These are unedited transcripts and may contain errors.

---

The IPv6 session commenced on the 6th of October, 2009, at 4 p.m. as follows:

DAVID KESSENS: Welcome to the IPv6 Working Group. Please, come in, sit down, we are ready past the official starting time so we'd like to start now, really start meeting now. We have a different, somewhat different set?up this time, somewhat different programme, occasionally I want to experiment a little bit with the agenda do some different things and see whether people like it or not so, let's just start.

The first important thing is to administrative stuff, despite the Working Group chairs every time getting rid of our blue sheet or attendance sheets they keep popping up and I like this /PRA digs so I am going to return them to the audience anyway. We have to verify that we have a describe. Excellent. The first topic on the agenda will be the up dates from the RIPE NCC, they will be fast and quick updates, I understand from James. After that, we will have a little bit of a discussion about future of the Working Group, Shane brought it up in the Working Group mail list and occasionally that is just a good topic to bring up, you know you can always keep continuing on auto pilot but sometimes it's a good idea to reflect a little bit, reflect on what you are doing and whether you need to change stuff and in any case, we looked at the charter and the head phone looks really outdated so it's time to do something about the charter in any case, whatever we decide.

Then recent topic that was brought up in the mailing list, people started discussing about experiments on the RIPE meeting network and I think it's just good if we can talk a little bit about that and whether there is any possibilities or opportunities that we can take advantage of having so many networking engineers in one room, whether we can do something in that area.

Then after that we normally have routing updates from Gert Doering about how does the IPv6 routing table look like, because we are going to have a little bit different meeting this time, we have that available on?line so if you are interested in that topic, we do have it available on?line but Guert is not going to give this update here today. Then finally, and that will actually be the main topic of this agenda, will be a panel discussion, Maarten Botterman has organised that fours and that is follow?up basically on the EU survey, we invited a couple of random individuals, perhaps not entirely random but somewhat random individuals to give us some guidance on the results of that survey but also to see if we can get some input from the audience and again we are going to do things a bit different. Maarten has prepared a few questions in on on?line survey so we can actually as a whole group give some feedback on this topic as opposed to going and stand at the mike and only give a few people the chance to give some feedback. That survey is available already on?line. Don't go and type in what you see here on the screen, it's much easier to go to the agenda web page of the RIPE ?? or the RIPE web page, there is a link there and click on that one and fill out that survey. We have developments initiatives in the RIPE region, we often have some interesting things going on like a conference or somebody wants to test something out or whatever and it's an opportunity for people to announce it. I have had nobody approach me yet but if there is anything then during one of the presentations get to me and I will give you a chance to announce that. And that is basically it for today. So let's go through to the first agenda topic and that will be RIPE NCC update.

JAMES ALDRIDGE: This is going to be a very quick update, I have only about ten slides. What we have been doing in the RIPE NCC v6?related since the last RIPE meeting. Basically /SPHREUTD into two sections, first what we have been doing in the RIPE NCC itself and then very briefly covering IPv6 at RIPE meetings.

I think the last RIPE meeting I said all /SPR?FRS v6 enabled, they still are. Still ?? some are still behind proxy rather than having native support and we are moving towards more native support. Within OPS we are still using an IPv6 assignment from SurfNet, we still see problems not honours the no ex export community. Now we have got a policy that allows to assign resources to ourselves we will probably put in a request for PI assignment for the NCC network. And then move over to using that and hand back the space that we have from SurfNet.

Recently we have within concentrating on improving the Monday /TOFRG our v6 services. We are looking at IPv6 generally improving monitoring. We looked about to deploy v6 NetFlow on the Juniper M7I, requires multi?services PI C to be installed so we didn't have, bought those, as soon as we get back we will be installing them. Hopefully we will have a pretty graph at the next meeting.

Elsewhere in the RIPE NCC the DNS group, K?root now has IPv6 at 10 instances, we have plans to announce v6 from more instances, Tokyo, Frankfurt, over the next few months. Query rate over v6 is between 150 and 200 queries per second which is about one percent of the total query rate: A graph showing the number of queries to K?root over a typical few hours.

Reverse DNS, all the reverse DNS servers in the RIPE NCC are available over IPv6. We see about 600 queries per second, about one percent of the total rate. And another graph.

IPv6 at RIPE meetings: When policy 2009 ? O2 was approved, and published, about 30 seconds I had a request in saying can we have some address space for the RIPE meeting. Finally last Thursday we got the/482001 ? 67 C64 which we are using for this RIPE meeting for the first time. Given it was so soon before the RIPE meeting, we hope that everyone had time to update their filters, but I haven't had any major complaints about v6 connectivity, some people even commenting it's better than in Amsterdam when we use SurfNet space. Got a few figures: V6 traffic peaked very briefly at around 1.3 megabytes per second on yesterday morning. Compared to a more steady 4.4 megabytes per second for v4. And we are going to cover this later in the this Working Group I think. Part of the job of the NCC is to support the network at RIPE meetings so if we are going to do more experiments, let me know and the earlier the better, really. Because it does take us time to make sure we have got the right equipment in place and everything configured. But we have a fixed ? addresses so we don't have to change as much every RIPE meeting this time. And that is about it. Any questions?

DAVID KESSENS: No questions. Thank you, James. James will actually stay here when we have our discussion on network experiments so don't wore /AOERBGS he is not going to leave us.

(Applause) (worry)

And yes, I was actually that person who said IPv6 connectivity etc. And v4 connectivity also was pretty good here so far and actually better than I have experienced before, for whatever that is worth.

The next agenda topic is the future of the Working Group. I have a slide for that, it's a very simple slide. It's really only one slide. The first item is general discussion. Shane on the Working Group, mentioned that he didn't see as much traffic on the mailing list as he would like to see or what he would expect from a Working Group or a RIPE Working Group to do and he want today bring out the topic of whether we still need an IPv6 Working Group, maybe we should continue with that, etc. (Whether) and as a follow?up item if we decide this Working Group is still useful, somebody mentioned that our charter was a little little bit out of date so that is obviously a follow?up item to discuss. So do I see Shane here to open this discussion item?

SHANE KERR: I am here. So this all started because I was looking at Geoff Huston's estimate of when we are going to run?out of v4 space and it occurred to me we only have a couple more RIPE meetings left before we are done with v4, I think it's something like three meetings before the IANA runs out and then one more meeting before the RIPE NCC's addresses run?out. So I got a little panicky because, I consider this a real problem and if we look at like James's presentation about the traffic going to the K?root, is about one percent, I think that is ?? I think that is actually quite typical for v6?enabled DNS servers and I think that is probably the best served corner of the world as far as IPv6 so, you know, everyone has been talking about for years and years there is a real problem and I think that is true. Fortunately, where it's going to solve itself one way or the other, so it occurred to me that the RIPE IPv6 Working Group, I thought ?? before I looked at the charter, I thought that the purpose of the group was to encourage IPv6 and help facilitate its ?? they say facilitate in there somewhere, I think ?? oh, no, anyway I thought the idea was to make IPv6 happen, basically. But it turns out the charter is just a kind of talk about stuff (to) and think about things and look at what is going on and I think the time for that is long since passed. I think, as I said on the mailing list, I think the time has come for blatant advocacy, I think it's time to say "you have to go to IPv6, do it now, it's the future. If you don't do it, you are going to suffer later." And keep saying it again and again. And I think that needs to be the goal. So I think the IPv6 group in its current incarnation is basically wasting everyone's time. I think we would have much better use of our time by pushing IPv6 out into the general world, and we see this as happening already. If you look at the mailing list, the most discussion that has occurred about IPv6 is on the Address Policy list which I think is appropriate and I think that is OK. So, I propose that we shut down the Working Group and make ?? take this time that we are all kind of sitting in here looking at the traffic changing from.01 percent to.015 percent and actually do more interesting and better things. So that is kind of where it came from. And we agreed to talk about it here. So do you want to ?? do you want to talk about this?

DAVID KESSENS: The mikes are open. Anybody has any comments on this? I saw Olaf very close to the mike.

Oh laugh: I was very close to the mike because I was trying to assess whether the direction you were heading had anything to do with this IPv6 Act Now effort?

SHANE KERR: It doesn't although I have to say I really applaud the effort, I think it's great. I think it's kind of a shame that it's necessary, given that we have a Working Group that ?? I thought in the old endays would you see documents like the DNS Working Group produced a set of recommendations for timers in your DNS header ?? SO A so it's sort of useful recommendations. And as far as output it, doesn't seem like we have many RIPE documents coming from the RIPE community any more or other kinds of output so I don't know, and if we look at the IPv6 Working Group mailing list, there is not much traffic, they are not discussing operations, they are not discussing anything, and I don't know if that is OK, just we don't do things on the mailing list any more but we used to, I thought.

ROB BLOKZIJL: I think there are various things that point in the same direction, you observed lack of technical documents coming out of this Working Group or the community at large. The 0 .1 percent traffic moving towards 0 .12 over the course of two RIPE meetings is another indicator of ?? there are a couple of other things. The outcome of the first results of this EU survey which were presented in the previous session, is another example and there are many examples of what I think that the community has not or the industry, not people here in the room, but the industry at large, has not got this message which came out of this community RIPE meetings, the RIPE NCC, their website, public statements we have done in the past, all telling, hey girls and boys, your happy days of IPv4 is almost over and almost went from six years to five years to four years, three years and now we are close to two years. I don't have the impression that the message has been received and digested at a large scale.

SHANE KERR: I agree.

