______________________________________________________________________ DRAFT TRANSCRIPT APNIC Member Meeting Session 3 Friday 3 September 2004 2.00pm ______________________________________________________________________ PAUL WILSON: Welcome back. The scheduled afternoon agenda is here. However, the good news is we have actually completed most of the later afternoon scheduled sessions, so we've got a big incentive to proceed through the SIG and BOF reports in an expedient fashion and I'd like to start running through in this order. If the presenters are willing to present if this order, we'll start with the policy SIG. Arano San is here. While he's starting I'll mention the wireless network is shutting down at 5:30pm. The terminal room is already shut. The wireless is closing at 5:30. For the purposes of this meeting, we were generously donated a couple of dial-up accounts for use as needed from Connect Fiji. They have even more generously offered to keep the accounts open for a week. So anyone who is staying on here beyond say tomorrow, if you arrange some working holiday, for instance, and you'd like to have access to a dial-up account for the next few days, Connect has offered that to you. So please approach Connie here or Save about the details of that dial-up. Starting with the report from the policy segment. Perhaps before Arano San starts I should mention the procedure we're going through here for the benefit of those who haven't been here before. The report is part of the policy development process we have within the APNIC meetings. What happens is we have the SIGs where concrete proposals can be presented and the role of the SIGs is to review those proposals in detail, to discuss questions and all the details and then to bring some consensus to bear on whether the proposal is to be approved or not. The proposal can be approved by the consensus of the SIG. It can be abandoned by the SIG or no consensus by the SIG means the proposal is generally dropped, or it might be sent back for some more work. But in the case of a consensus to make a policy change, generally a consensus to approve a proposal, the SIG's decision is not necessarily binding. For one thing there is an ongoing comment period. That comment period goes for two months after the meeting. we also ask that this meeting, as the meeting with officially this function, should review the decisions of the SIG and confirm them or endorse them by consensus. So where there's a particular proposal described in any of these SIG reports, I'm going to be asking you if you would raise your hand if you wish to endorse the proposal and if we have a consensus in this meeting, then it will be recorded that the SIG proposal having come through the SIG has then been endorsed within the member meeting. This endorsement, this voting, if you will, or indication of your support for a proposal is not restricted to members. So if you're here not as a member of APNIC you can participate as well. But the function of the meeting is to provide this endorsement. So I would ask you to please take time to just listen to the proposals here and please do make the effort to put your hand up if you do support the proposal. I'll also be asking if you don't support the proposal. So hopefully we'll have a good number of hands up for either or both cases and we can get some clear results out of this meeting. That's just so that you understand what's coming up over the next hour or so. With that, I hand over to Arano San to present the report. TAKASHI ARANO: This is the 10th address policy SIG report. My name is Takashi Arano, the chair of this SIG. Here we have the address policy SIG agenda. There are six or seven informational presentations, the rest policy. Session 1, 93, session 2, 92, session 3, 70. First of all Yong Wan Ju stepped down as chair. We called for the new co-chair. Three very excellent people nominated. Among the three, this candidate, Hosaka, was elected through that voting. This is the one big topic for us. Yong Wan Ju is just behind in the room. I'd like to thank him for all he did with this SIG. The first proposal was about dark IPv4 space and was presented by Gordon Bader. He participated by phone. His proposal was to withdraw address space from LIRs who route dark prefixes. A lot of questions arose from technical and procedural viewpoints. Some opinion says it is practically impossible. Some comments say it is not allowed, it should be originated, or something like that. Consensus has not been reached and probably we will continue to discuss these things in a mailing list if this proposal wants to proceed or be kept open. The next was IPv4 HD ratio, presented by Anne Lord. The proposal was to apply HD ratio for IPv4 allocation. Also there was a lot of discussion. However, there was no clear consensus. Actually some were in favour and some disagreed. I propose that the APNIC secretariat and NIR do a small survey on ISP address management problems - for large and small ISPs. My proposal is just to do a survey - I added not only APNIC secretariat but also the NIR secretariat to do this job, to help the APNIC secretariat to do this survey. The consensus for this part is to keep open the proposal and the secretariat and NIRs to do a small survey on ISP address management problems. Can I go on? PAUL WILSON: This is an agreement by the SIG to keep the proposal open. It's not a policy consensus so there's no need to look at it any further. TAKASHI ARANO: The third one is expansion of IPv6 initial allocation by Tomohiro Fujisaki. His proposal was to expand IPv6 address space for existing IPv6 address holders without satisfying the subsequent allocation requirements. There are some questions and answers there and the consensus was reached finally. The fourth proposal was IANA IPv6 global allocation by Paul Wilson. This should be a global policy. This proposal about how to allocate address from IANA to RIRs. Also there was a lot of discussion during the meetings - the proposal was changed several times. The consensus reached was on the latest proposal with some flexibility of parameters. Flexibility is left for consensus of other regions. This means that especially the number of parameters will be discussed in the mailing list. But the general framework or idea of this proposal was accepted and probably within that consensus was some extent of flexibility - we all agreed to this proposal, especially as it will be discussed in the mailing list whether the extent of the parameter is appropriate. There were many other informational proposals, but I will not explain these proposals again. That's it. That's all to report. Thank you for your cooperation. PAUL WILSON: I think we need to come back to that one. What you're saying, Arano San, is one proposal passed through the SIG with consensus so that's one we need to reconsider here now. TAKASHI ARANO: I don't think we have to discuss this now. What about we do a survey? PAUL WILSON: I suggest it's not necessary