
MR. SCHILLER: I would still be opposed to this policy, but I was hoping I could remove Leo's objection.
MR. CURRAN: Okay, thank you.
MR. HOSAKA: But that is a very useful suggestion. Thank you.
MR. CURRAN: Okay. We have -- the microphones are closed. We have just a few more comments. So one of them is from our remote participant and, Ray?
MR. PLZAK: This is from Tony Hain. He says, "We are to burn through just short of one-third of the remaining spaces here." Given that we are on a compound growth curve rather than a flat 10/8ths per year, there is not enough time left to implement this proposal. I understand Geoff's number show a longer timeframe, but 13, 15, 18, is what looks like at the moment. So to implement a two-year stop, the event would have to happen early next year and that's not far enough down the road for people to react.
MR. CURRAN: Okay, good to have. Far end microphone?
MR. SHACKELFORD: Scott Shackelford, Cox Communications. Rather than -- and I'm approaching this from a liability perspective, rather than trying to draw a hard line in a sentence that says this particular date, the T-date in this case, why not, I think along Jason's lines, have an element of publishing that comes out quarterly, monthly, whatever, just to state what the current churn rate is or the burn rate or the exhaustion rate or whatever you want to call it? Today we've got X-amount. A couple of months later you say, "Now we have this amount." Let people come to their own conclusions as to when they think it may be exhausted. Therefore that will motivate them in their own personal selfish whatever ways. And it seems, correct me if I'm wrong, to be able to remove the element of liability of saying, "Okay, you said this date, it didn't happen." It happened sooner or it happen later, what have you, so, thank you.
MR. CURRAN: Okay. Microphones are closed, and there's three people where I thought I had one. (Laughter)
SPEAKER: We're all the same person.
MR. SEASTROM: Not by mass either.
MR. CURRAN: Okay, so would the 3-odd speak?
MS. ARONSON: Cathy Aronson, ARIN AC, I actually had a question for the other person who wasn't supposed to be at the microphone. What -- my question was, what would you require -- what would IANA require in order to make the edit in the host file for the reserved space? Not that I'm in favor or against it, but what would that require in agreement with the RIRs or an Act-of-God or a government intervention or -- I know you can edit, but -- (Laughter)
SPEAKER: The last one is probably the best.
MR. CONRAD: It's actually an interesting question that I -- off the cuff, and I'll probably get beaten by 15 different people for making this comment, but my assumption would be that it -- for that sort of policy decision it would actually have to be a global policy. It would have to be approved, it would have to go through the ASO; it would have to be approved by the Board and everything, but that -- as I said, that's just off the cuff. So if you're going to do that, start now.
MR. CURRAN: Okay. Last speaker, center rear mic, RS?
MR. SEASTROM: Rob Seastrom from ARIN AC. I'm not in favor of this proposal though I am in favor of some of the things that it's trying to do. I'm completely in favor of the suggestions that people have made, the ARIN and RIRs publicize on a -- at a regular interval when they believe that we're going to be out of IP address space; however, setting an artificial time date where there is actually some space left is just, sort of, not available unless you're important enough, and everybody thinks they're important? Bringing that into the fray, bringing up issues of while there will be some trade-ins or some reclaim space, some space, yeah, that comes from companies that go out of business et cetera, it all blunts the message and it confuses the issue. And really the timing for roll- outs of new equipment, the timing for converting stacks, and all the money that that takes, is in the SG&A people's hands right now in all of our companies. Putting anything out that confuses the issue of anything other than, the time is coming and you got to be ready by that time, and we aren't exactly sure when the time is, but it's going to be somewhere in this range. Anything beyond that, I believe, is counterproductive to the interests of having workable v6 out there within a few days after there has been a tremendous train wreck, and, no, I don't believe the train wreck is avoidable at this point.
MR. CURRAN: Thank you. All right then, time to thank your presenter. (Applause)
MR. CURRAN: At this time it's an important function to provide the Advisory Council with input to its task of helping advance policy proposals. To that end, I'm going to ask for a show of hands of those in favor and opposed to advancing Policy Proposal 2007-12. So, everyone ready? Yes. All those in favor of advancing Policy Proposal 2007-12 please raise your hand.
MR. CURRAN: Higher, yeah, nice and high. (Laughter)
SPEAKER: 35 to nil. (Laughter)
MR. CURRAN: Raise your hand high; hold the high hand. Okay, thank you. All those opposed to advancing Policy Proposal 2007-12 raise your hand. Nice try, David, one hand. (Laughter)
MR. CURRAN: You'll probably get to vote in another forum too so -- (Laughter)
MR. CURRAN: Hold your hands high. Thank you. Okay, and we have a count? All right, number of people in the room, this excludes ARIN Staff, 93. All those in favor of advancing 2007-12, 4; all those opposed 47; thank you.
MR. PLZAK: Okay, Policy Proposal 2007-6 IPv4 Minimum Size Change introduced on 15th February of this year, goes into proposal the next day, and the first discussion was at this meeting. Reduces the minimum size that's hired for a multi-homed end-user from a /22 to a /24, shepherds are Alec and Lea, liability risks from Counsel, none. Staff comments, there is very little qualification prior to, which could lead to policy abuse by spammers. This could significantly increase the number of requests for ARIN services thereby requiring additional staff. The policy applies only to end-users, which could also be perceived as unfair to ISPs. It could also be to potential abuse of the policy if ISPs apply as end-users for a single /24 IPv4 address block. Unclear exactly how an organization can qualify for /24 address block under this policy. It appears that the utilization require -- rate requires 25 percent immediate, 50 percent within one year, would be the justification criteria. However, the reassignments to multi-homed downstream customers indicates that an ISP can reassign a /24 IP address block without regard to planned host counts as long as the customer is multi-homed. The question here is, does this policy allow ARIN to qualify a requestor for a /24 based solely on multi-homing or should host counts also be taken into account? The policy does not address requests for more than one /24 IP address block for multiple sites. Then it's also Section 4.4 Micro-allocation should remain as is since it is a policy section essential for micro allocation for critical infrastructure related request. In other words, this would modify section 184.108.40.206. Minimal resource impact to implement will require some software changes, changes to the guidelines and some staff training. 18 post by 11 people, 5 for, 2 against. If a user only needs a /24, meaning that it's enough for their addressing needs, and they can get it somewhere -- someone to route it, then why waste 75 percent. I don't think the reasons for this proposal outweigh potential spamming issues or run an IPv4 space. Discontinuing "class C" assignments was intended to overcome late-90s router memory limitations. Is that still a significant issue? The proposals in -- text is in your handout and is also at the relevant site, and I would ask then if David Williamson to please come up and present this proposal.