ROB BLOKZIJL: But one day, there will be hit, prepared or in most cases, apparently unprepared by the fact that IPv4 is gone, no more new IPv4. And then will come the big rush not to get IPv6 addresses because that is a very minor detail and it is very easy and very light and nobody is trying to keep them for OK. So it's not the Address Policy, we have done that; it is rolling it out and getting it and not in a way of an experiment by a few well meaning technical people in a dark corner of a large corporate backbone; it is a large roll out towards the end users, that is where it will be hit with lots and lots of operational questions which they might want to discuss in a forum like an IPv6 Working Group of RIPE. So I would strongly suggest not to shut it down, accept the fact that not everybody is moving as fast as we are or we would like to see them, accept the fact that probably between now and two years from now the relatively calm life of this Working Group will continue, but be prepared for a very big bang in about two years from now and maybe in the meantime, we can think about the charter which was written in the days, almost in the days where we are talking not on IPv6 but on IP N G which turned out to be fiction and oh, this is nice and people start playing with it, but it's not really serious because you don't need it for the coming ten years or so. Those days are over so that should be reflected, I think, in the charter. And it is technical coordination that is the cornerstone of RIPE and so apart from observing the flow of messages on the mailing list, I honestly think there is a need for technical coordination, people don't realise it, don't realise it yet, but I think we should be prepared.

SHANE KERR: One of the ?? I think you are right. There is going to be a large ?? I am sorry.

AUDIENCE SPEAKER: I am not really ?? Guert Doering. I am not really sure what to suggest for the Working Group as such. What we have seen is that we have lots of good presentation and operational stuff in the plenary regarding IPv6 and I think this is very important so the Working Group as a point where you could go to learn about v6 is not the central point of the Working Group any more, because that is happening in the plenary, getting reports on v6. I think the Working Group as such is useful to have things like the panel discussion that we are going to have today where we have real inter activity and questions pop up and people work together to find answers, and that may be even produce a RIPE document every now and then. I wouldn't shut down the v6 Working Group but the charter could need a bit of work indeed.

SHANE KERR: I I ??

ROB BLOKZIJL: I want to give one example which I have taken from this one?and?a?half day we are meeting already: Two presentations on implementing IPv6 on user premises and one access for all it, said we have looked at various CPE equipment and this particular model of Cisco box we throughout because it doesn't work and presentation by Portuguese research network and we chose Cisco and it works. That is two presentations in the plenary, fine. I would have hoped for an item on the agenda of the RIPE IPv6 Working Group to have a technical expose say on why the same box doesn't work in one network and works for the same applications in other network. That is I think a typical example of the role of this Working Group, not general presentations on "I have a network and I am playing with IPv6." That is very interesting, but that is for the plenary or if there is no time, we can skip it for the meeting.

SHANE KERR: Like with this presentation from Matt and Joe about signing the route, they gave a general presentation at plenary and they are going to talk about the gory detailed, the detailed stuff later on at the DNS Working Group. I think that is a pretty good model actually. I agree with that you there is going to be a kind of rush of technical questions and things like that later when ?? after the ?? after the addresses are gone, people are going to wake up and say what do you mean there is no more addresses, how can I function? What is going on. And then they are going to start coming. I don't necessarily agree that that means people are going to want IPv6, because they are going to do ?? in reality, people are going to use NAT, right, our C 1980 addresses and all the horrors that entails. Someone made a discussion on the list and I actually think it was a very good one, to have an IPv4 run?out Working Group, so if our goal is to improve and provide a forum where people can talk about their networking problems then maybe v6 isn't the real thing in that people are caring about; they are really caring about keeping their networks running in the future and if that means they have to do four layers of NAT that is is what people are going to talk about and RIPE can function in that capacity as well.

ROB BLOKZIJL: Absolutely. And nothing is ?? I want to reply on this and then ?? nothing is as easy as creating a RIPE Working Group if there is interest in the community. I mean, we have simple mechanism, like, yes, we foresee all sorts of new problems connected with the run?out of IPv4, people can start with finding like?minded people, organise a BoF, even this week, report on Friday morning and say "we think we have enough interest to create a Working Group" which can then be done at the next RIPE meeting. I also see that the RIPE NCC gets close to 1,000 new members per year. That is 1,000 ISPs starting up. This is not going to stop when IPv4 runs out.

SHANE KERR: It might.

ROB BLOKZIJL: I don't think so. I don't think so. Some of them will go deep into nothing; I mean, they will buy on the grey, black, white market, steal IPv4 address space, small chunks and invest in NAT boxes, or they will go for straight IPv6 but that needs to interconnect with the IPv4 world, so a whole new class of technical issues will come up and we should be ready, prepared to have a platform where people can meet and discuss this. And whether that is the current IPv6 Working Group or a slightly remodelled Working Group, I don't care, but I think we should realise that new things are coming.

DAVID KESSENS: OK.

AUDIENCE SPEAKER: I have a comment, this is /RAO*UPL a from RIPE NCC on Jabber. I agree that the Working Group has reached its useful life as a regionally ?? originally envisaged. Needs to evolve and I think through the charter what we need to move away from is repeating of the IPv6 situation in every single meeting. The Working Group needs to move away from repeating of v6 addresses are running out and on to more useful discussions. I also think ?? he will be here tomorrow ?? I also think that the new charter should be inclusive of other industry areas. They are lots of companies out there in non?ISP sectors that need to be involved in this. Thank you.

DAVID KESSENS: I would also like to give a few comments and this is David Kessens and I would like to give a few comments basically from the corridors, the back rooms of the RIPE organising committee, etc.. we have already seen over the years that actually the significance of the IPv6 Working Group has changed. I mean ?? and again that is not always reflected in the charter. But for example, when we started as a Working Group we also spent a lot of time on policy discussions, and that is one of those things that isn't even in the charter, but he spent a lot of time on it anyway, but we also came to a conclusion that it's becoming mainstream and IPv6 policy is now a normal policy thing and at that point we shifted it over and it became part of the Address Policy Working Group. At the same time, if you look back deep in the archives you will actually find IPv6 Working Group had longer time slots and part of that was we didn't give as much attention to IPv6 in the plenary discussions and the EUF committee and that ?? all the RIPE Working Group chairs and some other people, have already kind of pushed certain discussions more towards the plenary programme because it was felt that it was of general interest and not just for people who care about IPv6 or it was more often paper presentation and much less of a discussion item, that it ended up in the plenary programme and not in the IPv6 Working Group and that is just a little bit of background for people that, yes, we actually did evolve quite a bit already in the past few years and things have changed and that happens, you know, sometimes behind closed doors, whether that is good or not, but there is ?? you know a lot of stuff is happening in that area.

Now, of course, we want to get to a conclusion of this kind of discussion. What I am ?? of course I have my personal opinion but I am just trying to see if I can formulate a few questions so we can reach some progress because I don't think, you know, there is no point in discussing this for like hours and hours. We also have our other interesting topics for the Working Group. So one of the questions that I think we should ask is simply: Is there a point in continuing the IPv6 Working Group in one way or ?? in one form and then we can discuss in what kind of forum after that (forum) I think the first question, we have to ask that and see what people are thinking about that.

AUDIENCE SPEAKER: Just to say I am more or less on the same line than Rob, what he said, which is very unusual but sometimes it happens. If we are about to re charter the group, I would suggest to produce on trainings because a lot of people that needs training, operational training I would say, so I guess it was more or less a place in the EUF Working Groups, but now this U F has vanished so where in the trainingings can be held, I don't know. And the other thing is what you were suggesting about technical questions; as soon as people are about to deploy, they will face problems and who to ask these problems, to have a solution, and we need probably to focus on that as well.

SHANE KERR: I guess before we ask if it's useful to continue the v6 Working Group, I would like people to keep in mind that wave limited amount of time (we have a) and I mean, within a RIPE meeting, and that we have a lot of various interests, so, I think when you think about what activities you want to do it's not simply a matter of whether or not you find something useful or not; it's is it more useful than the other activities you could be doing? And that is kind of where I am coming from. And I don't disagree that it's possible that a v6 Working Group is a very useful activity, but if ?? but we need to make sure that happens.

DAVID KESSENS: Final comment on this one and then we are going to ask the question.

AUDIENCE SPEAKER: From /TKPWA*R. Another comment that is in the ?? about decide this part of the discussion because I am just arrive and when enter from the door I hear the IPv6 Working Group close. I think if I was a journalist, my idea is that RIPE close with IPv6, and so is a bad message, is a wrong message and so I think this is a point to consider about the life of the IPv6 Working Group. I agree with Rob that we have to change, we have to change with a ? training like Bernhard said but also we have to move from something that is an island for technicians and something that is open and that is open also for stupid errors or for simple ?? for easy configuration or for all things that are to ?? for people that are starting with IPv6, and one to discuss about what are work arounds for DNS or something else, that it's about IPv6.

DAVID KESSENS: OK. So let's start with asking this question: Should the IPv6 Working Group continue, one way or another way? And the other way we can discuss later.

ROB BLOKZIJL: I think the question is wrong. It's too complicated and too open so cut it into two questions: Question one is should the IPv6 Working Group continue as it is ?? has been? I think the answer to that, what I hear from the discussion, is no. So that is clear.

Second question: What should it be? Oh, yeah it's very efficient.

DAVID KESSENS: As I chair I do have my suspicion abouts certain questions and answers but I do have to agree with Bernhard that we officially have to ask this question and I am fine with splitting it in two pieces so let's ask this question: Should the IPv6 Working Group continue in its current form? Yes? How many people are for yes. Just a show of hands. We are not going to count the hands, we are going to look for the clear majority for one or the other.

AUDIENCE SPEAKER: Current form means quasi?plenary. We had in the bad old times we had actually parallel sessions.

SHANE KERR: It means doing what we are doing today.

DAVID KESSENS: We can fill in those details after it's clear whether we have to change the direction, yes or no. So, are there people who are saying that we should continue the way we are currently operating, continuing on with the current, charter, etc.? Is there anybody who believes that we should continue like that? And how many people are looking for change? (Show of hands) you see, clear answer. You don't have to count.

ROB BLOKZIJL: So, this is Rob, private citizen. I think we should be very modest in our aims and we should not try to reach the end goal without passing some intermediate steps. If we look at the development of the deployment of IPv6 and what Shane mentioned, that there will be some people who will say, well, not in my lifetime, I buy a couple of net boxes and daisy chain them and it might work or not, so why don't we try to focus on this world of what the Internet actually will be, a combination of IPv4 and IPv6 and it has to work together. I think that is where the real problems will come. Nothing is easier than having a pure but isolated IPv6 network that is not where the problems arise. It is where you want to have your IPv6 bit in the midst of this IPv4 ocean, that is where the problems come so why don't we first focus on that ?? not focus, we should not forget that that is part of the real world and should be covered by the re designed Working Group.