to call consensus on this. The consensus of the SIG was simply to keep the policy proposal alive, without either rejecting or accepting it, but to have further work. TAKASHI ARANO: I added the NIR in this action list item - so I will write to ask the NIR folks to agree with this action item. IZUMI OKUTANI: Yes from JPNIC. PAUL WILSON: Any proposals or suggestions before we move on here. So the APNIC secretariat with NIRs will do some sort of survey to be determined and that will be happening in due course. Thank you. TAKASHI ARANO: The next one is expansion of IPv6 initial allocation. PAUL WILSON: You're quite correct. That is a consensus policy decision of the SIG. Are there questions? RAM NARULA: A quick question here. If I apply for IPv6 space and ask for more, keep on asking for more, there should be some clear point where we can know whether it's a yes or no to get more. Do you understand? If I want to apply for IPv4 space I can just ask for more. What about with IPv6? TAKASHI ARANO: My understanding is if you already have a certifiable /32 you can make a request for the additional space with some deployment without actually using up the HD ratio. That's my understanding. RAM NARULA: With the deployment plan, I think it should be added. PAUL WILSON: The way this is expressed, I see the point of the question. It looks as though, the way this is expressed, as though there are no conditions on expanding an existing allocation. Could you phrase what the ... TAKASHI ARANO: Yes, I can do it with the deployment plan. This is just a summary - in short. The actual proposal was written by Fujisaki San. That was a formal proposal to be accepted for all. So I can do this. RAM NARULA: I just think it is not clear. ANNE LORD: Just in case there is any confusion, the actual criteria that are used, it's about using the IPv4 infrastructure as a means of justifying an IPv6 equivalent allocation. So just to clarify that. KOSUKE ITO: The plan or the infrastructure - some kind of supporting data - either or both or and/or - something like that. PAUL WILSON: Are there any other questions about this proposal? I am going to be asking for this meeting to approve or endorse the consensus of the SIG, so it is important that we have people who understand what's being proposed here. So if there are any remaining confusions, then please let us know. So in that case, I'll ask whether you will please raise your hand if you're happy to endorse the proposal. Could you raise your hand if you do not support the proposal. I think we have rough consensus. Thank you very much. Moving along to the next one. TAKASHI ARANO: IANA IPv6 global allocation. PAUL WILSON: As we've heard, this one also reached consensus for the latest proposal allowing some flexibility of the parameters, the flexibility being left in order to accommodate the consensus of other regions and also for discussion on the mailing list. I'll note again that discussion on the mailing list is a mandatory part of the policy development process, so none of these policies are finally approved until a two-month period on the mailing list, after which the APNIC EC is asked to look at the consensus of this meeting, look at any contributions on the mailing list and then to endorse that consensus and approve it, providing that there are no, for instance, major objections that are raised on the mailing list or other difficulties, other problems that arise. In this case, having seen this proposal reach consensus in the address policy SIG, I'll ask you are there any questions about the proposal, questions or comments of any kind at all. If not, then would you please raise your hand if you'll support the consensus on this proposal. If you do not support the proposal, please raise your hand. Thank you. That's rough consensus. Thank you very much Arano San. The next will be the database SIG. I believe there's a comment or question that's come in on jabber, however. Is someone able to read that. APNIC STAFF MEMBER: It was actually not a question to the policy SIG report. It was a question he wanted to be read out in open mike. Will I do that now? PAUL WILSON: Yes. APNIC STAFF MEMBER: It was another comment from Frank Martin. He wanted to thank APNIC for having its meeting in the Pacific islands and wanted to thank Fiji Connect for making it happen. He is looking forward to greater collaboration between APNIC and the Pacific islands. PAUL WILSON: Thank you, Frank Martin. XING LI: Good afternoon. The DB SIG - there is no formal proposal to present here so I will just summarise the items discussed in database SIG. There are several things going on for the previous open action items, action db-16-002: the update is to be implemented by the third quarter this year. A similar thing to db-17-001 - also it will be summarised later. There is open action item db-17-002: Proposal for IRR mirroring policy. This one the presenter postponed the presentation and it is not presented in the database SIG and is still open. The presentations carried on in this meeting. The first one is from NTT. The presenter proposed that APNIC establish a trusted IPv6 IRR that would also be promoted to other RIRs to contribute to the stable routing of the IPv6 network. There were questions and discussion. There was a show of hands in favour with no objections. However, the details need to be discussed in the mailing list. Also APNIC expected to have a new version of the RIPE software that would support IPv6 and the multicast routing registry by the end of this year. So this actually would fit the proposed time line. Also exchange cooperation with MERIT and other entities. So the action items - db-18-001: the status is open to be implemented by December this year. The privacy of customer assignment records project update was informational. The highlight is when systems are ready, all portable allocations and assignments will remain publicly visible. However, all non-portable resources will be made invisible by default. So if you inquired of APNIC Whois database you would see a difference. Some items you cannot see if you are not the owner. Second, the members of APNIC have to use MyAPNIC at this stage. Email amendments will be supported later. There was discussion on the details, action items none. The second information presentation is protecting historical records in the APNIC Whois database. The situation is when this is done custodians will still be able to use the resources. Those who wish to have their own maintainer applied to the resources will be required to open an account with APNIC with an annual fee of US$100 per year. There was discussion on this and some details - I mean some unexpected things may happen from this action. Action items - none. The modification of Whois domain object authorisation - the problem is reverse domain names. Because of the hierarchical nature of authorisation it can cause automatic submissions of domain object to fail. That's the problem. Currently it is maintained