MR. WILLIAMSON: Great, thanks. To start just with the proposed policy text, things of note, for those of you who haven't read it, is in the kind of gold colored tone is the exact changes in the section that I would propose changing. That's exactly two characters, that's changing 22 to 24. The intent is that the multi-homed, RPI minimum size, changes to 24; there is no other change. I suggest that since 24 is also the micro- allocation size and that is where all changes came in there, but those are not really necessarily part of the intent. A brief history lesson -- we used to do this, they were called "class C" back in the day. We found that the routing environment didn't really support that, and so back in the time of the minimum PI size was set at /20. In fact, that was the minimum allocation size and assignment size. Routing devices got better. 2002-3, change the minimum for multi-homed PI space to /22. A few notes that mostly are intended to address some of the staff questions. The utilization requirements still stand as specified in the rest of Section 43 for end-user sites. A few minutes questions about the routability questions, ARIN has never guaranteed that any assignments will be routable, although in practice /24s certainly seemed to be globally routable based on our experience throughout the use. Most of the people I talked to are surprised of /24 that's just arbitrary in the environment. It's not globally routable. So, while it's not necessarily guaranteed, it most likely is. It was actually tempting in preparing this proposal, to just simply remove the PI minimum entirely. The market will decide what's routable and what's not, and if you've got a small enterprise that wants to be multi-home with 28, if they aren't very interesting, then all of them won't get routed. If you had, hypothetically, Microsoft update homed in a /28, I think the market would find some way to route that. Still its -- while the market would sort that out, it's probably not really useful to introduce that much confusion in this sort of proposal. So, I kept that to a block size that's presently routable. Part of the intent here is to allow small organizations that feel the need for PI space to really get an appropriately sized space. If you only need a 24, there's no reason to either exaggerate the requirements and qualify for 22 and then waste the space, or be stuck with a PI space that you don't necessarily want. The requirements are just as stringent as it was with /22. The requirements that ARIN has set out are generally more rigorous than it would be for a small organization to get a 24 in PI space, obviously it varies by provider, but it's still the same level of stringent control imposed now by the policy -- or by the application process. My expectation is that this would have limited impact on routing table sizes. If you qualify under this proposed policy, you already qualify for a routing slot, through the PA. Obviously routing table can't tell PA from PI, that's all the same, but in this case it would be a PIs (off mike) slot and -- which shall be consumed in either case. There's some concern that this might make organizations that otherwise wouldn't qualify for 24 tempted to go out and, you know, become multi-homed such that they could get a PI 24. I think that's an interesting question. That's worth monitoring, but I still expect the number is fairly low. I want to talk about some of the negative responses from PPML and from the staff. There was a suggestion this might lead to a run on space. All of the rest of the RIRs have a similar policy of some sort, and that has not going to run on /24s in the other region that is detectable in the data. There was actually very little data available, to be fair, but on APNIC, for example, was able to supply me with a few numbers, and they showed that there's very little in the way of a run on space. There's actually very few 24s handed out. AFRINIC and LACNIC already have a minimum PI size of 24. RIPE has some different handling of how PI space is handled, to start with. But, there's a similar proposal under discussion in RIPE already, there's concern with the wording, but there seems to be some general support on their mailing list for the concept at least. And again, with the RIR, it doesn't allow PI 24s. So, is that spam thing, and I will profess this part of it with a note that I'm not deeply into the spam community, so I don't have a real good feel for some of the details in this. From what little I've been able to get from the other RIRs, there's little evidence that spammer are actually abuseless. It strikes me as unlikely that spammers would take the time in monitory investment to invest in being multi-homed and a small sized and then dealing with the registry and the justification required to get space versus their tendency toward -- hop on to the routing table somewhere, inject a horrendous spam and then leave again. In general, the type -- the spammer methodology seems to be more toward anonymity and the last point was something that someone pointed out to me. If you're a spammer, you know, your network ethics are somewhat in question anyway. I can imagine that they don't really care what the PI minimum is; if they wanted the space, they'd go for it. With that, get the Policy Text back up; any questions?
MR. CURRAN: Forum is open for questions. If you have questions about the policy proposal or would like to speak for or against it, please find a microphone. Okay, center rear mic.
MR. LEWIS: Ed Lewis, NeuStar again. On the previous slide, the first bullet you said that there was really no evidence from the other RIRs regarding spamming. I think I'm for this, but I want to clarify one thing. No evidence that this has occurred in other RIRs. I was curious to hear from RIRs who are here in the room, could give input to that?
MR. CURRAN: Point of inquiry. Are there any RIRs in the room who can speak about their experience with /24 requests? Is there a trend for more or less of them? Any RIR personnel in the room who can speak about that also with respect to the spam?
MS. YILMAZ: I'm Filiz Yilmaz from RIPE NCC. Well, first of all just a little clarification, we have a different policy for that in the RIPE region. We don't have a minimum size at the moment for PI space anyway. So we only issue space depending on the verification and, you know, what kind of information is provided for the network, network's need at the time. This one, the new one, for the multi-homing, I guess, is still under discussion, and it makes it more strict for us to go for one size for those. When it comes to the spammers' issue, I really cannot comment on that. We don't have any real substantial information or relevance in regards to spammers and /24s.
MR. LEWIS: I just want to ask you for that -- it's that, that would be the one reason why I would go against the policy. But right now I'm for it, but if there wasn't this demonstrable problem in the other regions, then I would change my opinion.
MR. CURRAN: Understood. Okay, far right microphone?