DAVID KESSENS: Any comments on that? No. I noticed a few distinct options. I mean, I heard something, we have to do more outreach in the forum, nearly propaganda ??

SHANE KERR: Can I interrupt for a second. I don't think we need to design a new charter on the stage here today, so I don't know what the traditional way of handling this situation is in RIPE; is it with a task force or done on the mailing list or perhaps tonight at the party?

ROB BLOKZIJL: I think the last option is is a very valuable one. The traditional way of doing this is you don't draft papers by committee, that is the most inefficient way. So you find one, two or maximum three, people, who say yes, I am very interested in this, I am committed to make it successful and I will write a draft for a new charter. And so, the call is for volunteers which is always easy because nobody volunteers and then we appoint a volunteer and that is somebody who voluntarily already spent the last half hour on the stage could be one of the volunteers. You might find a mate who can help you and the current chairman should be interested as well and then we have almost too many people already.

DAVID KESSENS: But one step first. I mean this is mechanics. There are a few things that I think are important to sort out and that is a bit of the direction to give the task force or whoever is going to write it, because just writing an updating charter in the air is is a little bit difficult. So one clear direction that Rob is suggesting is preparing for hybrid world. But I heard earlier was also do more outreach work in terms of pushing hard and warning people about IPv4 /KPUL /SHUPB, another topic was education/training, that kind of direction. I think it's useful if we just get a quick show of hands and this is not like a final determination because if we have like a proposed charter people can still vote things down or say it's a really bad idea but some general direction for a possible task force to work at or to go to. So, let's first ?? too many questions. Three comments and then we close the mikes because we also have to get moving, so first Martin.

AUDIENCE SPEAKER: Basically, I mean if you see this presence it's not like nobody wants to talk about IPv6, that is not the /SH*UFPLT /KWROPBLG the issue should be whether there is IPv6 Working Group or not but where you have IPv6 outreach Working Group or IPv6 issues workshop and you can see it becoming more dominant in terms of Working Groups as well, each for a specific focus. The training was takes place yesterday morning for ?, so it doesn't all need to be in one Working Group if you talk about IP version 6.

AUDIENCE SPEAKER: Private citizen as the term goes by now, I think. As much as I think we should continue to focus on implementation issues on IPv6 I would like to strongly reinforce what Rob said and start looking at how the hybrid world looks like and focus the Working Group in that direction.

DAVID KESSENS: OK. Let me ask that question: Should one work item in the charter be preparing for a hybrid IPv4/IPv6 world. Is that OK question to ask? Yes we should do that. That is quite a few people. But not as many as in the first question. Is there anybody who says no we should not have that as a work item? So it's pretty clear that there is consensus to have this as a work item.

Then this more outreach, pushing harder. I am struggling a little bit to formulate a good word as ?? possible work item there.

SHANE KERR: I would suggest that we are designing the charter on stage again. And we should ??

DAVID KESSENS: No, I think there is a few items we have to do here, otherwise no direction for the charter ??

SHANE KERR: I think we have had discussion that and can guide our direction.

ROB BLOKZIJL: A point of organisation. I think we are nearly at this stage, we can close this discussion if we have agreement from Shane that he will drive a draft, the writing and editing of a draft charter.

SHANE KERR: OK.

ROB BLOKZIJL: Thank you.

(Applause)

ROB BLOKZIJL: And not everybody who has an interest in the future IPv6 Working Group is present here so it will be on the mailing list.

DAVID KESSENS: Sorry about this, one question and I do think it's important to that first, and that is: Because there is technical focus and non?technical focus and that is really what this topic is about, do we need to do more on outreach activities and stuff like that and I think that is an important question and something that we want to give to the ? to write that up, whether that is useful or not to spend more time. I think that is a valid question to ask the audience here. And as the chair, I am asking that to the audience, Rob. Sorry.

ROB BLOKZIJL: As a part of the audience, I have a comment here.

DAVID KESSENS: That is perfectly fine.

ROB BLOKZIJL: Outreach is is a very fussy thing and I don't think Working Group is the place to do outreach. I think the RIPE NCC is much more suited and your website is an example of that. Training, now, I am speaking as chairman of RIPE, but I have a board member next to me. Training is a very touchy issue. We have, among our 6,000 members who do IPv6 training courses for a living, we cannot be seen competing with them.

DAVID KESSENS: OK.

AUDIENCE SPEAKER: I have three comments from three different people and I would like to add something myself. First of all, I have a comment from Marco /SR*ERT from Ben consult his comment I believe the focus for the new charter should be how to encourage conservation of the IPv4 and deployment of IPv6 because even if we get every ?? we are going to have the hybrid IPv4/6 world for many more years, no can drop IPv4 support and use only IPv6 so conservation is still one of the maimer issues.

Then I have another comment from Tony /HA*EUPB, and his comment is the real question about the charter is what is unique and not part of RIPE business as usual. Those items that are leading RIPE should be covered.

And then I just want today ask as my own capacity as training manager at the RIPE N C: Indeed we are doing IPv6 training but as Rob mentioned before, we have to be quite careful to not compete with our members, so our training is a little bit more administrative orientated and not so technical. I would encourage you to come and check one of our training courses out, free for our members and give any feedback or ideas or maybe bring more suggestions to the Working Group if you think more training should be done for the community. Thank you.

DAVID KESSENS: Last comment.

AUDIENCE SPEAKER: From IS production in France. I have a question as a newcomer: I wanted to know, it seems from the numbers in the budget of RIPE that the extra large category here in RIPE is managing, that means 67 companies are managing more than 50% of the IP addresses in RIPE. I would like to ?? but it's not clear because I don't have the exact number in the RIPE numbers; we only have the score, you know the score what is used to do the pricing, so I would like to know, maybe the first question would be would be first doing a real assessment of the IP number used by each category in RIPE and if we only have to persuade 67 numbers to go IPv6 to cover more than 50 ?? 50% of the IP address, it will be easy.

SHANE KERR: That is an interesting suggestion.

AUDIENCE SPEAKER: Second, it's like a comment that was in the remote audience; maybe we should work on the better management of IPv4 addresses, because I ?? if I understand well, the extra large numbers are using IPv4 as a free /PWUFT, they don't pay ??

DAVID KESSENS: I think ?? we don't have enough time kind of situation so let's first ask the question about outreach and then we are going to get a task force or whatever done and move on to the next agenda item. So who would like future IPv6 Working Group to be more involved with outreach activities, basically non?technical activities to promote IPv6 use in the European region or the Europe ?? sorry, the RIPE region, I said that wrong? Nobody.

SHANE KERR: I think it's like saying who is in favour of education and safety, well I am in favour that have of course, we need more of that stuff. I think we need more concrete ideas about how that can happen and the real way I would hope we are going to do is by activity trying to come up with more ideas, like this one that this gentleman suggested about talking to the top 30 ISPs which control have this space and maybe we can get a lot of progress that way. These are the kind of ideas that I think we can go very far with and people have made similar ?? not similar, other very clever suggestions on the mailing list.

DAVID KESSENS: I didn't see a lot of, I mean on this question, I didn't see a lot of support of straying too far from the technical world. Is there people who would specifically say like we shouldn't focus too much on outreach; we should stay close to technical issues that are related to, for example, in hybrid IPv4 and IPv6 Internet and related issues?

ROB BLOKZIJL: On a general remark. A charter is not a work plan for the next 12 months, so don't go into too much of these details. And don't let text slip in that would later prevent you from doing things that you have not thought of but are useful for a community.

SHANE KERR: OK.

MIRJAM KUHNE: It's a bit of a follow?up question. I am kind of new to this now but when I looked through the Working Group chair's list on the website, I noticed that you are the only poor chair who is still chairing a Working Group on your own; everybody else has Co. chairs in the meantime and one or two in some instances, I wonder if they could also be taking into account, when re chart erring the group if it would be useful to have a co?chair to with you for the responsibility of chairing this group.

DAVID KESSENS: That can always be considered.

MIRJAM KUHNE: I hope this can be considered by the audience and not just by the chair.

DAVID KESSENS: Let's first get back to the charter. So, forming a task form, is there anybody who objects against that idea? It's nothing to do with who is going to be a chair or whatever,

ROB BLOKZIJL: I thought we had just concluded this. Shane volunteered to do this activity.

SHANE KERR: I volunteered.

ROB BLOKZIJL: Does the Working Group agree with this?

(Applause)

SHANE KERR: Anyone else who would like to participate, please just talk to me and I will do whatever needs to be done.

DAVID KESSENS: Sorry, Rob, but as chairman of this Working Group, I am happy to ask that kind of questions. It's nice that you interfere with that duty and I appreciate your help but there is also a point where I can ask questions like that, sorry. So a task force, Shane is willing to do this. Anybody else wants to help? Bernhard. Marco. I am willing to help. By when should we have some result ready? And I have to ask Rob there because is there before plenary time at some point this week or what is the best way to do that?

ROB BLOKZIJL: No, I think we should give the people who are not here this week ample time to ?? it is an important exercise so we should not rush it through this week. I think at the plenary we should report this activity has started and I think it would be brilliant if we have agreed new charter and better understanding of how the Working Group will organise itself by the next RIPE meeting so it can be reported at the next RIPE meeting. And put in place, if the Working Group feels that in, let's say, in two he three, months from now, at that stage has been reached, we use mailing lists to take decisions but not later than the next RIPE meeting.

SHANE KERR: Just a clarifying question: Do you mean that we should be operating under a new charter if he next RIPE meeting? That is what I would like to see.

ROB BLOKZIJL: That is what I would like to see, yes, and then prepare the first meeting of the refreshed Working Group for the next RIPE meeting.

SHANE KERR: Sounds good.