manually. The presentation contains details of the procedural solutions for this problem. The new procedures are intended to be implemented by the end of this year. No questions, no action items. Finally the RIPE database software update. There are some new features which will be available soon based on RIPE database. The highlights are marked in yellow. The support for X.509 has been added and the organisation object has been introduced. Reverse DNS procedures have been changed - domain and DNS records are no longer maintained separately. There are no unprotected objects. CIDR notation will be used to create inetnums. There was no discussion and no action items. Thank you very much. PAUL WILSON: Thank you very much. There is one proposal or item there which I think formally should come back to this meeting and that is the proposal for APNIC to adopt the IPv6 routing registry. It may be regarded as a subset of the routing registry that we currently operate, but I think since it has been put to a SIG consensus and it may be seen as a policy matter, I'd like to ask you to give a consensus opinion on that. Are there any questions about the IPv6 routing registry? If not, could you please raise your hand if you support the proposal to implement the v6 routing registry. If you're opposed to the proposal. OK. Thank you very much. The next one is Joe Abley and DNS operation. Those watching the jabber feed online, let me know if there are any questions that come in in time for discussion. JOE ABLEY: Good afternoon. The DNS operations SIG meeting was held yesterday. We had about 70 people attend - it was pretty well attended. We had a full agenda. We filled our entire slot, but only with information updates - we had no policy proposals. We had a number of operational updates from APNIC staff and a presentation about global efforts in IPv6 and DNSSEC from Bill Manning. We had one outstanding action item from APNIC 16 which was to ask the secretariat to implement a proposal - prop4-v-001. That was held over from the APNIC 17 meeting. It's close to being implemented now. The work involved in carrying out that action item is now at advanced stages of testing and is ready to be turned on at the end of this month. I won't say more about it now because George made a presentation about it which I'll cover in a second. There were no new action items raised. The first presentation we had was from Terry Manderson. It was an informational update about the load of queries on the nameservers that APNIC runs - that's both master name servers which are directly related to registration activities and slave nameservers. The final conclusion of the presentation was that the infrastructure that's in place is performing very well. It looks like there are no problems whatsoever with scaling to much higher loads than are currently handled and the diversity of nameservers across the region is performing well. The next presentation was from George. This relates to that action item I mentioned before. As I said, the implementation of the lame delegation collecting is almost complete. It's running in semi-production. The only thing that's not really happening is the lame delegations which are persisting are not being suppressed. That's due to be turned on at the end of September. Terry again - George and Terry were oscillating mainly backwards and forwards from the mike during this meeting - discussed the DNS new zone generation system. The zone files relate to registration activity. The performance of the zone generation has been increased from around 40 minutes to around a minute. One of the reasons this is important is because DNSSEC is expected to place quite a processor load on generating those zone files. This provides APNIC with greater head room to be able to handle DNSSEC. George gave a briefing of APNIC coordination work with root servers in the region. APNIC plans to continue to fund root server deployment in the region and plays an important coordination role with all nodes in the region. He spoke of how it can benefit local Internets. George also presented some analysis to do with the implications of IPv6 transport for DNS queries in the reverse DNS. He described various risks that are possible when v6 transport is available from the roots down. His analysis concluded the risks were minimal compared to the benefits for v6 only. The work required at APNIC to provide this service once the delegation chain is available over v6 from the route is zero, since the APNIC nameservers - or some of them at least - already run dual stack. Bill Manning presented an overview of EP.NET deployment. He covered standardisation efforts and talked of the roles of ICANN and others in the rollout of DNSSEC from the root. It was noted that APNIC will have an important role in key management with members, as the allocations of addresses that happened from APNIC delegated space also have corresponding implications in the address under in-addr.arpa. The last presentation we had from Bill contained some of the same material Bill presented in the IPv6 SIG. It was a brief discussion about the introduction of v6 support into the root zone; in particular, AAAA records already exist in the root zone for TLD nameservers and more are being added. Doug Barton confirmed there is a growing number. AAAA records for root nameservers have not yet been added - that is still being discussed. Because the root nameservers are responsible for serving priming inquiries to 20 years of history to DNS clients it's important to get it right and not make changes which could arbitrarily break older equipment. The DNS SIG mailing list has not received very much traffic since the last meeting. I'd like to remind anybody interested in DNS operations that the mailing list is there. You're welcome to use it for anything relating to the DNS. Lastly I'd like to thank the APNIC staff involved in organising the meeting and presenting so many informative presentations during this SIG meeting. That's all. PAUL WILSON: You're very welcome, Joe. Thank you. KAZU YAMAMOTO: This is a report on the IPv technical SIG which was held on 1 September. We have no action items so far so we have 7 informal presentations by six people. The number is more than I expected, so that is very good. Let me summarise each presentation. IPv6 allocation status from George. This covers all RIRs and DB registration and routing tables. He explained that in a while /28 and 39 /32, is space already assigned by APNIC in this year. This is a very good number. The presentation provided by Hosaka San. He talked about IPv /48 assignment status report. He pointed out that only about 8,000 /48 registrations already count. He thinks this is too small. Actually there is more - other assignments on the bandwidth. So please register your assignment if you have one. I also think 800 is too small. There is actually other assignments out there, I guess. The third presentation was by Jordi. He talked about guidelines for