friday

Note: Please be advised that this an edited version of the real-time captioning that was used during the RIPE 56 Meeting. In some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the session, but it should not be treated as an authoritative record.

RIPE 56 Closing Plenary 11:15am, Friday, 9 May 2008

CHAIR: Good morning everybody. We have one minute and then we start. So if you could please find a seat.

Good morning. My name is Rob, I am the Chairman of RIPE and I am chairing this last session. Before we go into our regular agenda, I have one or two things I want to say.

I think we you can all agree with me that this has been an extremely interesting RIPE meeting. We have had many, many lively discussions on many important subjects. On Monday, when I stood here, I expressed our wish that our host organisation, DECIX, could ensure that we would have nice weather for the rest of the week. I think we can declare this project a success, which brings me to the happy task of saying a few more words of thank you.

DECIX has been acting as our local host and I think this has been an extremely nicely supported meeting. Thank you DECIX.

(Applause)

We had most sponsors for various bits and pieces of the programme, especially the evenings, this week, [ECO], Netnod and Colt, and we likewise thank them all for their very valuable contributions. Thank you.

(Applause)

And last but not least, I would like to convey our thanks to the stenographers, who have been coping very hard with our various interpretation of English and various speeds that's for sure and who provide a valuable support for many of us in this meeting. Thank you.

So, without further ado, there will be some another announcements later on, I would like to give the floor to the first presentation of this morning and this is Philip Smith who will enlighten us about some of the technical operational coordination activities going on in various other parts of the world. Philip.

PHILIP SMITH: Okay. Already, good morning everybody. This is going to be a fairly quick fly through some of the regional operations groups around the world. Basically I am involved in quite a few of them so I thought it might be a bit interesting just to let folks here, just see what other activities there are around the globe.

I usually get asked in some of the operations meetings I'm at, what's happening there, [unclear] there was anything going on, and so force, so I think pulling them altogether in one presentation I thought would be interesting for some.

Apologies for the picture in the middle. Some of the operations groups meet in quite difficult places to be. Mind you it was beautiful and sunny here as well but that's a Pacific network operators group and that's probably half the Pacific Islands operation community last year.

So, you pretty much know what the operations groups are about. Pretty much like the meeting here. There is a lot of business relationships are worked on, there is all the peering stuff, industry relationships, technology discussions and so forth. I have put a list there. It's not complete, but it's generally what I tend to find in most of the meetings I am participating at.

So I want to look at the regionals quickly. I promised Rob I'd go through this presentation quite quickly. Apologies to the stenographers for that.

I think most of us in the room probably have heard about NANOG and know what it is. It came from the NSF regional telecommunications or evolved out of that in 1994. The north American network operations but there are quite a lot of Europeans and folks from Asia and Pacific who come to that meeting as well. Three meetings a year, one of them is held along with NANOG. Around about 500 attendees. And as we found out from the recent NANOGs, getting the programme out early ensures that we have better attendance.

Two days of plenary, one day of tutorial, approximately. There are also panel discussions, there are various BoFs, peering and security have been two very longrunning events.

Moving over to the Asia and Pacific region. We have got APRICOT which is the Asia Pacific Internet conversation operational technologies. Someone figured that one out a way back in 1996 when it started. It's colocated with one APNIC meeting, so APNIC has actually formed a fundamental part of APRICOT since it started.

It moves around the Asia Pacific region. The next year it's going to be in Manila in the Philippines. Year after in Kuala Lumpur. 2011 we are looking for potential venues and so on into the future.

Attendance is around the 600 mark. Again it depends where in the region we are at. How difficult it is to fly to and so forth.

Programme: Big change here is that we have workshops. Workshops started in APRICOT in 2000. We have five days of five workshops. They are very well attended, usually full to the door. Then we have four days of tutorial and conference and then it finishes off with a one day of the APNIC member meeting.