DAVID KESSENS: There is no real reason why we cannot finish something like that on the mail list and in fact it's the preferred model in many ways.

ROB BLOKZIJL: We might use this Internet thing to progress this work.

DAVID KESSENS: So anybody who isatology volunteer, please show up after this meeting here so write down your names, e?mail addresses, etc., so we get this done.

Next agenda item ?? Kerr Kerr thank you very much, David.

(Applause)

DAVID KESSENS: And thank you, Shane. So next agenda item, we don't have a whole lot of time to discuss it. Perhaps we could even also move that to the mailing list at some point or form a small group, or whatever, to think about issues like that, but recently we had on the mailing list we had some discussion about experiments that could be done on the RIPE meeting network to see if there is anything that we could learn from IPv6 deployment in that network or the opposite, learn from IPv4 net deployment in that same network. I don't know if there is any ideas in that area or things that people say, hey, we really should think a little bit more about that because we have, you know, a large group of technical people present, we have a network, we have people who care about IPv6, considering the number of people in this room, so why don't we do something in that area and I am not saying we should do things like disabling the network or something like that or forcing people to use IPv6, because that is fairly counterproductive, but do people believe we should be thinking about a future experiment again in a different way than what we have done before like an IPv6?only hour?

Gert Doering: I would like to actually disagree with you. I think forcing v6 on people is a must because if one of these certainties at this point in time is not able to reach his home network over v6, we are all doomed, because this is the crowd that actually does v6, so everybody else is lagging behind. As for the specifics, I could imagine something like having a v6 only network available, again for the whole week and for a given point in time for well?defined half hour, one hour, or so, switch off the v4 on the main network, no big reconstruction, no big outage, just stop routing v4 and and see how far you can get with v6 only.

DAVID KESSENS: Yes, so one thing I have is that we already did an IPv6 hour, for example. And I think it's actually if, we do an experiment I think we should do something different. Again that is a personal opinion, this is not like as a chairperson talking to you.

Gert Doering: We did that experiment but the scenario has changed since Berlin. There is new operating systems get rolled out, new versions of the stuff you have on your laptop guess rolled out and people have homework in the network so the scenario might give a different result. I wouldn't actually run NAT/PT this time N Berlin we did. I would run v6 /WOEPBL no NAT. Either you have v6 or not.

DAVID KESSENS: So there is another thing that I discussed that actually with James right after lunch, there is another topic that I thought was fairly interesting that is another thought is actually, not so much testing v6 but actually doing the opposite and testing and just seeing with your own eyes how NAT will perform if do you it in real constrained situation so a situation where there is actually lots of users and very few IP addresses available for the NAT box, etc.. one of the interesting things that I think there is, for example, many people ?? many of you probably have seen this example of a Google maps running behind NAT box where some of the tiles in the maps don't load because of resource constraint issues on the NAT box and that might actually be interesting for some people to see this and experience this in real life as opposed to seeing this in presentations, etc., but that is just another idea. I see somebody in the audience shaking and say no, but ?? Marco

Marco: I agree. Instead of having the v6 only network, having a network which is strictly netted with one on 8 IP addresses or something, that would be useful. Regarding Gert's comments and things I have seen on the mailing list to make the meeting v6 only and keep in mind that we are largely dependent on sponsors to give us decent connectivity at meetings and although IPv6 connectivity is more or less with /KPHOD good you may encounter places you have to tunnel stuff to get v6 at the meeting. It happened before and it might happen again. And there is a risk there, so actually by going v6 only we are actually telling our sponsors yes you are welcome but you have to do v6, yeah that might abmessage to give ?? you might miss out on particular deals or places to actually hold a meeting and trying to outreach to all parts of the region, not only Western Europe where v6 is around the corner.

DAVID KESSENS: So I am going to close the mikes. Everybody who is at the mike now can still say what they want but after that we simply don't have more time.

AUDIENCE SPEAKER: Dave Wilson, /H?FPL A E net, I can we should rerun the IPv6 hour from Berlin, in the spare /EUT of what I was saying yesterday I have different expectations for it, from what we did last time /AFRPBLTD I think it was worthwhile last time around and it gave us an idea of what is broken. This time, I would like to do exactly the same thing, personally I'd keep NAT/PT but we can argue over that over lunch but out of it I would like to get a very clear detailed picture of what we expect to work, what doesn't work and what we need to do to fix it. Those ?? that specific /TPH?FBGS is what I need I think at this point to move forward in my own network.

DAVID KESSENS: OK.

AUDIENCE SPEAKER: From J.P. nick from Japan and the situation in Japan and Europe might be a little bit different so it might not be necessarily applicable but just for information about our situation, we are trying to consider what would be the specific issues when the existing IPv4 network coexist with IPv4 ?? v6, excuse me ?? and since not all of the networks can do the experiments, we try to do, we create test bed for each case, cable TV providers, ISPs, data centre providers and test what would be the issues on our v4 and v6 coexistence and then publish it and share with the rest of the community, and another thing we are trying to do in parallel is consider other technical ways of addressing the exhaustion such as various NATS tech /TPHOPLTS,atology 444 so that ISPs a way of what would be the best choice on how to implement their network to face v4 exhaustion, just for information.

DAVID KESSENS: Thank you. Let me ask a very quick question because we are running out of time: Do people believe it's useful to rerun an IPv6 only hour with a condition that we have a more clearly laid out test plan and see what we want to learn out of that? Do people want to rerun that? (Show of hands) is there people not interested in doing that and it's just a bad idea or for other reasons would like to do something else, for example, whatever?

OK. Dave, could we perhaps interest you to work out a plan with James, does that work, James? People in the Working Group agree with that? No objections? Thank you very much. Then it's time for our panelists, please come to the podium. There is not a whole lot of chairs here but we can probably fix that in short order. Bring a chair with you, for example, that seems like one good option. So again that is little bit different approach than what we have done in other IPv6 /W?PS. Usually we talk technology, sometimes I think it's also good to reflect back a little bit on you know what is going well, what is not going so well, since we have had this survey done by the ?? for the EU regarding IPv6 deployments, I got an /TPRAUFR Maarten to actually get a panel together and discuss these issues here so you know why not, let's do it, it's a bit different from what we normally do but a bit of experimenting is always health /AOEFPLT

One good thing that Maarten proposes /S?D to ask some on?line feedback from everybody in the room. Normally we always get people standing at the mike, you just hear from a few people, but instead Maarten proposesed, of course not replacing this, people can still come a mike and have comments, let's see if he can get some general feedback out of the audience by doing an on?line survey and we hope you will all participate in that or already participated in that. That survey is available on the, from a link on the agenda web page on the RIPE website, it's very easy to find. Go to the agendas and you will see it there. Please fill it out and then using the survey we have a few short presentations, not everybody has a presentation but let's start the panel and just enjoy and feel free to interrupt, ask questions or whatever. Final remark, if you don't get thrown out of this room I do not have a problem if we run over time a little bit. And that is ?? again that is one of those things we used to actually plan the IPv6 Working Group a bit at the end so we have the ability to be a bit more flexible with time, especially of course if there is a lively discussion going on.

MAARTEN BOTTERMAN: Thank you. Well, thank you all for being willing to join me here. This is the panel that will help to focus the discussion, they were recommended to me as an expert panel who talk what they talk about, so but next to that, very much we will ask you what you think about it, already seven people found the link on the agenda today towards the question fair and we have three questions and upon the responses we will try to live go to the panel and have their input on what that means.

A very short introduction for those who miss it had this morning or late they are afternoon, we did look at the survey both in the RIPE region and APNIC region, from the Respondents they are actually quite active in IP version 6 compared to the rest of the world. On the 33/37 percent was not doing anything with IP version 6 of the Respondents. Main drivers was really to be ahead of the game and make sure IP version 6 is ready to be supported by those people. Why not considering because there is some of the people who responded don't consider it yet mainly because the urgency is not there yet although there is some people who run into problems either technically or with support from the infrastructure or even the ex /PEPBLGSs not being justified.

And we'd like to go a little bit deeper under the surface here, that is why we are here. The hurdles in vendor support and cost is something that the panel may comment on and the technical problems as such that have been experienced, as well want to see the economics behind getting things started.

So, we also collected prior to this is statements of what people say. Now, in the preparation of this panel discussion, already had some people say let's stop talking and let's really look at what is going on and nevertheless these things have been said not just by people on the street but by all kind of people. First of all, we have heard we don't need to deploy IP version 6 because, for instance, IV version 6 is not the ultimate solution and hold out for IP 4 until better solution is available. My connection works, why should I use IPv6? I am not in trouble I am already served. People say NAT will do the trick and people do say that out there, maybe not in this room but they do say that. And IP version 6 doesn't provide backwards /KPARTable with IPv4 and we have heard things about IPv6 and IPv4 translation and the other way around here, that people out there haven't heard. So I want to give you this of people out there because it's not only us here who are making it happen; we depend on the wider community beyond this.

The second group thinks very strongly that deployment is cumbersome. We saw one of the big initiatives was Google who first tried to do IPv6 on its website and you can still do it but have to go to IPv6 dot google.com and people say that is cumbersome. I would say this is transition period. It works, I get ?? why should I get myself in trouble? I don't know how, I think you raised that. There are still many people out there who have no idea what it's really about. They haven't seen the good practice examples. Also, in those who consult in this, there is not a lot of availability yet of really knowledgeable people who are actually done it, experience is still little. What we know from those who follow the IETF discussions as well is that the transition mechanism, translation mechanisms aren't ready yet, and widely heard thing, if you talk to small or medium businesses or government agencies, is ISPs don't support it yet. And a bit more important argument that comes across and it really stop people from considering this is there is a security threat there. You have heard those arguments as well I am sure and these arguments go as far as even official reports from IETF who says that IPv6 addresses entering is much more error prone than entering IPv4 addresses. Things like this come across everywhere.

So, these are some of the falsehoods I want to take you out of the things that takes place in this room and I would like to ask our panel members to give a short consider of what they consider to be the key things that keeps IPv6 from deploying and what needs to happen very shortly.