MS. AZINGER: Unfortunately I have couple -- Marla Azinger of Frontier Communications; I have a couple of questions. One, related to the spam, but it's going to require ARIN staff member to stand up. If we were to do this and we discover we're having spammers, which we will okay, and get the space, how do you expect to clean up the space because right now I know how I do it and other ISPs do it, we disconnect them, and everybody learns what happens and then we have to go clean up our block. But there is no control measure for ARIN to really clean it up as far as I'm aware of.
MR. CURRAN: You want to respond or -- Leslie? (Laughter)
MS. NOBILE: Okay. Okay, so right now, oops, we have spammers that come in, we catch them, we shut them off. We've sicked our lawyer on some of them and he may want to comment on some of that. We actually are actively tracking spam activity. We will shut them off; we will take the space back. If you submit fraudulent information to ARIN and we can prove it, we'll take it back. We spent a lot of time on this actually.
MS. AZINGER: So, thank you. That answers the first question. I didn't know that. It's always good to know, but then that raises a concern of how much that will increase in time and money for ARIN to be doing it versus many, many ISPs working on it as opposed to just one single RIR working on it. Jumping to another, the fudging factor, which being people applying to get a /20 or more, yes, people can lie, but it's a lot harder for them to make that information for a /20 versus a /24. That's just plain fact; at least it is to me. Hopefully, it's apparent to other people. The other thing is -- this could get us into the run down of IPv4 and into the black market area that I brought up earlier, that being, Joe decides, "Well, you know what? I've been paying attention to grant a v4 space, so I'm going to -- because we changed this, I'm going to get my /24 granted. I only need a /28, but if I get a /24, when the v4 space does run out or, well, hell, then I can just sell it on the black market to other people. It would be pretty simple for them to do that, and if we change the policy, it would make a little bit for them to do that even we set a rule that they had to have contracts that show, you know, they are multi-homing. All they have to do is get a contract for one year that doesn't cost much. But some will Podunk ISP here and there, and then after that one year they still have that, both /24, and then they can do whatever they want with it. So, sorry, those are my comments."
MR. CURRAN: So given what you've learned --
MS. AZINGER: I do not support it.
MR. CURRAN: Thank you.
MS. AZINGER: You're welcome.
MR. CURRAN: Okay, microphones are being opened, center rear.
MR. LOCH: Sure, Kevin Loch again. It's the first time on this proposal. Dave, this does not apply to ISPs only end-user assignments?
MR. WILLIAMSON: Correct.
MR. LOCH: Why was it written that way because I would personally like to see it if we do it for one, we do it for both.
MR. WILLIAMSON: There's some thought given to that. The feeling is that an ISP with a /24 probably doesn't really have enough space to have a viable business in the current, kind of, economic space. So there's not pressure to really do that for ISPs. If people thought that there were smaller ISPS, that really /24 was enough space to get them started, I'd probably, you know, be in favor of that proposal, but it doesn't seem viable, just economic reason.
MR. LOCH: Okay. And also I just wanted to point out that it wasn't mentioned in the presentation that this does intersect with v6 policy for better or for worse; that's the way it was written, so that this will affect the threshold for people to get a /48. So, I just wanted to point that out to everybody.
MR. CURRAN: Okay, thank you. Microphones remain open, center front?
MS. SCHILLER: My name is Heather Schiller with Verizon Business and I do not -- I'm opposed to this policy. So, the comment that I wanted to make is that with assignments it's my understanding that ARIN does not reserve address space. So they give a -- they would give a 24, but they don't reserve anything around it, so that the -- they could grow into it, which, if any year they did need to come back for additional address space, they wouldn't be able to grow into anything, like -- so it would add more rout to the routing table.
MR. CURRAN: Correct. Okay, in the queue, Alec?
MR. PETERSON: Alec Peterson, Message Systems and ARIN Advisory Council. I have two comments really. One, I am looking through the proposal I did not see what the problem that is hoping to be fixed by this proposal is aside from just making it so that more people can get more address space and other RIRs are using it, or other RIRs are doing it or considering it. So that's one comment, and another is just responding to the comment that -- about different policies. I agree that they shouldn't be different and I was not aware that a matrix for success for an ISP is how much IP space it uses. Thank you.
MR. CURRAN: Okay, the first point, you're asking for what other --
MR. PETERSON: Well, why -- whenever the -- whatever we're looking at a policy proposal, generally that should be some problem that should be fixed. Clearly, we used to get out /24s and other RIRs were doing it, but as far as the "why," I didn't really gather that from the proposal.
MR. CURRAN: Why this proposal?
MR. WILLIAMSON: It levels the playing field for small organizations. At present if you have a small business with a critical online service and you wished to be multi-homed in order to sustain your business, you're beholding to a PA space provider in order to have a valid IP space. For small organizations that don't need as large as a 22, this policy would level that playing field and allow them to have PI space.
MR. PETERSON: May I respond to it?
MR. CURRAN: Yeah.
MR. PETERSON: I'm not quite sure how I see how having PI space or how, to put it the other way, how not having PI space, put somebody at a disadvantage.
MR. WILLIAMSON: For small -- or small online services on variety, if you're in a PA space from a routing standpoint you're not disadvantaged, but from a business flexibility standpoint you're disadvantaged compared to a service that's slightly larger and you have no way to get out of -- on a contract with an ISP that sucks, or, for example, I had one person write to me when I proposed this and said, this is perfect because I'm a small provider and my upstream ISP, one of my two, the one I had my space from, went out of business, I had to renumber in one week. So there's clearly some small businesses out there that would have a lot of interest in being our PA space not to avoid certain business hassles that currently don't apply to larger enterprises.
MR. CURRAN: Okay, Alec?
MR. PETERSON: Okay.
MR. CURRAN: Okay, I'm closing the microphones. Please approach the mic if you want to speak in favor or against; center front?