Moving into south Asia. Now APRICOT actually covers the whole punish Asia Pacific region which included south Asia, but the needs of the south Asia region and the logistics for getting to APRICOT meant that it seemed very, very obvious that a south Asian network operators group could address the needs of south Asia. So, that was started in January 2003 by Gaurab, who is sitting somewhere here in the audience. He is the chair of say nothing. It's basically the Sark region, so from Bangladesh in the east to Afghanistan in the west.

Two events per year: January and approximately July. Early days attendance was sort of high, well high sort of 70, 80s, low hundreds, we now seeing regularly 250 people coming to that. So it's got its own momentum.

SANOG 12 and 23.

We haven't been to the Maldives yet and we haven't been to Afghanistan either.

Format is similar to APRICOT. We have workshops. Workshops we are turning people away every single workshops. It seems as though people sign up and they tell their friends. So it's good to see such great interest.

We also have an APNIC Open Policy Meeting showcase. Similar idea to the RIPE NCC regional meetings I suppose where APNIC spends some of the time during SANOG to explain policy issues and what else is happening in the Asian region. We tend to try and have some [unclear] meetings happening alongside too.

AFNOG has been running since 2000. Major focus of AFNOG is still on hands-on workshops. We are starting to see a few more tutorials and of course the conference now. There is one AFNOG meeting a year, that's co-located with one of the AfriNIC meetings. This moves around Africa. If you can see the logo on the screen, the green pieces are AFNOG has been. The yellow piece, this logo was from Nigeria where AFNOG was last year. The idea is really to make the whole of Africa green, then we have been everywhere. That will take I think about another 30 years or something.

The programme, as I say, five days workshop. There is a lot of investment and time and effort actually to build the infrastructure. Quite literary [unclear] to go and build networks before the actual workshop can start. But it's pretty exciting for everybody involved.

Moving into the Pacific. Again, the Pacific Island nations, on the periphery of the APRICOT region. They rarely could make it because flying from Pacific Islands tends to be through New Zealand and it's painful to get to. The APNIC meeting was held in Fiji in 2004 and the folks from this is a good idea, maybe we should carry on. So the fourth pack nothing will be in Vanuatu at the end of June. The programme is like a five day workshop and two day discussion, conference recollect it's reasonably informal because the focus is on workshops, the training and so forth. So we have instructors, supporting from NSRC, from Internet New Zealand and various other places.

MENOG is probably the newest regional NOG that has started. This was launched in April 2007, following the RIPE NCC regional meet indicating Bahrain in November 2006. The RIPE NCC has been fairly fundamental in I suppose opening eyes to well trying to make a network operations group function within the Middle East region. So this is currently meeting twice a year. April-ish, October, November-ish. MENOG 3 was in Kuwait last month. It was probably about three weeks ago. Next meeting locations are in Egypt and in Oman and of course MENOG will be actively participate indicating RIPE 57 in Dubai later on this year.

Programme is still developing. We started off pretty much with just a conference. We have started add indicating tutorials. It's still very early days. And there is still quite a bit of future expansion being planned. A lot of good ideas being thrown around, so, we will see how this moves forward.

The two country NOGS, [unclear] going to put country NOGS in this. It's two I am just fascinated by and I think they deserve special mention from other parts of the world. First one is NZNOG, now I think of New Zealand and, it's quite far away. It's only 3 and a half hours from plane where I live, but it's pretty much a long, long way from here, 30 hours travelling. It meets once per year, one day of tutorials. Two days of conference. Look 120 participants. There are 4 million people in New Zealand and they can get an amazing turnout for the size of the country and the size of the industry for a networks operations group like that. If you translate that had into back other Council tree operations group, you'd be almost in the thousands and so forth. So, that's a pretty impressive nothing, they don't fear about what they talk B I still find NZNOG is the most interesting one that I have ever been to, just with the breath, depth of the discussions, so pretty interesting to see there.