After these five minutes statements, we will go towards the questions that you will find on the agenda under the heading "take the survey." So panel members. Who may I invite to talk first.

GERT DOERING: Just as a matter of introduction, my name is Gert Doering. Working for small ISP and have been interested in IPv6 since very long time, mostly from a technical background because I am just sick of haggling with customers over/29s, 28s, 27s, if they just want enough addresses. The customer doesn't care how many addresses he gets as long as it is enough and v6 is going to do that.

Regarding the statements why people cannot deploy IPv6; all these statements are true, actually, in a way.

AUDIENCE SPEAKER: No.

GERT DOERING: Yes. You can do everything ? if you are willing to pay the price. And the thing is really what sort of Internet do we want to have in five years. Do we want to have an Internet that works like television with broadcasters and everybody else sitting behind a NAT and it's just a consumer? Some people certainly want that. Or do we want to have an Internet where people can communicate with other people? "I want that." So this is why I think direct addressing of end devices is important and regarding the major hurdles, I think well, it's the big large scale ISPs that really need to roll out v6 to the end user without waiting for end user demand because these sorts of end users are not going to ask for it. They will happily use it if they have v6 because they don't know what it is, they will just use whatever Internet you give to them.

MAARTEN BOTTERMAN: . Thank you I will give you an opportunity to react, in order to keep the speed of things going, I would like to have the three views rolled out first.

GEOFF HUSTON: Flight crash investigators have an amazing job, you know, planes, are fan /TA*Sically complex systems, and when a plane falls out of the sky, the basic theory is it fell out not by accident. There is a reason, and their job is to find it. So, you kind of look at the situation we are in now and you think, is this an accident that one of the largest activities on the planet and one of the more valuable industries has managed to wedge itself so badly, is that an accident or are there basic causes here? And if we understood why we are in the situation we are in, would it mean that our next steps are a little easier to understand? So, I am not an economist and I actually am playing one, so watch out, I am an amateur, I am wondering if the reasons we are where we are are economic in nature. I was struck about by this quote, it comes from down in the southern hem see if again, New Zealand minister for communications, Internet was a /TRAO*EUL child of deregulation in the late 80s, early 90s, vibrant competition, fantastic innovation and the old Telco style of networkinging was destroyed by the Internet because of competition, intense vibrant competition and deregulation and he is saying here and many of you I think believe it and many policy makers believe it, that this next issue we face with v6, we only need to repeat what we did before: Have courage, have faith, competition will solve this, that intense vibrant competition will see us through. The private sector has the problem and they will see it through. I am wondering is he right, because the situation we are in looks pretty /WAOERLD. Remember that one you saw this morning, the whole idea as the Internet grows green, the v4 address pool, blue, /SH*EUPBGS, but we are clever folk, we are planners and professionals, the red, v6, will come up and just before the blue hits the bottom, we will have managed to create a v6 Internet. Fantastic. So let's take this theoretical view and put some time lines on this. So if we start at 2004, no real v6, less than one percent. 2006, less than one percent. 2008, less than one percent. And we are pretty sure that you guys are going to hit the blue line to the bottom in 2011, so you have got to get from less than one percent to not just 100 percent but 110 because the Internet will be bigger in two years' time. You have got to get there in under two years. You are professionals, you can do this. Or not. Because what is more likely with all those 100s of millions of routers, all those lines of code that need to change, all those little bits and pieces, is that you are not going to be running dual stack over the next eight years, with a fractured stack, the Internet will still try and grow, the green line will waiver a bit, the blue line will hit the bottom in two years' time and somehow you have got to make that red line go up. That is hard. It's hard for anyone.

So, we didn't do it when we could. That is tough. And now we are going to double the pain;, you know, give us a real challenge, fine. Do it with dual stack without any v4 addresses. Feeling brave yet? So this is /WAOERLD. Because what it seems that short?term self interest and long?term common good are different. We are where we are because individual industry players are following their own short?term self interest and the long?term interest of all of us are different. And that, to me, is market failure, where the supply outcome is inefficient, self interest doesn't lead us to where we want to go to. So you go into the economic textbooks or if you are just feeling lucky you go to Wikipedia full out a few definitions, v6 transition as a transition is non?excludable, I can /TRAPB circumstances you can later it makes no difference and non?/REUFRL /RUS, it doesn't stop you and exclude you from doing it and all of a sudden then you start to think that this actually isn't a normal good; it actually has a whole bunch of properties that are more like public good and conventional market dynamics fail. They always will in a public good situation. And the corollary is you have to look elsewhere if you want this transition to fire because what it really says is that if this transition has all the characteristics of a public good, then at some point you have to go into a solution book that includes public good solutions. Assurance contracts, we all say to ourselves "all right, no mucking around, nerve this room signs a /KPHAPB contract, we will all do v6." Crazy in solutions which is a similar kind of thing, go into government enterprising solutions where the government say we will order v6 as a mandatory, that will kick the industry on. I am old, I am too old, I remember gossip. What didn't work 20 years ago probably has even less chance of working now. So, do we do the thing we did with digital television, on this date your carry a licence depends on v6? Or with tax incentives: If you deploy v6 you can write down the deappreciate ration on your equipment double the normal rate. You can move your equipment around. Do we give consumers tickets to buy new CPE worth X 100 dollars redeemable. There are many remedies to a public good problem. It doesn't necessarily have to draw a line in the sand but what it does say is if this /TRAOUBL is is a public good problem then the solution doesn't lie slowly sole flee a private competitive market, it does involve a public solution. Thank you.

MAARTEN BOTTERMAN: Thank you for that. I am not sure you were always so supportive to governments playing a role in this area.

KURT LINDQVIST: I work for NetNod, I am really tempted to go after Geoff presentation but I promise to say ?? so I work for a company that run the exchange point in Sweden and one of the famous 13 DNS route servers and rerun DNS serve services for a number of TLDs around the world. We have done the v6 deployment for us and David asked me this morning, saying it would be good if we if I could present we are non?for profit so commercial drivers doesn't apply. And that is not true. If you think of commercial drivers maybe not necessarily as monetary but in having or providing ?? meeting demands of our customers, that was clear requirement both from exchange point customers and to get v6 services which is why we have done this and why we deployed this and one thing that occurred ?? that was our we wanted to be quite early on doing this because we wanted to have our v6 services work exactly the same way as v4, treat them the same way and monitor and have the same operationalry tunes as for v4. What we did discover a lot of this, maybe operational functionality or routines wasn't available and we believe one of the benefits of being one of the early adopters we can hash out a lot of things out and work through a lot of issues before Geoff's nice crash happens where sort of we don't have ?? worry at the same time. At least not for the same reason. So that was that was made us do this adoption. The other thing is important, coming back to the slides before; you often hear this about the cost of deploying v6. For us I think we had, I would say shorter time we spent doing the configuration, wasn't that long, we didn't do any investment we wouldn't have doing otherwise. There was no additional cost. I do realise you can't translate our scenario into all the other networks but we didn't have any additional cost of doing this, this was part of our upgrade cycles and part of the normal purchasing we would do. And coming back to the survey, I do have a bit of issue with treating all networks the same, because in Geoff's crash I don't think all of you in here can probably get v6 deployment two years if you know you are going to go bankrupt otherwise, the millions of DSL users out there can't or won't or doesn't understand.

MAARTEN BOTTERMAN: Thank you very much. For you the last chance to come into the survey itself, the first question that you will find on the survey is about doing it. Like for you and the peers like you, what would you gain by postponing or what would you gain by rather stepping up to the plate earlier than later? And the options there are really to postpone to save costs now or to wait until your vendor is ready. Or you believe that you will never be able to make that switch. Or by stepping up either to do that because then you have to time a space to make that transition. For people here that is still possible. Or to really support your brand's image and attract customers or maybe prevent that others attract your customers. Could you please cast your votes then and I will show the results in five seconds from now. Anybody still thinking about it? Still responding, OK.

GERT DOERING: We seem to have 25 answers so far. Please stop reading your e?mail and answer the survey. 39 now. So we are moving forward.

MAARTEN BOTTERMAN: For the first question, we close the thing. We take the numbers. I will show you the outcome. What you will see is that the postponers, one post /POEPBS because of to save costs; one to be able to rely on vendors and one postponers will never switch to IPv6. And the bottom part is stepping up to the plate, so I have time and space to do the transition, is the highest score, together with the stepping up to the plate and really confirm our brand's image as state?of?the?art and about half that number of votes is for attracting customers. So from 40 votes, 10 say none of the BoF and I would like to hear your comments on what you mean none of the above, but I first wanted to give the panel the opportunity to respond to this score.

KURT LINDQVIST: A bit curious about the person who would never go to v6. But besides that, I don't think there is anything really surprising in this. The people who are doing it now to say, operational ready when it happens, I think that is what we are seeing. I don't think there is many of you in this room who doesn't have plans to have looked at v6 and know exactly what you have to do. I don't think the problem is the providers, is the end users the enterprise and the people who will be worst hit by this. How do you give people incentive to change DSL CPE, I think we can study ??

GEOFF HUSTON: I disagree and either you guys are all liars or a very non?representative sample. Five percent of the autonomous systems out there advertise v6, the rest don't. That is 95 percent. Even if you take the small class of ASs which are transit it's still only 15 percent advertise v6; the other 85 percent don't. Now, as far as I can see, all of you are doing it, so either you cram into that 15 percent congratulations, or you are liars. There is a huge problem out there and almost despite those numbers, the measurement evidence suggests this industry is in a postponement mode and the industry is still waiting. Exactly what they are waiting for, I don't know, but the numbers suggest that they are.

GERT DOERING: I am not so surprised by the numbers and they don't really disagree with you, Geoff, because those that do their planning and their homework today are not those that you are going to see in the BGP tables. If they are prepared their network ready, getting equipment tested, everything in place so they can just switch it on next year, you wouldn't see them today. They might have allocation and what we see from the allocation to routing statistics, we know that about half of the allocations are not visible in the routing system, so some of them are dead because people went bankrupt and allocations just lost but I am sure some of them are actually working with that and getting things ready eventually.