ISPs on IPv6 assignment to customers. He pointed out that multiple /32 to ISPs is not convenient. For example,increasing routing tables. So he thinks that shorter prefixes would be better. Fujisaki proposed to relax the old initial allocation policy. Please read that discussion. The next presentation was by Bill Manning. That title is an update on operational recommendations for IPv6 for TLD and other servers. He explained the one name and two machines approach for actual operation. That is very interesting. The next presentation was about IPv6.kr DNS deployment in Korea by Billy. He explained that Korean people finally installed g.dns.kr for IPv6 transport, so we can resolve name and .kr by IPv6 transport. That is very good news. The next presentation is from Geoff Huston, updated on the multiple solution from IETF. Multihoming solutions from IETF. He described the functional goals for multihoming and he explained very interesting approach which is called "ID layer" between IP layer and transport layer. The last presentation is about deployment plans behind larger IPv6 allocations by Jordi again. He pointed out larger assignment behind /32. For example, Vodafone has a /31 and TeliaSonera has a /20. He interviewed with them on the business plans and got an answer. That is very interesting. That's it. PAUL WILSON: Are there any questions regarding the IPv6 SIG? OK. If not, thanks very much, Kazu. APPLAUSE Next up Akinori followed by three presentations by Philip Smith. MAEMURA AKINORI: My name is Akinori Maemura again, on behalf of the chair of the NIR SIG. Finally, for those of you not familiar with NIR SIG, I'd like to explain what is NIR and what is NIR SIG. NIR stands for national Internet registry which is serving foreign country's Internet community. For example in the case of JPNIC, JPNIC serves the Japanese ISPs located in Japan. That is mainly covering the variety of the languages and cultures which are in Asia Pacific region. NIR SIG is for the discussion on the NIR operations and policies. This is my first slide so it's OK. NIR SIG is for the discussion of the NIR operations and policies. Usually the NIR SIG has a lot of informational presentations for instance, producing the operations of each NIR. For example, holding the open policy meeting of NIR and database deployment or something like that. The NIR SIG is open to everyone. So please join if you're interested in NIR issues. This NIR held yesterday, Thursday, September 2 from 4pm to 5.30. It gathered roughly 20 attendees. We have two proposals and four presentations. I am so happy to introduce you about two new co-chairs. One is David Chen from TWNIC and the other is Izumi Okutani from JPNIC. Please stand up. These two are the new co-chairs. APPLAUSE we had two proposals. For the first time for a long time we had two proposals. One is the proposal to abolish redundant charges in IPv6 allocations and the second one is changing the NIR fee structure. I am introducing to you the proposal for changing the per address fee for IPv6. The proposal is to revise a method of calculating IPv6 per address fee so that multiple fees charged for the same address range will be abolished. This proposal reached consensus. I have two slides to illustrate this proposal. Currently for those who are not so familiar with the policy of per address fee. It is the additional charge which is charged only for NIR to adjust the difference of the ordinary APNIC members and NIR members. It is a per address fee, so it is charged by a one host address each. In IPv6 per address fee policy when the one member applies for subsequent allocation, for example one RIR already had new allocation of /32, and if they need a subsequent allocation, additional /32 it makes a /31 and in this case, in the current - according to the current policy the per address fee for this subsequent allocation is the amount of the /31 address block. So this proposal says that this portion of the per address fee is redundant. This proposal proposes the per address fee for the subsequent allocation to be the difference of the per address fee of the /31 and the /32. As I mentioned in the previous slide, and this proposal has reached consensus in the NIR SIG. The second proposal, which is applying an upper limit in per address fee, proposes set per allocation fees charged to NIRs and NIR members at a reasonable level even when large allocations are made to NIR members by setting an upper limit to the fee. Almost all people here are not familiar to this situation. This is a table comparing the per address fee for a certain size of prefixes and also APNIC's annual member fee for such prefixes. Per address fee is the one-off fee. Member fees are the annual recurring fee but in case of the smaller prefixes are held by the NIR, the per address fee is not such a big amount. But if the prefixes get bigger the per address fee significantly increases. For example, in case of the /13 the per address fee are as much as annual fee for the corresponding member fee. In case of a /10 it is much higher - four times higher than the annual member fee. The proposal argues this situation is not so reasonable. We had a big discussion here and had this discussion exhausted all the time. So the discussion was quite active. The discussion is regarding the fairness or the concerns for the financial of APNIC. Also, how to resolve this situation which JPNIC faces, by some possible working around solution. That kind of things are discussed in this agenda. Consensus was that there is no consensus reached for this - no, the original proposal didn't reach a consensus. The consensus was a more fundamental review of NIR fee scheme should be necessary. We had one consensus out of this agenda to say set up a working group to discuss NIR fee scheme with the APNIC secretariat. Actually these two presentations used up all the time of the NIR SIG this time and that is my fault due to my lack of coordination on behalf of the chair, but we would have had four other information informal presentations from CNNIC, TWNIC and JPNIC to present us an update for their operations, and also from KRNIC and we would've had a Whois update project. But these presentations are already available at APNIC web site, so if you have any interest on that, please visit the APNIC web site and see them. I am so sorry to the presenters. Out of this NIR SIG we have the two action items. One is the proposal -022version001, which is changing the per address fee, reaching a consensus. So we should process the other process from this AMM and within the secretariat. There is the other item nir18-002, to set up the working group for the NIR fee structure. That's it, thank you very much. Do you have any questions? PAUL WILSON: Can you bring back the slide on the fee structure? MAERUMA AKINORI: Yes. PAUL WILSON: Are there any questions about Maemura's presentation? There's one consensus from this meeting that involves the IPv6 per address fee structure. This had