JANOG, I haven't actually yet been to, but it's another country network operations group that's been running for a long time. Since 1997. It meets twice per year. It's in Japanese. We have some of the organising committee from there with us. Around 3-400 participants. They have two days of conference. Website of course is in Japanese with an English subsection. I have actually been invited to that one, coming up later on this year, so I am quite excited, quite privileged to have been invited today participate in that.

So, more information: Well, various places that keep track of all the different NOGS, well one URL that I use is that one, I think [unclear] maintains. As for the meeting planning, most of the operations group meeting planners that I am aware of tend to use the calendar that was set up and maintained by the network start up resource centre, the NSRC. That's what we try and tend to use to really to try and avoid conflicts and overlaps between the different NOG meetings. It's pretty hard. You still end up with inevitable overlaps depending on venues and so forth. But we try and use this as our planning year calendar.

New NOGS: It's pretty much been a recent phenomenon. Up to 2000 it was pretty much there was NANOG, there was the RIPE meeting that we have here and, well, there wasn't well a great deal, APRICOT of course but there wasn't a great deal else. And since then, we found that more and more one trees and region are start to go look at network operations groups. Language issues of course, local needs, something is acceptable in some region, but some other regions want to talk about things that we probably wouldn't talk about here at a RIPE meeting. And of course local cultural issues as well.

As for the models I tend to fine that SANOG and NZNOG had common models. There is one or two really dedicated people who make the thing happen. And everybody else just helps out.

Some of the newer ones are very worried about setting up lots and lots of bureaucracy. Please don't do this, it makes thing a lot more complicated. Potential new regionals, they are at various stages, central Asia, Caribbean, Latin America, at various stages of looking at trying to do something.

Now, the second part of the presentation was looking at some of the European NOGS, so rather than having people run up and down from the podium here, I have just gathered the slides together. So we will run through some of the European NOGS as some active examples. I apologise if it's not complete.

First one to cover is UKNOF, that have established in early 2005. They have had ten meetings so far. Roundabout just under the 100 attendee mark. I have manage today make it to one of them and again it was, fillet it was a similar thing to NZNOG, it was a fascinating meet, no holds barred about what could actually be talked about about and so forth. I really enjoyed T similar structure. You got the programme committee who pull the whole thing together. 8 people funding Council. Keith is the chair in the audience here.

Charter: I suppose the main thing would I like to highlight here which is a common they'll with all the NOGS. Sharing of knowledge about how to operate networks. Nonprofit. Again that's fairly common with all the operation groups. No registration fees to attend UKNOF, obviously some operations groups have fees but again mostly these fees are cost recovery.

What else should I cover? There is also some outreach to the UK ISPs who are not in a position to attend some of the international meetings. Airfares or even getting permission to travel internationally can be quite hard for some organisations.

There is also coordination with other Internet bodies, getting again sort of fairly important part of any of the operations groups.

Generally keep a nice open mind about things. As I say, tend to find at some of the regional NOGS are more focus regional issues. At least with a country NOG you can drill into things that don't quite fit in other places.

Upcoming meetings: Well, you can see advertising there for UKNOF 10 which is coming up real soon. 11 later on this year and 12 coming early next year. So any UK folks or even others from this region interested in going, you know where to see.

The next one is Swiss Network Operators Group, thanks to Freddie for these slides. There is no legal organisation or entity really for Swiss NOG. But, again, it's an operations group for the service providers within the Swiss Internet industry.

So, usual website, mailing list, IRC channel, the various submailing lists as well. The two annual meetings in about Bern. Bern is sort of in I suppose a neutral part of Switzerland really for the different pieces of the country, so the meeting is in Bern in a brewery. I was interested to see that because NZNOG makes a big issue about beer being a fairly important part of their function as well. So if you are interested in this, then, well you can go and indulge in your love of beer also.

Upcoming events for SWINOG. There is the one on the 14th May is also closed. It's next week, isn't it? Whereas, SWINOG 17 in October in Bern, that's again just before the RIPE 57 meeting in Dubai. So maybe somebody could fit that in on the way if interested.