Well I am optimistic today.

KURT LINDQVIST: AS number is really, really poor measurement of how ready we are to deploy v6 towards end users. I would like to have graphs, how much these v4 space they AS have allocated to them.

GEOFF HUSTON: Happy to go there. I really need assistance with getting data and folk who are running dual stack websites that have v4 and v6 aligned, I need your data to generalise from one or two sites is tough but the more sites we get, the more accurate a picture we have of the entire industry and you are right it's more the ISPs, it's end?to?end.

MAARTEN BOTTERMAN: Thank you for that. Now, we saw ten of the people thought there was none of the above. I would like just for the count of time only allow three people to react now to the panel discussion or to the nun of the above question. Could I see some volume tier.



RUDIGER VOLK: Not taking a position what the actual status is. The question of kind of assumes that, at this time, for all the relevant players, there is actually still sufficient time to do all the nice planning and I am quite sure there are parties out that have so much that, need so much lead time for doing the proper job that, well, OK, you can't answer the question, the questions with the answers that you offered. We are already damn late in this game.

AUDIENCE SPEAKER: If I had to manage this panel, I guess I didn't focus on what you did, sorry. But you, the title of the panel is what are the bottlenecks and what is missing to deploy IPv6, so I am not looking for the causes where if we are late or not and the causes where we are here, just what is miss missing and to me ?? as an example of what is missing, if you are asking an ISP to deploy IPv6, but say, at the same time, tools like IPv6 MIBS and not flow monitoring is not available you will laugh and he will right. So vendors have to be also shake answer bit.

MAARTEN BOTTERMAN: Thank you. That is the next slide.

AUDIENCE SPEAKER: I have a question on Jabber from Prins from the Netherlands and his comment is there will never abuser demand for the simple reason that the user does not care about how he is connected as long as he is connected. The drive should be own only by the ISPs and companies. Besides that in the big companies we see big hesitation to move to v6 mainly because of security issues. Thank you.

MAARTEN BOTTERMAN: Thank you very much.

GERT DOERING: I want to respond directly to that because one of the problem with there is no user demand is there is ?? or there is no customer demand is not all customers are equal and this is often overlooked. If you are a backbone provider there is lots of customer demand from smaller ISPs buying from the big ones so they definitely see customer demand. If you are a DSL provider providing connectivity to home users like my parents, they will never be demand for anything IP related because they don't know what IP is so they are not asking for IPv4 either. Regarding the enterprises, this is an interesting problem because they could profit a lot from v6, with regard on what the global Internet looks like, but enterprises are traditionally very, very slow in changing anything and I think this is actually where 95 percent of these ASs are hiding, this is enterprise stuff because the ISPs are likely to be there.

KURT LINDQVIST: I actually also want to address this. First of all the issue of the security concerns. You also brought up in initial slides and I think it's kind of fun we are now for ten years listening to the IPv6 advocate saying it's much more secure and saying now it's not ?? I think neither is true, both are as poor. I think ?? as Bernhard, the what sort of bottle next are doing this, the lack ?? now in 2009, for the first time we have commercial available DSL mod ohms that support v6 and I am into the windows user, was weeks away from having the most used operating system on the planet supporting v6, that is actually pretty scary guys. That is probably one of the big reasons why we don't see more update. We can we can build as much backbone as we want but no one there to ? backbone.

GEOFF HUSTON: I still think the basics here are fundamentally economic. You are all fine and all of your customers, today, are all fine. They actually don't have a problem. So this is an /*EBG /TUS situation the onus on one group but the costs imposed on everyone and players need to act in the common interest and make investments which, in the short?term, yield no return. That is amazing difficult for a deregulated industry. It's easy for government, difficult for deregulated industry, because shareholders want returns this quarter, not in five years' time. This is why I believe that the situation we are in now is no accident. The situation we are in now was I suppose predictable ten years ago if we had thought about it. It is a difficult problem that will not, I believe, get solved purely by this industry. We can be moan SNMP and all kinds of things about CPE but even when you fix them it won't fix this problem. It's deeper and more /STRUBGT /TRAL than that.

MAARTEN BOTTERMAN: Before we move on to the next question, is it ?? there is more questions coming and at some point we want to leave for dinner as well. Same for you. Three more questions and then the next question.

RUDIGER VOLK: Geoff, I think I am very much with you in pointing that ??, well, OK, the economy is the background. However, I would like to have an opinion, well, OK, is actually everybody in the game right and fine right now or are there a few zombies who just do not know they have no chance to survive the next two years?

SHANE KERR: I agree with ?? I basically agree with Geoff. This is Shane Kerr from ISC. But I think ?? I think if you view it could you actually take a slightly more insidious view because all the ?? everyone operating a network today will be in a better position after IP exhaustion than any new entrant so it actually, it's actually against your commercial and other interests to encourage IPv6 adoption, and even if this is in isn't a conscious decision on your part acting against it or at least delaying and not doing anything, is actually, it's going to serve you as an organisation making profit, it's going to help you in the future. I think for the Internet that is a horrible horrible thing. One of the reasons the Internet has been so successful and has brought so much good to the world is because it's quite /T*URB ail to start a new service. I mean, if Google couldn't have create add service that could be viewed by everyone, they wouldn't exist, right, or they might exist in some other service, surdescription, but that is basically my point is that ?? yes.

MAARTEN BOTTERMAN: You know people how to get people on the edge of their chairs.

AUDIENCE SPEAKER: I have a comment from a different point of view. I am from national research network, so you talk about economic questions and so on, but we have IPv6 on our network and our user has done for free. So, but we have a few universities that, now, are using IPv6, so this is an interesting point, because they have the possibility to use IPv6 at the moment, but ?? so the economic point is not for these people and but they are not using that, so, we are trying to understand why we ?? they are not using and one main point is the problem for the lockal area network, that (local) maybe an obstacle for these special entities. And another point, are the advantages for IPv6 in terms of application for the level 7. We are doing a big ? inside the community to move the middle to IPv6 and this is a big point because many of our institutes and universities are using /TKPWR*EUD middle wear but it's a fight to move the fiddle wear to IPv6 and it's the same for the applications, so maybe a point of...

MAARTEN BOTTERMAN: Thank you for that. I think that leads us to the next question, indeed, and I have asked the panel to hold their comments for the final round until the end. The the next question is, in fact, about hurdles, and in hurdles we distinguish experience from perception because many people think what it is and what we have seen in the survey as well, people just think what it is and people experience what it is. So first and for all the first answer could be as there is very little customer demand I can say what I want but I can't convince my management to make the investment. That is one.

Two and three is the experience or the perception that costs are really substantive, and I can assure that you we have had different answers there. The same thing for technology and the third thing for vendors. Could I ask you to close your votes in there and again, we are ready to move towards the real live answers on that. People still working on it? If I see people typing I never know whether they are filling in their e?mail. We are closing it now and getting update and this is what we see. Actually, there is ten people who find that whatever can't convince their management today; there is /PHAOEUB or thes but also people who experience that costs are substantive; equally perception and experience, that is funny; technology, in fact, is the same; and there is a great ?? technology is the biggest score here. I have the experience that vendors cannot deliver what I need." "None of the above" is is a pretty high score as well. It's what we call in Holland understatement. Please can I get your comments on this and after this the floor is ?? again response. Doering Doering if I would be asked to answer the questionnaire I would most likely click on "none of the above" because the options are pretty extreme, like very difficult, very expensive, vendors cannot deliver. I would want to actually sign it in that extremity because, yes, they are technical problems and they need to get solved; yes, there is cost involved. You might not need to buy new gear but your people spend time configuring the gear, working out addressing plans, fixing the problem, so there are some operational costs attached to it but it's not very high if you do it not in a rush but sort of phasing v6 in, so yes I would click on "none of the above" because in these extremes it's not fitting my experience.

GEOFF HUSTON: From my experience doing this it was actually interesting that when we first rolled out a technology, be it dial up or DSL, a huge A the product cost was actually in servicing customer calls, which if anyone has done this would realise it costs an enormous amount of money to have the phone picked up by someone intelligent 24 hours a day. It really is your major cost. You spend a lot of effort training your customers not to call you and that means better equipment, service, etc.. now the thing about v6 is, all of a sudden there is this prospects of a whole bunch of, again, be mused, amused and be if you had he willed customers, it's the not the cost of kit, it's the cost of customer base that changes dramatically, how can I offload that cost and get rid of it. Your question here is phrased on infrastructure, but some of this, particularly in the mass market, the numbers involving millions, is actually more about the process of servicing a customer requirement because you have just changed everything from underneath.

MAARTEN BOTTERMAN: Thank you.

KURT LINDQVIST: I don't think I have anything to add to what Guert said before. I think none of the above seems more likely, Geoff illustrated there is a lot more components to this than what you give for here.

MAARTEN BOTTERMAN: OK. Thank you. Well actually, I must say that the panelists are not responsible for the questions; you can only blame me for that. Please, I would like to open the floor for explaining what you feel about this and I think "none of the above" is clearly what is the real issue here. The vendor seems to be a real issue still, even maybe not as extreme as here and convincing people who are not in the trade, maybe another thing.