the consensus of the NIR SIG to change. What we do need to do is assess whether there's consensus from this meeting to change this aspect of the fee structure. I would point out that traditionally major fee structure changes have gone to a full vote of APNIC members, members being the ones directly affected by the fee structure. So in the past when we've made major changes we've gone to a formal vote where we have ballot papers and cast votes in accordance with membership tier. I would suggest in this case - we certainly haven't had sufficient notice or preparation in order to prepare a formal vote on this issue, so I would suggest that we put this to a consensus and through the normal process at which point the APNIC EC will make a decision as to accepting this in their capacity as member representatives on behalf of the membership and they will decide accordingly after the comment period. Are there any questions about that? RAM NARULA: Can you please explain the difference between the current structure and the new proposed structure? I don't quite understand. PAUL WILSON: I'll start at the beginning. The per address fees for NIRs are the one-off fees paid for allocations that go to members of NIRs through the NIR. In the case of IPv4 it's a simple formula where the number of addresses in the block allocated is calculated and the per address fee is applied per address. In the case of IPv6, every allocation of IPv6 is in theory an expansion of a previous allocation and for every allocation that's made the entire - the number of addresses in the whole prefix is calculated again, even if that prefix encloses a previously allocated prefix for which charges have already been paid, the charges are levied on the entire prefix. And this is where, in this diagram there's an illustration of a /32 having been allocated, charges imposed for that, a /31 being allocated and charges are imposed on the entire 31, not just the difference. So that first /32 is paid a second time in addition to the new portion of the allocation. If that is expanded to a /30 subsequently then the entire amount is paid, including this component for the second time or for the third time, I'm sorry. This component for the second time. Is that clear? So there's a double charging that's going on which is arguably unfair and inconsistent with the per address fee structure that's imposed for other allocations. RAM NARULA: Could you please show it in numerical figures? It would be much clearer that way. PAUL WILSON: Izumi had a slide of numerical figures. Let me just point out - I think you're asking good questions and it is difficult when we have SIGs that run in parallel for everyone to be at every one of the SIGs. What we normally try to avoid going through the full justification for particular proposals at this point in the meeting. But I can see it's being set up, so you've asked a reasonable question. I think we have time to look at it. Randy. RANDY BUSH: Beware we don't pierce the fantasy in which we all indulge that what we're paying for is the service of the allocation, not the address space. PAUL WILSON: I think it was in recognition of that principle - I'll call it a principle if you don't mind - that the per address fee was charged on the entire size of the allocation. It's inconsistent with another principle, which is the per address fee is as it's called - a per address fee, which is a special fee that is - a fee that's been imposed for many years on NIR allocations as a one-off fee by consensus decision of the membership community some years ago. Izumi, would you like to explain it? IZUMI OKUTANI: Let's suppose that a particular LIR has requested for /32 allocation initially, then they will be charged for just for a /32 address space, which would be calculated as 7, 132 multiplied by the per address free. And this will be charged for this space and the calculation for this will be... Mmm, it's difficult to explain. (Pause) PAUL WILSON: I have been reminded it's been suggested that I remind the audience that this fee structure applies only to allocations made to the members of NIRs through the NIR system. It does not apply in any way to the conventional normal APNIC allocations to conventional APNIC members. IZUMI OKUTANI: How per address fee is calculated in IPv4, IPv6 allocations are - we look at the HD ratio chart that is listed in IPv6 policy document, and look at the threshold. The equivalent threshold of an allocation. So for example if you received an allocation of /32 then you look at the table of HD ratio and the figure for /32 says 7,132. So you multiply this figure by per address fee, which is in JPNIC's case 0.03. So this is how we calculate per address fee for an allocation you receive. Under the current fee structure - let me just go back to another slide. (Pause) So for /32 allocation it would be 7,132 times per address fee, and when a member has received subsequent allocation of /32 it is only actually this pink address range that this same LIR has received as an additional allocation. However, they will be charged for this whole of /31, so the charge would be 12,417, multiplied by per address fee, whereas it actually should be this 7,132 multiplied by per address fee should be charged. If you add these figures up for the case when an LIR has received /32 to begin with then expanded its space to /31 and finally to /32 the total fee charged will be 41, 168 multiplied by per address fee. However, if you compare this case with a case when an LIR has received an allocation of /30 to begin with, the final address space that these LIRs get at the end is actually the same, however, this LIR will be only be charged for 21, 619 multiplied by per address fee. So even if you're getting the same address space the amount you pay is different. This is shown in the figures here for case 1 and case 2. I don't think I could explain very well, but... is it more clear to you? RAM NARULA: Yeah, much more. I just want to be clear since I would like to exercise my vote. Thank you. PAUL WILSON: Thank you KENNY HUANG: The fee structure proposed here only applies to an NIR member. It's not to the APNIC member? MAEMURA AKINORI: You're right. PAUL WILSON: Thanks very much, Izumi. I think you did shed light on the issue in question. IZUMI OKUTANI: I hope so. PAUL WILSON: Are there any other questions? Any questions on the jabber? No, OK. Let's consider this question then. Could you raise your hand then if you support the proposal to change the per address fee for IPv6. (Pause) Thank you very much. Could you raise your hand again if you disagree with this proposal. OK, thank you very much. That's a consensus. APPLAUSE I think that's it for the NIR SIG, thank you very much. MAEMURA AKINORI: Thank you very much. PAUL WILSON: There was one BOF not listed - the NIR systems BOF is now listed, which our friend Mr Shin will report on. PHILIP SMITH: I have three - I'll be very quick with these - at least I'll try to be very quick. The first