MR. DeLONG: Owen DeLong, Jittr Networks. I support the policy to partially address Alec's question. There are a number of business continuity issues, there are a number of marketplace perception issues and worst of all, there is the whole issue of the number of places your addresses end up in that are not under your control in terms of configuration files and such. For example, one of the clients I recently worked with deployed a site in Japan and we have an APNIC/24 for that slide. If we were in the ARIN region, we would've been forced to get a /22 and we would've probably purchased enough extra machines to qualify for a /22 just to qualify for a /22 in order to meet policy instead of getting this policy and only using a 24. It's that much of an issue for them. They have a number of content providers who open up holes in their access list and who restrict down the number of machines and the addresses from which they can come, and it is not feasible for them if they want to switch providers to have to change addresses. So because of the way their business model works, they have to do whatever they have to do to be able to qualify for PI space, and they view that as a cost of doing business. Fortunately they were deploying in Japan this time and so they didn't need to worry about buying 300 extra machines to qualify for a /22; however, I don't really think that that's the kind of effort ARIN should be making. In terms of the ISP question, it is in a metric of how much IP space you consume that makes your business successful or not, but it was hard for us to fathom a meaningful or viable ISP that could only handout 15/28s and keep one for its internal infrastructure, which is what you're talking about when you give an ISP, a /24, to reallocate from. It's really hard for us to imagine a design of an ISP that would not need at least a 22 just to get through the next six months in any meaningful way and we felt that the aggregation was not useful to the community.
MR. CURRAN: Okay. Leslie, do you have a point of information?
MS. NOBILE: I'm sorry, I did. I'd forgotten that Leo Bicknell had posted on PPML, in relation to this policy he had enquired about the rate of -- you know, the burn rate of /21s and /22s. Since we implemented that policies, since we lowered the allocation as -- to 21s and 22s, so I went back and got some stats just to see if there was a run on the space as he said, and just so since the policy was implemented in May, we issued 607 /21s and /22 to ISPs and we issued 368 /21s and /22s to end-users. So -- and the larger picture wasn't a huge amount, just --
MR. CURRAN: Acknowledged. Okay, I am closing the microphones. I'll take comment from the rear microphone.
MR. VIXIE: I'm Paul Vixie from ISC, also member of the Board of Trustees. A /24 and the, you know, some reasonable ISP who won't need more space than that in the next six months. This is a can of worms, and we probably -- you're about to, I think, all agree, that we don't want to go down this rat hole, but there are a lot of ASPs, and there are a lot of managed rack space providers and virtual machine providers, a lot of them use NAT, because they consider ARIN's address space to cost too much, or because they can't multi-home or whatever. Some of them do multi-home, some of them just do e-mail. There are a lot of people for whom a /24 is enough to make a fine cottage industry to support a family for a period of years, and I don't think we want to either change the definition of an ISP to include ASPs and people who can live with NAT, but I do think we want to recognize that there is a valid need for multihoming that does not require more than a /24 and serves the Internet industry and the community just fine.
MR. CURRAN: So you would support this?
MR. VIXIE: Grudgingly, because I'm concerned about scale. If we had 10,000 cottage industries like this being created every year, then where would we be, that sort of thing. I was more comfortable with it when it was a /22, but that -- my grudging support of it makes me wish that it had a sunset provision where we would look at it in six months, and see how it had gone or something like that.
MR. MANNING: John.
MR. CURRAN: Okay. Yes?
MR. MANNING: I did have my hand up --
MR. CURRAN: Yes, Mr. Manning.
MR. MANNING: I did have my hand up, but if you want to go first --
MS. SCHILLER: Go ahead.
MR. MANNING: There was a discussion earlier before the break, where there was some indication that it might be the case where ARIN would be the happy recipient of some number of very small non-aggregatable chunks of address space, which you wouldn't know what to do with unless we had a policy sort of like this, and so I would be -- I am in favor of this proposal to the extent that I'm in favor of any proposal.
MR. CURRAN: Okay. (Laughter)
MR. CURRAN: Does that mean when the vote comes, you will raise your hand to the extent you'll raise any hand? (Laughter)
MR. MANNING: Maybe. We'll wait till the -- we'll see when the vote comes.
MR. CURRAN: Thank you. Okay, remind people microphones are closed, that includes the illustrious head table up here. Center front.
MR. SCHILLER: Jason Schiller, Verizon Business UUNET. Three points, the first thing I wanted to talk about was Section 4.4. You said that maybe the editorial changes could be abandoned, I would highly recommend that. Section 4.4 puts aside special space for core DNS servers as well as special space for exchange points, and the fact that these come out of a specific block and its known as important to some people. So I would not recommend removing that text and just allowing core DNS servers and exchange points to get their space under the PI /24 policy if it passes. The second point I wanted to make was about the ease of multihoming. One of the concerns that I had was that people who are not at all interested in multihoming would just go ahead and multi-home in order for them to get a PI /24, which would increase the number of routes in the table. And since they're actually not interested in multihoming, they -- after a year, they may decide to stop multi-home, or in fact just cancel their contract before the new circuit even gets installed and not ever multi-home, in which case they end up with a PI /24 that has to be routed even though they're a static customer only connected to a single upstream ISP. One of the ways to combat this problem is to raise the bar just a teeny bit. And one possibility is to require that anyone getting a /24 under the PI policy do it after they've demonstrated that they have been multihomed for say six months. This would also probably combat the spam issue, because it would require spammers to buy circuits and pay for them for six months. Hopefully, if they don't pay the bills, they would get turned off prior to six months.
MR. CURRAN: Jason, that is a interesting proposal, was it one you're going to make to the Advisory Council?
MR. SCHILLER: I doubt it, but I suspect someone else might.
MR. CURRAN: You should work on that. In the absence of that proposal, would you be for this proposal?
MR. SCHILLER: I'm against this proposal.
MR. CURRAN: Continue.
MR. SCHILLER: My third point is on the -- leveling the playing field for small organizations. I think, if you are trying to level the playing field for small organizations then you have to offer a /24 PI space to any organization that asks for it. And I certainly would not be comfortable with such a policy, so I'm not sure how much more level the playing field is now as opposed to before.
MR. CURRAN: Okay. Microphones are closed, but I see someone back up the mic. Owen, you're -- the microphones are closed, sorry.
SPEAKER: I kind of wanted to respond to Jason's comment.