The Italian peering forum. This is brand new. That's coming up Pisa, end of the month. The main idea here is to create an Italian network operators group. So that's something again if you are interested in Italy, you can go and have a look at that.

Finally, getting on to ESNOG. This is brand new Spanish networks operations group, and Joao has been one of the main helpers with getting this up and going. There is a little participation from the local community indicate wider industry fora. Just again the usual issues of language barriers, travel restrictions and workload and all the usual things. Unawareness of what is out there. People tend to think that they have their own little island they work in and there is really nothing else happening out there and they are quite surprised when I talk about this sort of thing of all the regional operation groups activities.

So a few got together, starting a mailing list, got a venue together, find some sponsors and ask people to come along and people came along, 65 attendees. They filled the room completely. Which is great to see, so very good participation, lots of contributed material. Very lively discussion and as it says there, thanks also to the RIPE NCC and IANA/ICANN for joining in the next meeting, Gore 2 is going to be held in September or October, I guess it's a case of watch this space for more information.

That's my presentation. Are there any questions about any of this?

AUDIENCE: Just to mention that because we here in Germany, there is a DNOG.

PHILIP SMITH: Thank you. I know there is.

AUDIENCE: And on the same token there is actually an LNOG, which is a mailing list they once met in Perth, so...

PHILIP SMITH: Maybe next time we can do a summary and if we get more time in the programme, we can do that.

CHAIR: There is so many NOGS, maybe we can have sub meetings of NOGS, the ones that meet in pubs and breweries. Thank you Philip for a nice overview.

(Applause)

This morning, we sprinkle the agenda with a few unscheduled announcements, and the first one is by Paul Wilson.

PAUL WILSON: Thanks very much. This is just a quick ad hoc announcement or presentation that is fairly topical at the moment. You'd have been partying quite hard this week to miss out what happened last weekend with the cyclone. It was a really major natural disaster and if you haven't been watching the news, there is a few clips here showing inundation in the country, this highly populated southern part of the country which sits about two or three meters above sea level and had a cyclone go straight over the top of it with a 5 meter store surge which flooded these vast tracks of highly populated very poor areas. Picture down the bottom there shows a few human corpses and dead animals lying in the rice paddies where they will probably be for quite sometime. The infrastructure in this area has been completely wiped out and the disaster is going to go on for quite a long time. Estimates up to maybe 100,000 people killed.

So, just this week, a bit an initiative is coming together to deploy this system called DUMBO. I did ask Andy if he could possibly present this things this week because it's a pretty interesting mobile ad hoc network which has been developed by AIT, the institute of technology in Thailand, which has been deployed in test and live situations and so we are now getting together a crew of people who are doing some training at AIT in Bangkok early next week and are going to start sending some Burmese people back home to do some deployment and training and so on.

And that's something that's going on with AIT and Telecom on front [unclear], WIDE and APNIC. This is a very new initiative, but if anyone has any ideas of contributions or help that could be given, or if you just like to know a bit more, then there is a bunch of links there. Please get in touch with me if you do have thinking no contribute at all.

Randy did have a very nice presentation which some of you may have seen showing a bit more of what was going on after the tsunami and these are quite fascinating pictures of how this thing actually looks and how it gets set up. That's the first hop hub in this wireless network. Moving to the next one, the next one is up in that tree. And that's what it looks like. So anyway, hopefully that will be happening over the next week or two. Like I say, this disaster is not going to end next week. It's going to be going on for a long time. So hopefully they'll start letting people into the country. They are not letting too many people in at the moment. Hopefully [unclear] get people in to do this and set it up.

Thanks for your time.

(Applause)

CHAIR: Paul, before you walk away, can you make sure that your presentation, especially your last slide with all the relevant links is somewhere reachable on the RIPE website.

Thank you. Jim Reid, are you around here? Then we continue with the report from the Data Protection Working Group. Is there any from the Data Protection Working Group?