one is the routing SIG. That took place yesterday 4pm to 5.30pm. We had four items on the agenda. Tim Griffin talked about BGP wedgies, basically routing policy issues and interactions which could easily be debugged. He gave quite a good presentation on some of the issues there. If you are an operator and you are multihoming with several different providers it's probably worth having a look at the presentation. The second one was by Randy Bush on happy packets. Some initial results from research that he's been doing with others. Basically happy packets are those packets that get from the origin to the intended user without really being mucked around by the network, so Randy has installed BGP beacons and is measuring the fact of withdrawal announcements of prefixes and so forth on these packets. There were some results from that research work. Then to finish off he repeated or updated a presentation he gave at APNIC16 in Seoul about basically BGP is chattier than you think, where the impact of announcing a single - sorry, withdrawing a prefix resulted in about 40 or so updates arriving at the actual destination. So again information about that. The third one was the IPv6 routing status report contributed by Gert Doering. Gert couldn't be with us, unfortunately, so I gave the presentation on his behalf. Gert has been doing this presentation for the IPv6 working group within the European community, the RIPE community, and we invited this presentation here, which we duly did. Again that was an interesting presentation on what's happening on the IPv6 Internet. So if you're working in the IPv6 Internet at the moment it's probably well worth you having a look at that. I should take this opportunity to apologise to the stenographers 'cause I rattled through lots of the numbers and so on probably far too quickly for them to actually follow me but you should actually go and have a look at the stats, they are quite interesting. Finally, Geoff Huston gave another of his BGP status reports and supplemented that with BGP, the movie, complete with opening and closing credits and titles. It's probably the first one that I've seen like this for several years. I think the amount of interest that we saw afterwards, after the SIG closed with people clustered around Geoff and so forth to look at the movie and so on. There's quite an interest in what it shows. It basically shows the history of introduction of address space in the Internet probably from as far back as early 1980s to the current time. It is absolutely fascinating to have a look. We had around about 50 attendees, we started more or less on time. We overran about 15 minutes. I have to thank Chris for minuting the meeting for us. That was the routing SIG. We now have the Internet exchange point SIG - IX SIG. This one took place on Wednesday right after the opening plenary. We started about 15 minutes late, we finished bang on time. It was disappointing in a way that it proved really, really difficult to get anybody to volunteer contributions, even with lots of personal emails to exchange point operators. So I was quite disappointed with - we tried reasonably hard to get people to present but Toshinori Ishii and Gaurab Upadhaya are very reliable contributors and very kindly gave an update on their activities on JPNAP and from Gaurab a lot of the exchanges in South East Asia. Ishii San was asking if there was a comprehensive list of IXs in the region so a prefix list could be built for filters. There was an address block, announced by one of the participants which caused them some operational issues. We had some discussion about whether this list existed and what existed and maybe it could be a job for somebody to put that together. Gaurab came to the APNIC meeting here fresh from Bangladesh where he helped get the Bangladesh IX up and running as the first few participants there. It was a nice presentation on that. He also gave us an update on the NPIX in Nepal. 6 participants, the most recent joinee has been Nepal Telecom. WE also talked a little bit about NIXI, we didn't have a presentation on NIXI but we talked a bit about the National Internet exchange in India. Some of the issues there and also a short discussion about the Vietnamese exchange point - the VNIX, the main Hanoi exchange and the fact that a new one has opened in Ho Chi Minh as well. Gaurab raised the issue of whether IPv6 address block could be assigned to IXs when they receive their IPv4 address block. Some of you may remember this was introduced as an address policy a few APNIC meetings ago but it didn't actually go anywhere. So the suggestion was that maybe this could be reintroduced at a future APNIC meeting. We had about 30 attendees, it was reasonably well attended considering the competition we had from the IPv6 special interest group here. Thanks to Chris, again for minuting that meeting. Finally - APOPS. Just quick information about that - APOPS is a birds of feather meeting of the Asia Pacific operators group, I suppose, for want of a better words. Hideo Ishii look after that. There is a mailing list. If you are not aware of the mailing list we'd encourage you to join and contribute traffic because it has very little traffic apart from routing reports and CIDR reports. We try to have a birds of feather meeting alongside each APNIC meeting. We try not to have too formal an agenda so people should feel held back from contributing, but we had a very fascinating presentation from Kazu Yamamoto about some of the devices available these days are that - that are actually using IPv6. Things that you can actually buy off the shelf. Randy talked about some of the tensions between innovative Internet and security. Gaurab gave us a little update on the INOC-DBA system, basically authentication has been added to that. And the other parting note - update on SANOG, he discussed what we had at SANOG4 in Kathmandu last month and that we've had to move SANOG5 from clashing with the SARC ministers meeting to a new date early in February. Sandwiched between NZNOG in New Zealand and APRICOT in Kyoto. We started late, had about 20 attendees. People felt they had to rush off and catch the bus to the social so we were a bit jammed in between two major events. Apart from that we had a nice meeting, useful meeting, we'd like to thank all the presenters for that. And that is all I have to say. Are there any questions for any of the two SIGs or the APOPS? Nope. That is good. Thank you. PAUL WILSON: Thank you very much, Philip. Three times! APPLAUSE Jong H'ng : while Champika is setting up the laptop for me, let me very quickly, and I wouldn't take very long, this is the report on the NIR training BOF. It's a very informal meeting. It was motivated primarily by a number of successful collaborative training sessions that we've organised in conjunction with the OPM meeting of a number of NIRs on their invitation since 2003 and the beginning of this year. I thought it was good for this APNIC 18 where all