MR. CURRAN: I responded to Jason, so it's okay. (Laughter)
MR. CURRAN: Okay, so at this point we need to get advice for the advisory council on advancing this policy proposal, and I'm going to ask for a show of hands, and we've had discussion on both sides. The -- I'm going to ask a show of hands of all those in favor, and then all those opposed. And so to it, all those in favor of policy -- advancing policy proposal 2007-6 please raise your hand, nice and high. Okay, thank you. All those opposed to advancing policy proposal 2007-6, please raise your hand. Okay, thank you. While we're waiting for the count, I'd like to thank our presenter David Williamson who escaped from the stage without a round of applause -- (Applause)
MR. CURRAN: Okay. On policy proposal 2007-6 IP minimum size change, number of people in the room, not counting our staff, 93, all those of advancing 2007-6, 19, all those opposed, 21. Thank you. I like folks like this, the AC's job needs to be challenging on occasion, and this is the thing that does it.
MR. PLZAK: Yeah, the AC had it too easy, everything was pretty well lopsided; they were just going to have dinner tonight, now they got to work.
MS. TAYLOR: I was getting really excited.
MR. PLZAK: Okay, we are now to open mic and we'll turn it to the chair. Remember your meeting courtesies and rules of discussion, and if you're not sure what they are, look in your program.
MR. CURRAN: Okay, this is open microphone, and this is the opportunity to bring any discussion of any topic relevant to the membership on policy proposals, this includes changes to policy proposals that have come by, deviate changes in wordings. We don't revisit exact questions asked, but certainly there have been on occasion topics that have come up that people would like a show of hands for, this would be an appropriate time. Mr. Bradner.
MR. BRADNER: I'm Scott Bradner, I just want -- a nit what you said, and it's not a concern to the membership, it's a concern to the public.
MR. CURRAN: Oh, of course, yes, it's actually the -- to the full public.
SPEAKER: Members of the public.
MR. CURRAN: Right. So microphones are open, and I got some people over here, and some folks there. Okay, far right.
MR. BARGER: Okay, Dave Barger, AT&T. A comment was made, I believe, yesterday during the open mic session about the desire and the need for ARIN to develop some sort of a web interface that folks can visit to monitor utilization -- I forget the exact context of it, but I would like to encourage them to do so, and would be more than happy to help provide input as far as, you know, what kind of content might be there, it would be very helpful for us to be able to go in and proactively look at our 50 percent plus SWIPs and who is that we have out there, and check that utilization against what our own internal database numbers are showing, and that would allow us to track things a lot closer, allow us to look at, you know, specific things, so -- not to get into a lot of detail here, but I'm very much in support of that.
MR. CURRAN: Okay, so along those lines, the -- if I hear correctly, you'd like to make sure people understand you're in favor of a web interface to the ARIN data, something friendly, that would allow you to see your scripts and your allocation data. I'm going to note -- certainly people hear that, additionally, the consultation and suggestion process, if you feel to submit it that way it will get a formal consideration and response.
MR. BARGER: Absolutely. Just one other thing I wanted to add, related to that too, I mean, you know, this could also include things like, you know, statistics, you know, how many, you know, /20s were allocated, you know, signed out of a certain allocation, you know, things like that, so thank you.
MR. CURRAN: Thank you. Okay, center microphone, back.
MR. DARTE: Hi, this is Bill Darte among the ARIN AC and I note that a number of the policy proposals that emerged for this particular meeting had to do with editing the NRPM, and I think that's a worthwhile goal, and I think there are many edits that could take place in there, and I wonder if there isn't a better way of consolidating a lot of those rather than one policy proposal per edit. And I don't have a suggestion exactly on how to do that, but I think it'd be more efficient that way.
MR. CURRAN: In terms of efficiency, you're saying that having a big editorial set would be better than having a bunch of small wordsmithing ones.
MR. PLZAK: John.
MR. CURRAN: Yeah -- one sec, I just want to -- could you answer that -- Bill?
MR. PLZAK: Yes.
MR. CURRAN: Okay, great.
MR. PLZAK: To answer your question, Bill, we are also, from the staff perspective, continually going through this, we're trying to figure out a right way to give it to the AC in small enough chunks to digest to work through a way through, and then they can decide whether or not to craft policy proposals, or whether or not it's strictly an edit.
MR. DARTE: I understand --
MR. PLZAK: So you will have more work to do, it'd be gainfully employed and have lots of job security.
MR. DARTE: -- I accept that role.
MR. CURRAN: Okay, thank you. Far left microphone.
MS. MURPHY: Sandy Murphy, Sparta. I made a comment on the discussion yesterday, and I was -- suggested that it might be an open mic question. The question had to do with certificate authorities. The PGP discussion has a lot of material in it about the trust chain. The question is, do the membership care to express an opinion as to whether ARIN should accept X.509 keys that are signed by other certificate authorities, and if so, would they want to express which certificate authorities?
MR. CURRAN: Okay, people who have opinions on the topic of which X.509 keys ARIN should accept signed by other authorities for purposes I presume of interacting with the registration data --
MS. MURPHY: I don't think the X.509 policy proposal specifically said what the keys would be used for, I'm certain that they were intended --
MR. CURRAN: Please find your way to a microphone, if you have a -- if any of the members have a feeling on this one way or another, we'll handle this topic, please. If you're not speaking about this, step back from the mics for a moment. Okay, on X.509, other certificate authorities -- yes.
MR. LEWIS: Ed Lewis, NeuStar. I -- what I want to add is I know that ARIN currently has an X.509 policy document on the web site, and I think it's already addressed in there. I don't know the policy proposal this time has much detail as that document. I mean, I'm not going to express an opinion on who they should go to, but I think that -- there's information already in ARIN that's under -- in production, that's already there to consider too.
MR. CURRAN: Right, there's a policy document that covers our existing use of X.509 certificates, I don't think there's one, or I haven't seen one that covers what's proposed in these three policy proposals.
MR. LEWIS: Right, not in the proposal, but I know that ARIN currently has a document saying, "Here are certificates that ARIN's currently processing." I don't know how actively, but there's a current policy in place, which says that only certificates issued from ARIN -- or ARIN only relies on its own certificates.
MR. CURRAN: Correct.
MR. LEWIS: So there is something in place already, this is not -- it's not like this hasn't happened in ARIN before, this is actually an in-place policy too.