Wilfried: From the Austrian research network and looking at the situation at Vienna university. I stopped participating in this exercise because I would have to click on /KWR?GS none of the a/PWAOF" for all of the questions, the reason being that my view on the whole thing is that we are completely off track by looking at implementing IPv6 or making an organisation IPv6 capable or not, by looking at that as a black or white or as a digital decision as 0 or 1, because from my point of view, and national research network is probably not very ?? not a very good example for other types of industry in the IT area, but some of these things which are required are easy, some of them are very easy. Some of them have been done years ago. There are probably a couple of folks around who would agree with that. There are others which are sort of medium difficult but you can find a solution for those; it's going to be cumbersome and take time, and take some investment and take to you wait for a while until you get to the next cycle of rolling sort of ripping out the old stuff and pushing in the new stuff. The biggest problem from my point of view, actually, is somewhere in the layers, seven and eight, sort of going towards infrastructure, going towards applications, and that sort of things, and the really hard stuff from my point of view is those applications or those operational pockets where new stuff has been put in place for the last couple of years and for those projects, nobody was clever enough to require IPv6 capability for those things or even if people at the decision points were clever enough, they had to take the decision to not implement that service at all or to implement it IPv4 only, and one of the example that is we have on our desk at our university, is for example, a platform for e?learning. If you would have required two or two?and?a?half years ago to have a full?blown IPv6 capability going out for a call for tender, the number of offers would have been digitally 0, so it's just reality. And at the same time, if you decide to go with an IPv4 only solution, you end up with new boxes in the network, you end up with more complexity in the network, like boxes to support load sharing. That was a functionality that wasn't there two or three or four years ago, so if we would have managed to go IPv6, then we wouldn't have to deal with a situation right now where the IPv4 based Internet hassy involved, IPv6 has not evolved, we are in a position where we have much more complexability right now in IPv4 and there is no (evolved) obvious solution to that. If you look at the backbone, job done, doing it since years, and even there, you run into problems with operational management, like trying to find from one of the well?known vendors to find an operating system urgent for the backbone nodes which does dual stack and give you traffic statistics, flow data, whatever you call it, you have a hand picked set of ?? that is the reason for me why I sort of dropped out of that whole examinerise, because it's not easy to say yes or no, different easy, it's different for different pockets.

MAARTEN BOTTERMAN: The other two comments first.

Marco: I am also one one of the "none of the above," I agree with Geoff, the cost isn't in the hardware, it's all capital expense, it's five year writeoff and actually, there is no ?? there should be no price difference for hardware doing v4 and v6 and in fact a couple of load balancers last year and they didn't get any more expensive because I required IPv6. Yes the choice was a bit less and spent more time testing them and convincing the vendor it should be IPv6 and it's all operational. And in fact, if you do pay extra for IPv6, basically you should say no, IPv6 is not a product F a vendor tries to sell you IPv6 licence tell him to go away and go to his component, IPv6 is critical for infrastructure, don't pay licence, just tell him to go away.

AUDIENCE SPEAKER: I was one of those who click "vendors doesn't have products for me." And I need to justify myself, because I am responsible also for security boxes and so on. I need to say that in this piece of market it's a very big problem because both well?known brands and so the smaller ones doesn't offer full supper for v6. Sometimes it's a problem because having the data sheet separate yes and that means they can fully pass through v6 packets or fully block them and that is v6 ?. And sometimes it's only in routing mode, which is not an option for me, and few of the vendors elsewhere, as we heard on our road map for 2013 and maybe 14 or something like that because there is no demand from the market. So if you as a customer want a security product with some extra functionality with some non?standard things with v6 of course, it is a problem. I do not know right now any brand who fully support bridge model v6 and VLANs and some more extra functionality and that is VLANs and bridge mode, those are normal things we are using for years. And as Wilfried said I think, load balancers also. There is a problem, still, and application of firewalls also has a problem and web caches for big portals, they still have a problem. In Poland there was a PL NOG meeting and I think on the last or previous one, we discussed that, and guys from huge portals, they said basically that they don't because of ?? they don't ?? not they don't care ?? they have a problem implementing web application cachers and so on because they are not existing. So basically speaking, vendors just sell routers and they do not sell anything else.

MAARTEN BOTTERMAN: Thank you for that. For me, it's all help me to put this question into perspective, which was useful, even when the answer was "none of the above." Gentlemen, for your last comments and then we move on to the next one. Doering Doering I need to agree with part of what has been said. Some of the vendors are really a bit slow in taking up, to put it mildly. And we have a fairly complex network with about every single piece of hardware ever built by anybody and none of this is fully v6 capable, as we would like to have it. The consequence that we took for us was to, well, fix one bit at a time, so we look at what is working, what is not working and what needs replacement anyway and then we try to get replacement hardware that is v6 capable or we yell at the vendor until they fix the box, so it's really not like you can turn on v6 on the network overnight for everything, but that would be high cost anyway, so doing it gradually and trying to kick out complexity that is v4 only, works for some pox and for other pockets ?? if the /PHRO /SES is going on and you can replace the v4 only stuff next year and everything else is v6 ready then you are (pockets) already sort of in the fast lain

GEOFF HUSTON: Spoken like a true customer I want everything and don't want to pay. You can get anything you want from a vendor if you are willing to underwrite all of the costs and of course, what the vendors are really saying is that, at the price point, this market is putting itself at, which is commodity based cheap networking, the Internet really is cheap, then we as an industry are making it very difficult to get massive amounts of customised engineering whacked into products in short periods of time because that is expensive and it's kind of I want it all and I don't want to pay for it, really says nothing is going to happen, so part of this really is, if you want a solution, you have to put money on the table, but the money isn't coming, hence the wedge. Doering Doering I am hearing you and you have a point here. On the other hand, the amount of free product testing and debugging I have done for certain vendors over the last couple of years are not paid by these vendors technical crew either, so I am giving my vendor some value for their software.

KURT LINDQVIST: Can I make the oaks Geoff, there is actually product offering very similar to v6 that the vendors are throwing after you whether you like it or not and that is MPLS. Most of it comes for free, but v6 you have to pay for.

GERT DOERING: Kiss sew changed this,

MAARTEN BOTTERMAN: Please, those two comments here for closing on this question.

AUDIENCE SPEAKER: Geoff, what short period of time are we talking about? I don't think the v6 features set is really catching vendors by surprise. You described it as a short period of time, that it the vendors had to implement to v6.

GEOFF HUSTON: Oh, I was hearing that the complaint was it was taking a long time to get the features in, that this wasn't coming and that wasn't, and I have to wait for the vendor to get equipment and my point was, it's a problem about how much you are paying your vendor. If you put a large sum on the table you grab their attention. If you are saying "well, I will put it in my next RFT and I am going to buy three units next year you haven't really got attention have you?

AUDIENCE SPEAKER: It's not a question of next year, it's the last ten years. We have been spending money with our vendors over the last ten years.

KURT LINDQVIST: Trying to build stable v4 routers since '91/2.

MARCO: I have seen occasions where we had the full working v6 implementation and all of is sudden so the cost was already made and all of a sudden it got popular and charged people for it so it's a different way around. On the other point I want to make on your side, how much development have been doing in IPv4 over the last 20 years and how much money did they get for us from that one already and still getting for it.

GEOFF HUSTON: We gave that to the shareholders already, that is gone and then we blew it up in a bonfire last year called global financial crisis.

MAARTEN BOTTERMAN: Thank you for that. This point on who is going to pay the bill for the last ten years on anything is the last round of remarks. This is not going to be "none of the above" for anybody. What do you think? IPv6 has no value over IPv4 in terms of deployment costs and risks today, is that valid for you? It has value already today; it's unavoidable in due time; you will probably never deploy IPv6 and the last two questions we have been discussing already so I won't repeat those. I am going to close on this in about five seconds and then ask for the comments again. OK. "None of the above," only seven this time. The biggest responses were here for the deployment of IPv6 is unavoidable which is not a surprise but by far biggest people say, most people say technically everything is in place to deploy IPv6. What do the answers here tell you, panel members?

GERT DOERING: Why are we here?

GEOFF HUSTON: I would offer there are two views out there and hopefully not diametrically opposed. If I was working for the company I worked for the last ten years, my answer would be realistically I will do whatever it does to maximise shareholder value. That is the mission of the private company. So is v6 in that or not, I have no idea per se. If, at the time, it looks to be maximal shareholder value we will do it, but quite frankly if some other mechanism gives me more money, traps more users in my cage, offers me the ability to sell for a higher price, or do out something else that maximises shareholder value, we will do it. On the other side, is the thing that we are actually in the business of public telecommunications. And all of a sudden you are in an entirely different area. What maximises the common good? What allows innovation, creativity and folk to play? And the answer is, openness. The answer is preserving a platform that doesn't trap users and trap providers. So if I am looking for a solution that maximises the common good, I really am faced with the issue that the only way of doing this is to keep a network open and the only way I can really do that with today's technology is to keep addresses open and v4 isn't going to do that. So, does the private interest coincide with the public interest? Right up until this day and this point the answer is they don't look like they are working in the same direction, that is a problem.

KURT LINDQVIST: I was a bit surprised by one of them was that they were the second most popular answer "I was deploy it in due time." Given what Geoff showed us on the slides was due time is pretty close. We had this over lunch, planned this panel a bit, me and Geoff were having what sounded almost like a disagreement, we don't need more information campaigns about deploying v6 because you people already know that it's really close in time to do this but this slides tell me you need to start doing this for address presentations at every single meeting again.

GERT DOERING: More train wrecks, please.

GEOFF HUSTON: Do folk honestly believe that this industry is uninformed? I must admit I find this very difficult to understand, that the major players out there, and there aren't that many, who have millions, hundreds of millions, 1.6 billion customers, are uninformed. I truly believe that the answer isn't more train wreck slides, much as I love doing it and you love hearing about it, I truly do believe that this is an informed market, and that where we are is actually a process of self?interest being expressed, that optimising shareholder value has got us there. So I am dis ?? I disagree, I really do think we know what we are doing.

AUDIENCE SPEAKER: Sorry, I would like to answer that, because I guess with my moore words, Geoff you are polluting this debate with economic conscience, with mass markets, we do not care at all, and for that reason, let me take just an example: When the ?? when the vendors decided to go for mobile telephony, do they do what you are suggesting for IPv6? Do they ask for the version of the protocol they would like to have or what kind of ?? what kind of ?? what amount of money they are ?? ready to pay for? Not at all. Now, ask your grand ma, which IP version will you want to use, or would you like to prefer? She doesn't care, and me too, because my customer mr has nothing to do with the IP version. There was one comment somewhere about Jabber, I guess, saying "unless I am connected to the Internet, everything is fine." It's not my business to decide what kind of protocol I have to use. And that is really the point we have to reach. We don't care about the version of IP. And it's not a mass market decision here.