the NIRs are present to get all the NIRs together to share experience and comment. That was very interesting - a lot of discussion, a lot of experience sharing. I was particularly interested in determining from all the successful sessions that we've had so far whether there was any merit in setting up a working group among the NIR training coordinators to further explore possibilities of further collaborative effort. So that was the intent on my part. Attendance for the meeting was 14. Like Phillip Smith's APOPs we were caught between a routing workshop which ended a little late and everyone was anxious to go to the social, so there was a little bit of sandwiching between two major events. But the attendance was quite promising, 14, and we have participation from CNNIC, JPNIC and TWNIC. As you can see from this slide, that was the participation we got. The outcome of the meeting was, first, there is no need for a working group to be set up. And the general consensus from the people who were present that the discussion can be done via email or mailing list and this is what we will be continuing to do. The meeting ended at 6:45 - in time for the social. Minutes for the meeting are already on the web site - or will be available on the web site - for those who are interested and for those NIRs who were unable to attend unfortunately. Thank you. PAUL WILSON: Thanks, John. I didn't call for any questions on Phillip's presentation - were there any questions? Oh, I did, did I? Good. At least someone did. This will be the NIR policy process BOF called to discuss common issues between the NIRs in policy development. IZUMI OKUTANI: We held our NIR policy process BOF. This wasn't our pre-planned BOF, but we thought it was a good idea to exchange information about the NIR situation, about how they operate the policy development process in their own country and also exchange information about how they try to collaborate with the APNIC community. It was held yesterday from 6 o'clock to 7 o'clock. We had roughly 15 attendees from NIRs and APNIC. It was quite a cosy small session. It was very good that we could discuss it more freely in this format. We didn't really have presentations. We simply discussed verbally about each other's situation. The situation in each NIR varied quite a lot. Some NIRs currently don't hold open policy meetings. However they are planning to hold one in the future. At this moment, they would like to just focus on trying to implement APNIC policies and while two of the NIRs already hold open policy meetings regularly - and this is JPNIC's case - in JPNIC's case we have a mailing list to have discussions between the meetings, just like APNIC. We also translate APNIC proposals and send them to our mailing list and do the necessary coordination. So even though the situation was different depending on the NIR, we all had common concerns: how to have more participation from our community. In some countries the concern was to have more comments on the proposals and in other countries how to have more proposals coming from the community. But the basic concern was the same. Another thing was how to reflect the voice of our countries' communities into the voice of the APNIC community. In JPNIC's case we use the mailing list for collaboration, but there could be better ways. I think other NIRs are looking for ways to represent their countries' voice in this policy forum. So I think it was a good opportunity to have such discussions and we concluded that we would exchange information through email when ever necessary. Thank you. PAUL WILSON: Thank you. Questions? Thank you. I might just briefly mention about the BOFs, these are informal sessions, Birds of a Feather sessions, sessions that are intended to allow anyone in the group that comes to the meeting to carry out an informal meeting to get together with people who have common interests, taking them as a group. In future, if any of you would like to propose a BOF, please feel free to do so, preferably before the meeting, so meeting rooms can be arranged. But certainly, rooms permitting during the meeting, you can certainly do the same thing - have the opportunity to stand at the podium and report back afterwards and you get to have some interesting discussions. SHIN YAMASAKI: I'm sure we all want to get out as soon as possible so I will try to keep my presentation as brief as possible. NIR system BOF is mainly discussed about IT system issues among NIRs. We had a one-hour meeting on Tuesday. 13 people from APNIC, KRNIC, JPNIC and TWNIC. We had two topics on the agenda - the first one is the NIR perspective of the new DNS generation system from Terry from APNIC. The second topic was APNIC registry system update, from Sanjaya from APNIC. So Terry's presentation was quite similar as Sean in the DNS corporation SIG but forecast for NIR. Since NIRs and APNIC exchange a lot of data that will benefit us a lot and we won't cause any negative impacts. Sanjaya explained the overview and status of APNIC's new registry system as well as XML interface between APNIC system and NIR system. We still need to give a lot of input to APNIC. After those two presentations, we didn't have enough time to discuss other topics and we didn't receive any raised issues from NIR folks, so we're going to use our own mailing list. So that's it - any questions or comments? Thank you very much. PAUL WILSON: That is the last SIG BOF or other report that we've got on the list, but if there are any reports which have been forgotten, please feel free to come and have a talk. Thanks to all the presenters once again for both presenting and chairing, for arranging the sessions in the first place. It's very much appreciated. These sessions are a critical part, in some respects the main part, of the meetings. So it's good to have comprehensive reports and a good range of sessions to hear back on. I'd like to open the floor for anyone who would like to make any suggestion or proposal or request regarding the meeting arising out of what we've heard or otherwise. Moving right along, meetings coming up. Gaurab, did you want to say a few words about your coming SANOG meetings. Gaurab Uphadaya - it will be in Dhaka from 6 February to 20 February. The other will be in Thimphu, Bhutan, July 16 July to 23 July 2005. All of you are welcome to come. You can visit the web site. We have had quite good meetings in the past. So hopefully we will see more people from other parts at this meeting. PAUL WILSON: The address is www.sanog.org. Those meetings will be listed on the APNIC web site, those meetings. The next meeting of APNIC will be held during APRICOT 2005 in Kyoto in Japan. It's a wonderful city. Many things to see there. I'm sure it will be very nice at this time of year as well. It's spring, right? Nearly! Details Of our meeting within the APRICOT program are on the website here. You can see the URL there. There is also APRICOT.net where you can find APRICOT's web site itself. The one to