MR. CURRAN: Correct, the question I believe, that's being asked is can that be expanded to other X.509 entities, do you have an opinion on that?
MR. LEWIS: No, I just want to make sure we knew we already had something --
MR. CURRAN: Oh, no, no -- far microphone is on the topic of X.509.
MR. MINGS: No.
MR. CURRAN: No. One moment, let's finish that up.
MR. MINGS: Okay.
MR. CURRAN: Other folks on the topic of should ARIN open this up to other X.509 certificate authority chains. I guess, you asked the question, is there a set that you'd like to see ARIN recognize?
MS. MURPHY: I did note that there were some comments that the procedure for getting a certificate involves a rather lengthy process of proving your identity, I'm not exactly certain why proving identity is necessary in this case, but if other RIRs are handing out certificates and someone needs to get -- go through their identity certification process requiring -- again, if they come to ARIN, seems to be an unnecessary burden.
MR. CURRAN: So you're -- are you -- is it your suggestion that ARIN accept other RIR certificates?
MS. MURPHY: Yes.
MR. CURRAN: Oh, okay, that's a pretty clear statement. Does anyone else want to speak about ARIN accepting other RIR certificates? Yes.
MR. BRADNER: Scott Bradner, I was going to limit it to other RIR certificates, I think the key point is that there is -- has to be a -- when you're doing an X.509 certificate, you publish a set of procedures you use to qualify those that receive them. Any decision on ARIN's part to accept some other person -- some other organization's certificates would involve ARIN staff reviewing those procedures and making a determination of whether those procedures are one that -- ones that ARIN believes are strict enough in order to assure the identity of the correspondents. That adds process to the ARIN staff, it's not necessarily huge. Accepting on blind faith that the other RIRs --
MS. MURPHY: Yes.
MR. BRANDER: -- RIRs are doing the right thing is probably a -- perfectly reasonable thing to do, but going beyond that, I think in theory it's reasonable, in practice it may not be.
MR. PLZAK: John?
MR. CURRAN: Yes, Ray.
MR. PLZAK: In the research that's going on in the certificate area right now, that is something that we have been looking at as far as accepting other certificates, so it's not a new topic to us, and it's not inconceivable that something like that would happen.
MR. CURRAN: You're saying as a result of the activity right now, working on the --
MR. PLZAK: Design --
MR. CURRAN: -- signing --
MR. PLZAK: The signing, yes.
MR. CURRAN: Okay. So there's work in progress that may, at least with respect to other RIR keys, allow some commonality there.
MS. MURPHY: Sounds good.
MR. CURRAN: If you -- I -- I'm going to go back and I'm going to -- this is going to sound like a broken record, but if you'd like a formal review and answer, I put it in the consultation and suggestion policy, and you'd get a response back.
MS. MURPHY: Okay.
MR. CURRAN: Okay, thanks. All right, microphones are open, far right microphone.
MR. MINGS: Yes. Sorry -- good evening, Stanford Mings of (off mike) services. I would like to open a discussion whereby you have a scenario, whereby here in the Caribbean where we have -- you have a LIR, it's almost like an ISP, that has been sanctioned as a monopoly by the local governments and this -- the scenario would basically -- would be that these LIRs, they're not able to provide certain services. A good example would be the provisioning of IPv6, IP addresses, and I don't know if there's anything in the policy whereby an organization -- who is -- whom -- maybe because they're either too small, or they don't fit certain requirements based on ARIN's policies, if there's -- if there's a possibility of policy changes so that these organizations that -- can come in and provide the services such as IPv6 around the existing monopolies.
MR. CURRAN: So I want to make sure I understand the question, you have an organization that you -- a small organization that you believe presently cannot get address allocations, for example, for the purposes of IPv6 services through ARIN, because of its policies, but also is unable to get it through the local Internet registry, because of local regulatory conditions.
MR. MINGS: Yeah, not necessarily due to regulatory conditions preventing IPv6, but the LIR -- either they see no need for it in terms of its, there's no -- it's -- costs too much, and that the regulatory restrictions are basically there because of the size of the community, for example, in the Virgin Islands, it's just -- it's a rural community based on SEC regulations.
MR. CURRAN: So your -- the organization that wishes to provide, for example IPv6 services, that currently the policies don't exist to allow that because it's too small.
MR. MINGS: That's exactly --
MR. CURRAN: What change to IPv6 allocation policy would you need to make that happen?
MR. MINGS: I'll be honest, I don't know. I'm just basically putting it out there to see what -- to see if there is something there.
MR. CURRAN: There is nothing, ARIN doesn't recognize anything about why you can't per se get addresses from your LIR, what you're looking for is a changed IPv6 policy for the purposes of getting them for your organization. And you should discuss that with some of the members of the AC and see if they want to suggest the policy, Lee.
MR. HOWARD: I also want to say that if the issue is that you can't provide Internet service or whatever kind of service because of local laws, then ARIN is not going to attempt to facilitate that because we will comply with all appropriate laws. But absolutely the policy process is open and if we need to change it, right now it's assigning, you know, 200 -- applied to assign 248s within five years, and we've talked about that today, so of course the policy process is open. And other than legal restrictions, we would certainly be entertaining further policy proposals.
MR. CURRAN: I would hunt down those folks who have AC badges and talk to them about what you need. Does anyone want to make a response on this point, I see -- I see someone, okay. Very good. Microphones remain open, thank you, front.
MS. SCHILLER: Heather Schiller, Verizon Business, but also the AC, and so I had a question about the AC and the consultation and suggestion process. And I was wondering if it would be helpful or useful to suggest or put out there for consideration that some of the suggestions that are of a lot of interest to the community, if the AC could help, like review and shepherd them similar to the policy -- the like policy proposal process, and that way -- and that they would get equal time at meetings, you know, for discussion.
MR. CURRAN: I'm happy, the Board would need to discuss this to get a Board opinion there, but I'm happy to give you my own personal opinion.
MS. SCHILLER: Sure.