GERT DOERING: That is what I tried to bring across a couple of questions back, that there is different sorts of customers. If you have a web?hosting customer who has to upgrade the load balancing gear, which costs money to have their website v6 capable, they need to make an informed decision. If it's the end user in the home that will just plug in the CPE that Telco provides and will use whatever IP falls out of it they don't want to make an informed decision, they just use Internet. I think I see some one of the big Telcos having comments for us.

RUDIGER VOLK: Private citizen. Well OK, Geoff, I wanted to jump on your talking about is the industry informed. Well OK. The industry is flooded with information, is the industry when we are talking not just about, well, OK, what common frenzy is is a single party jumping on, but well OK, what is the direction that the collective is moving towards, well OK, there should be also the question well OK how enlightened is it moving. And well OK, just looking at there are things in this industry or very closely related industry where huge amounts of money are spent on technologies that is pretty recently defined in detail, and well OK, so, are we seeing that, say, free GP P, which is absorbing lots of money and which is absorbing that money in a time frame when, well OK, the information available to the industry certainly included that, well OK, IPv6 is coming and IPv4 is going away, is it really enlightened that well OK, free GP P doesn't do v6 investment?

GEOFF HUSTON: Yes

RUDIGER VOLK: Is it or is it not?

GEOFF HUSTON: It is, it's all about customer lock?in and not opening up the network for other folk, it's all about putting boxes in the way and then renting that space. Unfortunately self?interest wasn't in common interest in that situation, sad but true.

GERT DOERING: When rolling out 3G networks v6 would have had a nice vendor lock in because there is nowhere you could go with v6.

MAARTEN BOTTERMAN: Can we go to the floor again?

BRIAN: Brian from HEAnet. There is a phrase which is used to repeat the same activities under same conditions expecting a different outcome is a sign of something bad, bad thing, and having been to quite a few RIPE meetings now and IPv6 Working Group sessions and plenaries and heard Geoff, among others, speak on this topic and we are here again and asking why people aren't deploying IPv6 and we are getting not dissimilar answers to every single time we have asked this question before; I am wondering what ?? maybe there are people in this room who are getting something amazing out of this session and out of sessions like this and maybe there is some pearls of wisdom which are propping, because I have heard too many of them or because we have had a fully dual stack network since 2003, that I am missing the point, but I am just really kind of wondering, again as Geoff said, is the industry not informed? Is there anyone in this room who is sitting here going, that is amazing, I never realised that? Is there anyone who is technical in the wider Internet who suddenly goes, my God, I have v6, I must buy it now, I never heard of this before"? So what is going on here, really, and are we going to continue to have sessions like this again and again and again until ??

KURT LINDQVIST: But isn't the problem it's not the people in this room that need to be informed, it's when people go back home to their companies and organisations, they go to the CFO and explain unless we do this v6 thing we are not going to have addresses to send to our customers. When? Geoff tells me 2011?12 whenever it was. CFO I am trying to figure out how to pay the mortgage for next month. It's just not on the radar screen. I think in a way, maybe the problem is the time?line is too far away, they have much more pressing issues.

MAARTEN BOTTERMAN: What we can do as well, looking back at the whole thing, let's do that after. We also looked at the current options that are out there, the activities that are out there to support roll?outs, OK.

GEOFF HUSTON: I would suggest the folk who are trulily uninformed come from the regulatory sector. Private industry will solve this they are saying and private industry is not and they are sitting there going "I don't have a problem" and I think they do.

MAARTEN BOTTERMAN: We will come back on that.

AUDIENCE SPEAKER: I wanted to go back to the mobile operator case and to go to the same way that Bernhard said before. Do you see a lot of mobile operator deploying IPv6? No. The reasons are simple. Everything is running very fine in IPv4 and we have not identified any applications today and because we have lack of these applications, we have no reason, it's just a cost to to deploy IPv6 but even if the problem is is a lack of IP addresses because most of your mobile operator will have a lack of IP addresses for the terminals, we still faced with the problem of chicken and egg with the vendors of terminals and it's really a problem, you know. And it's four years we have heard about IPv6 terminals, I talk about mobile terminals, not PC because on the fixed part it's easier PC is running and IPv6 is ? but in the case of terminal vendors, it's not. It's still not.

MAARTEN BOTTERMAN: Thank you for that. David just pointed to me we are flying through time so all the comments now will remain for the last final round.

The last question is about actions that matter and in fact, some of you may have seen the action plan from the European Commission. There is different actions in it but the commission thinks that should be done to make things happen, anxious to stimulate IPv6 availability, consent service applications and they want to make government information available via IPv6, so government websites and portals whatever, encourage content providers to offer their content fee, they are not very precise on doing that, the same to standardisation bodies.

The second line of generate demand and through public procurement. Public procurement. To ensure timely preparation for IPv6 deployment, they see awareness campaigns. This is the action plan was published in 2008. Standardisation processes, and in fact, make sure there is some knowledge, knowledge pool developing in Europe and thirdly, actions to tackle security and privacy issues.

I added an option, other actions than those mentioned in the current action plan and I ask you to vote on that for the speed of time, I already saw many responses were in. Get your input.

GERT DOERING: I just start with a comment. Geoff said that governments try main date ? and it failed. Well, I am not so old, when I came on board IPv4 was already a big success so I missed that experience.

GEOFF HUSTON: Lucky you.

GERT DOERING: I think it would help to have bodies like the EU stop talking about how great IPv6 is and just demand it from their vendors. That is what some of these operators seem to be waiting for. It might fail in interesting ways but at least one of the excuse would be gone.

MAARTEN BOTTERMAN: Just to go over, and these are literally the actions that are in the action plan: The headings are big difference, some difference but not much, no difference or counterproductive. What you will see is that, actually, there is quite some applause for the action to stimulate IPv6 accessibility, that they can make a big difference. There is a little bit less but also high for government information reachable. There is the highest is to encourage providers to offer the content via IPv6. Standardisation body encouragement is not seen as very effective. Actions to generate demand is interesting but even more so to have governments to take up IPv6 within their own networks and service contracts. As has been said here several times, if you look to the rest, seen as somewhat interesting but the most extreme are the three points mentioned above. In fact, 33 percent says that other actions than those mentioned in the action plan are, they don't know about it. So there is not a strong proposal for other actions than mentioned here. Taking this into account and looking at the highest scores, could you express your opinion.

KURT LINDQVIST: So I think the problem with this question it's not contrasted to anything. There is nothing on here that I think is doing inherent harm, I don't think anything here is really bad, when you ?? would you rather see this being done or do you think some of the things that Geoff mentioned with creating real stimulus on the market, is that better? Would you rather the easy spend your money on doing surveys or upgrade to CPE to capable device? Doing things that mattered would take political leadership and courage, that is a dying breed of politicians, they don't exist any more, at least not in Europe. This is the problem, all of this is very nice and sounds a like very nice words and go nice on PowerPoint, it doesn't really change anything. It's hard to be against it.

GEOFF HUSTON: How to get infrastructure providers to change their basic infrastructure? Allow to write off in taxation depreciation on their existing equipment. How do you get customers to upgrade their CPE? Have a list of devices that they can buy and give them a voucher redeemable for CPE. We have tried to before in other industries, it works one way or another. It's not a case of saying we must do this and that is ?? that isn't enough. When you strike a problem that is akin to a market failure, you have to then be rather creative and understand that it needs an amount of push from the public interest to then make the private sector respond in the way you want and you can either use a stick or carrot, you can play on fear or greed but we are basically optimistic folk and basically greedy. Carrots always seem to work. And if you can change the investment profile for ISPs they will change their equipment faster. If you can make it for consumers they will do so. If you fold your arms and say we will conduct a survey next year and see how you are going, quite frankly the numbers we see today are the numbers we are going to see in 12 months' time and that is probably not the best piece of news.

GERT DOERING: I am not sure what to add to that. That is a really good closing word. Maybe let's try with a quote from Randy Bush, who is not here, unfortunately, and when I met him at one of these IPv6 conference thingies, what he said was let's stop moving power points, let's start moving packets. That is basically falling in line. Do something and stop talking about it.

MAARTEN BOTTERMAN: Thanks. Well, actually, just for the sake of time, I think, we have to bring this to a closure now. I would like to say is, one of the things that I experienced last time was my first time at a RIPE meeting, the least thing will happen now, and it may not be new facts for all of you but I think we have discussed it in a way that decision?makers, policy people may understand the discussion as well and if you want to get policy people to understand and act, you need to talk the language they understand and act. And if this discussion and what we wrote up on this may lead to a coherent message, then I think we have made a little bit of progress.

I would really like to thank you for your active participation for RIPE NCC to give ?? and for David to give the opportunity and the panel members for what I think were very constructive and coherent answers. So, please, gentlemen, let's give the floor back to David.

(Applause)

DAVID KESSENS: Thank you very much, again. I think this was a bit different meeting than usual, was more a meeting of reflection and not moving packets. I think we have got a clear message of the audience and community and we as a Working Group again to get back to work and work on technical issues, next meeting get a new charter in place before the meeting so we can deal with pragmatic and real IPv6 issues and perhaps coexistence issues that are going to be relevant by then. Thank you very much.

One final agenda item is always, is there anything happening in the RIPE service region that somebody would like to announce that is interesting for the community? I didn't see anybody coming to me. So then we are done with this meeting. Thank you very much. One item. Chris Buckridge: The social tonight is at the Budda bar and being sponsored by RIPE NCC Act Now. The buses are leaving between 7:00 and 9:00 from the terrace side of the hotel so if you are in the lobby you will see where they are going from. The social itself starts at 9:00 but there won't be food so if you would want to eat beforehand and there will be buses coming back to the hotel between 10:30 and, I think, 1:00 a.m., so every half hour or so. Looking forward to seeing you there tonight.

DAVID KESSENS: I heard you have to bring one non?IPv6 capable item as a donation to the RIPE NCC. That is just rumours. OK. Thank you very much.