follow that, we don't have a location yet, so we're calling for proposals from today. The call for proposals document which will be posted in the near future will give you all the details of what's expected and what's requested in a proposal to host an APNIC meeting some time around August-September around this time frame in 2005. So please anyone who is interested in assisting with one of those meetings, we do rely on good hosting such as we've seen this time. So you're encouraged, please, to consider that option. The one after that will be - we do know the location of the 2006 meeting - that will be in Bangalore, India, in February-March 2006. That will be the first time APNIC has met in south Asia so we're all looking forward to that as well. OK, that brings us to the end of day 4 and to the end of the APNIC member meeting. I'd like to give a few words of thanks before we run to the beach. We thank very much China Netcom Corporation for offering support as a gold sponsor for this meeting. That's a substantial amount of financial contribution and it's very much appreciated, as are the contributions of three NIRs, all of whom came together, each of them a silver sponsor - CNNIC, KRNIC and TWNIC. Thanks very much for your support. We also should thank all the sponsors who've come to APNIC 18. We count Telecom Fiji and Connect not just as meeting hosts but as special goal sponsors for having contributed financially as well as in kind with staff and facilities and a great deal of time and effort. So they once again are thanked most sincerely for their support. If a representative from Connect, Alfred, is in the room, I'd like to just present a small token of appreciation, Alfred. Thanks. The APNIC graphic artists who generally do a fantastic job with our logos and websites and things have created a certificate of appreciation for hosting the meeting and there is something to really enjoy. It's Australian wine. Thank you very much. Similarly, a representative from Telecom Fiji. Thanks to Telecom Fiji for all of your help. Thank you. The other gold sponsors for the event have been the ITU-T who contributed substantially in funding fellowships to the attendees who came along to the workshops last week. By all accounts it was a considerable contribution to all the expenses of those participants - money well spent, I hope. ITU-T's contribution there is very appreciated. Softbank BB contributed as a silver sponsor on Wednesday and also a silver sponsorship to the fellowship program to assist in bringing people and covering some of the expenses of approved fellows to this meeting. CNC group, of course, today. FINTEL sponsored the international banquet, TransTel the opening reception, Xceed Pasifika the communication systems listed here, the local connectivity, the fibres or the 2MB lines that were brought into the site. So all of those are considered as gold sponsors. Please show them all your appreciation. We thank the silver sponsors - TWNIC, CNNIC, JPNIC, KRNIC and Internet New Zealand. The ongoing contributions of the NIRs are much appreciated. Most years we are supported in this way. I'd like to thank all the SIG chairs who presented and asked you as a group to come up and receive a small gift. So please Arano San, Yamamoto San, Maemura San, Xing Li, Phillip San and Joe. Thank you. Your contributions are invaluable. Thanks. Do you want three, Phillip? You can only use one! The guest speakers who came and presented at the plenary, thanks to all of you. I think Jordi is still here. Jordi San - Jordi is particularly appreciated for stepping in at the last moment, literally, when one of the plenary speakers couldn't come. The stenocaptioners, I'm not sure you had the chance to meet them during the breaks, you wouldn't during the meeting, but Kathleen and Lorraine, from the Australian Caption Centre, who've provided this service to us. APNIC staff have been here working very hard, as I think you've all seen, a fair number of us, but these days there is a fair amount to do - too many to name but I'll give it a go. Vivian, Connie, Save, who've done a lot of work on the local arrangements, George, Terry, Darrin, who set up the technical systems for us, Frank who also was involved there, the documentation minutes team and web team include Gerard and Chris and Yurim. The hostmasters, Son and Elly. There's also Naurine and Rahni and Anne, the trainers, Miwa, Champika and John H'ng - why don't you all stand up? - a couple more there as well. And also to the APNIC EC, our RIR colleagues, who presented - Ray and Axel. And last - but certainly not least - to all of you who attended, more than 130, which is a pretty good number in all, thanks for coming all this way and via quite complex routes in most cases - even from quite nearby locations in some cases. There were flights out of the Pacific to Hawaii and back into the Pacific again and all sorts of things in order to get here. The effort is appreciated. I think it's been a pretty good meeting. For the next APRICOT, in order to encourage you and to fill you in a little bit about the next APRICOT, Maemura San, would you like to say a couple of words about it. This is APRICOT in Kyoto In February 2005 when it will be getting warm and so on, I'm sure. MAEMURA AKINORI: I'm afraid not. Thank you very much. On behalf of APRICOT 2005 local committee, I'd like to give you the warm welcome from Japan. This APRICOT will be maybe our 10 years - so from the beginning of APRICOT, 10 years will have passed. We are now working very hard to make a program. It will be quite a different one from the previous nine meetings. APRICOT was originally programmed for catching up with technology from the US and Europe,but 10 years have passed and right now Asia is the leading region in terms of broadband and other new things in the Internet. So we'd like to make the coming APRICOT 2005 as introducing the cutting edge technology, or new age or whatever, from the Asian region to the world. I think this will be quite an epoch-making effort. We are aiming for that. For the name of Kyoto- I don't think anyone who doesn't know - people know of the Kyoto name. It is the ancient capital of Japan and the most Japanese city in all of Japan. So we have a lot of temples and shrines and kimono girls - that could be inappropriate! Anyway, you can bet that you can enjoy quite the Japanese atmosphere in the Kyoto city. In February, though, Kyoto is very cold. Please prepare for that. Please, though, never fail to come to Kyoto. Thank you very much. PAUL WILSON: One last thing which - well, another thank you, which is to the Sheraton resort here, which has done a lot to help with this meeting, providing the venue, providing a couple of surprises too. Particularly I think we've all seen the multitalented staff, who are going to come back to wish us farewell. A nice surprise has been provided to us by the hotel. So please enjoy. (Fijian choir sing) PAUL WILSON: See you in Kyoto. Meeting ended at 4pm.