MR. CURRAN: We want to make sure -- I'd like to make sure that the consultation and suggestion process has a visible public law, the way for people to see what's been suggested. Some of the policy -- some of the suggestions can be handled directly by staff or the executive team. Some of them require Board discussion, but the AC's got plenty on its plate to do good policy, and there's no reason to burden AC also with what a lot of case is, operational and execution tasks within ARIN. I personally would be against it. When the AC finds themselves bored with absolutely nothing to do five years from now, I'd probably change my mind, but frankly, I think we have a lot of policy issues that completely occupy them. Their charter per our bylaws is consultation on Internet resource policy formation, and that's not ARIN operation, its Internet resource policy.
MS. SCHILLER: Okay.
MR. CURRAN: That's just my take. Other Board members have other views.
MR. HOWARD: Yeah, maybe we should talk about that some more.
MR. CURRAN: What?
MR. HOWARD: Yeah, maybe we should talk about that some more.
MR. CURRAN: Yeah, we're going to -- probably going to talk about that. The Board is actually staying in wonderful raining Puerto Rico for another 12 hours to talk about some stuff on Wednesday, so it will come up there.
MS. SCHILLER: Okay.
MR. CURRAN: Okay. Microphones remain open. Front, center.
MR. DeLONG: Owen DeLong, Jittr Networks. This may be a little bit out of bounds in that it goes back to the last policy proposal we discussed, but I think that the, at least one of the major objections that I didn't get a chance to respond to can be addressed through an operational thing that doesn't need to be part of the policy and I would like to propose that. Certainly, it was not the intent of that policy proposal for people to be able to do it an end run around on the multi-homing requirement. I think the intent is pretty clear in the proposal and that this in no way modifies that, but I would suggest that it would be possible for ARIN staff at each renewal of those resources to require verification that the site is still multi-homed or refuse to renew the resource. And I think that that would address Verizon's stated concerns operationally without requiring modification to the policy. I don't know if there's a way to solicit additional input for the AC, whether that meets the requirements or not.
MR. CURRAN: Actually, in operational practice, to verify and then not renew, I guess, is how you phrase that, not renew resources, probably is a great topic to have in the NPRM, because the nature of the removing resources. It's -- and thank you, right. So the -- there's no reason why that shouldn't be suggested as a policy proposal and go through the process. I would be hesitant to see staff verifying multi-homing and not renewing, unless we knew we had a clear policy as a community on that particular aspect. Okay, front center.
MR. SCHILLER: Jason Schiller, Verizon Business/UUNET. One of the things that came up in the policy discussion today was about end site assignments and how to judge utilization of end sites or what's the appropriate size block to give to an end site. And I wanted to suggest that what's very important here is that we set up a process where we strike a balance between aggregation and efficient use. Obviously, in v4, we don't have a lot of space, so we're driven simply by efficient use in terms of the number of hosts in 30, 60, 90-days, what have you, and obviously in v4 we're not talking about hosts we're talking about -- in v6 we're not talking about hosts, we're talking about subnets, but what I'd suggest is we should have a guideline that is based on the utilization of subnets in a given period of time, multiplied by some sort of over sizing factor. And I wonder what people think about that, and just to throw in a number to start, may be we take the number of subnets an end site is expected to have in the next 10 years, and double that number and then round it up to the nearest v6 CIDR.
MR. CURRAN: Okay, comments on that. Comments on Jason's suggestion. Yes.
MR. BICKNELL: I'm concerned that we've never had great luck at predicting things 10 years out. Scott might be the only exception with his 2008, plus or minus three years. Spot on there Scott, but in general we don't seem very good more than about two or three years out predicting stuff.
MR. SCHILLER: Number of subnets in five years quadrupled, and then round it up to the nearest CIDR.
MR. BICKNELL: Maybe, I like your concept, I think you're kind of on to something in that concept, but the timeframes seemed long --
MR. HOWARD: You could take a hint from the HD ratio and go logarithmic instead of multiples, if it makes you happy.
MR. CURRAN: People who have opinions on this policy thought. Jason it's a probably good one to try it out, put into the fray. Center rear microphone.
MR. LEE: Hi, Louis, ASO AC. It's I guess my duty. In the last 24 hours two policy proposals in the LACNIC region were put on their main lists, and they could potentially become global policies. One of them has to do with IP exhaustion for v4, when it gets near to the end. Basically, five blocks of equal sizes should be distributed to all the RIRs, and the other one has to do with ASN assignments from IANA to the RIRs, has to do with when IANA would be given a 4-byte and -- or stop doing 2- byte and start doing 4-byte ASNs.
MR. CURRAN: Okay.
MR. LEE: And if you don't -- if you don't want to figure out where the links are to these policy proposals, just e-mail me, I'll send you the links directly, louis at equinix.com.
MR. CURRAN: Well, it points out that we have some policy proposal at LACNIC, that could be showing up here in another form or already have. Yes, Mr. Manning.
MR. MANNING: Aren't all ASNs 4-bytes? That was my thought looking at the spec, which is that we only allocated off the bottom half of that for ever and ever, and many people just assumed two bytes.
MR. CURRAN: Another comment at the end, Ray.
MR. PLZAK: That policy proposal regarding AS numbers happens to mention 4-bytes, but that's not the intent of the proposal from the way I understand it. It is the intent to provide to the IANA, a policy that is -- that tells them how to interact with the RIRs with regard to AS numbers, the same way they have a policy for IPv6 allocations and IPv4 allocations.
MR. LEE: Thank you for that clarification.
MR. CURRAN: Okay, Lee.
MR. HOWARD: Louis, would it be reasonable for us to hope that these policy proposals would be on the NRO NC website sometime in the near future. Should I ask the NRO secretariat?
MR. PLZAK: The answer to your question is that they'll get proposed and passed in each of the regions. And it will get to you. At some point it will get to that -- to the site, but what will happen to ASO site, is it will just be links to the various RIR sites, that will get there.
MR. CURRAN: Right, when it is available in all five regions as a policy proposal, then it is qualified as a global policy proposal.
MR. LEE: Until then, if each of will still track it informally.
MR. HOWARD: It's already a proposal, it's been proposed as -- to be intended as a global policy proposal. Do we have to wait for it to be sent to all five mailing lists?
MR. PLZAK: It will get there, believe me.
MR. HOWARD: I have my doubt.
MR. CURRAN: Presently the global policy -- the global policy process says that it is a global policy candidate when it has made it to all the RIRs.
MR. BICKNELL: I think what Lee is trying to say is it might be useful if the NRO had a webpage that tracked attempts to do that whether successful or not. So if people wanted to find failures to use as a guideline for what to do in the future, they were in one place.
MR. CURRAN: That's an interesting thought, so Ray, can you discuss that with the NRO folks.
MR. PLZAK: Yes.
MR. CURRAN: Thank you. Okay, microphones remain open. Center back.
MR. LOCH: Okay, Kevin Loch, Carpathia Hosting. I wanted to -- there's been some discussion in the hallways, and this also relates to Jordi's policy earlier in the week. There's an incongruency between the recommended assignment size that ISPs should give out in v6, /56 in the general case, and the existing language that the initial assignment be based on 200/48s. So I think we need to propose a new policy to modify that. I'm just wondering what people's thoughts are. One possible way to reconcile that would be to say 200 or x number of /56 or larger blocks, so that it depends on the type of blocks that an ISP is giving out.
MR. LOCH: Do have a --
MR. CURRAN: Okay.
MR. LOCH: Response there?
MR. CURRAN: I have none, people want to comment on the inconsistency, and need for reconciliation, find a microphone, if you would like to speak on this. I see two, Jordi is first. Go ahead, Jordi.
MR. PALET MARTINEZ: Well, I'm drafting some text as I think I mentioned in the mic, to correct all the criteria not just Section D, I think it was yesterday. So, this is one of the points that I'm already considering.
SPEAKER: Okay, thank you.
MR. PALET MARTINEZ: I will probably instead of sending a proposal, first, do some kind of trial balloon in the list, to ask some questions to the people and then send the proposal.
MR. CURRAN: Or you could solicit here people who want to help work with you on it.
MR. PALET MARTINEZ: Of course, yeah.
MR. CURRAN: Could find you?
MR. PALET MARITNEZ: Yeah.
MR. CURRAN: Okay, Paul.
MR. WILSON: Yeah, Paul from APNIC. We've passed policy that makes the length of the IPv6 assignment discretionary up to the ISP, and the corresponding part of the policy says that they need to have made 200 assignments.
MR. CURRAN: Okay.
MR. WILSON: Of any size.
MR. CURRAN: Okay, so that's the practice at APNIC, okay?
MR. LOCH: Could I respond to that for just a second? I like that approach, the only concern I would have is that an ISP that is giving out /64s or /112s, that might not be what want to do. So may be set a minimum size that's smaller than 48 bits --
MR. CURRAN: That counts for purposes of allocation. If you have a recommendation in this area, Jordi is in the second row, 1+2=twice the force working on it. Okay, thank you.
MR. SCHILLER: John.
MR. CURRAN: Microphones remain open.
MR. SCHILLER: John, I had a comment on that also.
MR. CURRAN: Yes, okay, comment.
MR. SCHILLER: I think you could easily resolve it by saying 200 customers who are 200 unique customer end points. And not really specifically deal with whether they are making 48, 56, or, you know, /32 assignments to end sites. Well, 200 customers, I think is, yeah.
MR. CURRAN: Now we have three people interested in -- we honestly have a design team now. (Laughter)
MR. CURRAN: Do you folks want to run off and sit in the bar? I'm sure we'll get a good policy proposal shortly, Leo.
MR. BICKNELL: I would like to make a quick comment on that one other idea that has come up in the hallways; I have discussed with two or three different people. And it actually came out of the staff comments in the open policy hour about what constitutes an organization, since we have used that several places. But the suggestion was, since ARIN requires paperwork from people to prove various things, require 200 contracts to provide IP services. Because that would exclude for instance the business that just has, you know, three sub companies and they provide service to each other, not because there is a contractual ISP relationship, but just because they are related and do things for each other. So -- and it's also something very easily intangible for staff audit, they get a stack of contracts and count and go 200, good.
MR. CURRAN: Okay, we got four people for the design team. (Laughter)
MR. CURRAN: I'm actually closing off discussion of the hypothetical policy proposal that hasn't been proposed yet, but if you four get together I'm sure you can come up with a good one. Okay, I'm going to be closing the mic shortly, all other discussion -- anyone else wants to comment at the open mic, please approach the microphone. Yes, far left side.
MR. POMEROY: Hi, John Pomeroy, Cablevision. I just like to put a final cap I guess for the day on the concept of shall we say preserving the IPv4 space. You know, I feel a little bit like (off mike) from your example earlier in that we do use millions of addresses for internal use, and that's kind of where we're trying to go with this. And we're not, you know, any of the large MSOs are going to have the same problem, so we're just having this massively accelerated depletion of the v4 space, it seems like a ridiculous concept. Maybe it doesn't help some of you, but it does help us. So if anybody else is interested in discussing this and figuring out if there is anyway we can do this, please grab me in the bar or wherever, thanks.
MR. CURRAN: Very good, okay. Microphones are closed. Last speaker, Marla.
MS. AZINGER: This is just an announcement for the Advisory Council our meeting will be promptly 15 minutes after we stop this meeting. I'm not going to give you a time on my watch because god knows what's on yours, so 15 minutes after this ends, thanks.
MR. CURRAN: Okay, and that thus ends our open microphone session, thank you. (Applause)
MR. PLZAK: Okay. First of all let's thank our sponsors and special participants. (Applause)
MR. PLZAK: Don't forget to complete the v6 workshop, and ARIN XIX meeting surveys are online, and get yourself a chance to win one of those. Reminders tomorrow for the Members Meeting, the breakfast is at 8:00, the meeting is at 9:00. Even though this is a Members Meeting it's open to anyone that wants to attend, the agenda is in the meeting bulletin. If you're not going to attend and you do have a wireless card, return it to the registration desk before you leave, or it is yours at a cost. And the wireless network will shut down after the Members Meeting tomorrow. So, with that I say thank you very much. We have had a very fruitful two days, and thank you for making this a very interesting discussion.
(Whereupon, at 5:45 p.m., the PROCEEDINGS were adjourned.)