DENIS WALKER: Hello, I am Denis Walker from the RIPE NCC Database Group but today I am representing the RIPE Data Protection Task Force Group. We had a very long meeting on Monday about five hours, going through the legal matters around the RIPE database in a very thorough way.

Several issues were discussed. The RIPE database terms and conditions: Removal of personal data procedure and bulk access to the RIPE database.

The terms and conditions: We went through these very thoroughly, discussed all the legal implications and the wording in particular. In the end, the terms and conditions were approved by the data protection task force subject to a few corrections.

We'll publish the final document shortly after the RIPE meeting when we have done all the corrections and had a final legal check on them, because wording matters very much in these documents.

The main issues: All users will have to adhere to the terms and conditions. By user we mean anyone who queries the database, anyone who updates the database. We have been defined the hierarchy of registrant, maintainer and user and what their responsibilities and roles are. Only maintainers are actually allowed to update the RIPE database. The maintainers are responsible for the accuracy of the data which they maintain. We have defined the purpose of the RIPE database quite strictly, I recommend you read the document on this published because this is the only use by the RIPE database is allowed to be used for.

We have a link to an acceptable use policy which still needs to be finalised but this will be finalised by the time we have published the document. This basically lists certain parameters like the number of queries you can make will be unlimited, number of updates you can make is unlimited but obviously the number of simultaneous connections you can have the database and the number of personal data seccing you can retrieve have finite limits.

We have also, the RIPE NCC can now remove data. This is a legal requirement, particularly if someone says remove my personal data from your database.

The implementation of it: We'll publish the document shortly after the RIPE meeting. There will be a one-month period for comments. Then we will review the comments and the document will become effective.

Removal of personal data: Again, this procedure was approved by the Data Protection Task Force. The final document will also be published shortly after the RIPE meeting, subject to one or two minor corrections.

The main points: Any requests to remove personal data will first of all always be referred back to the maintainers of that data. They are the people closest to those, the one who wants the data removing. They are the ones who actually know who it could be replaced with.

The RIPE NCC will only take action if the maintainer doesn't respondent or satisfy the user request within a period of about two weeks. This is a legal requirement as the data control of the RIPE database, we must take action if somebody specifically requests that their data is removed.

Some perm data will not be removed. For example, if somebody has PI space and they are the only registered contact on that registration, and then they say they want their personal data removing, it will not be removed. We will not allow a anonymous registrations by removing personal data.

Bulk access: We have a new NRTM procedure now for personalist data. That is the data sent without the personal data. This was also approved by the task force. This is actually now in use. We have one or two people now using this new stream. There is a link there which is a starting point if you are interested in looking further at this.

No bulk access will be allowed for the personal data for any purpose. That's something which will change and we can discuss it later.

What next for the Data Protection Task Force? We need to publish these documents that we have now approved. We need to finalise the RIPE NCC's privacy statements. This is in advanced draft stage with you we need to review it in view of the terms and conditions we have now written for the database. Make sure there are no conflicts. So that will be publish in the near future.

We also need to rewrite the AUP for proxy access to the RIPE database.

Any questions?

CHAIR: Any questions? About the lifetime of the task force, it seems to me that a final report at the next RIPE meeting will take place and that's the end of the task force after all these documents have been published, approved and implemented?

DENIS WALKER: If we see no further use for it well yes, but if there are further uses

CHAIR: If there is further use, then either we continue the task force or we turn it into a working group.

DENIS WALKER: Yes.

CHAIR: If there is continuous use for this activity. A task force is created to do a task. It seems that the task

DENIS WALKER: Is almost complete.

CHAIR: Is [unclear] complete. So give it a thought at the, between now and the next RIPE meeting, whether there is enough new activities.

Wilfried.

WILFRIED WOEBER: Actually to that very item. We were asking this question on Monday to ourselves, whether we should sort of declare it as done. And there was the common feeling that we should at least still be in existence till 56, sort of to be available in parallel with the implementation and then my feeling would be that we would declare success and go away.

CHAIR: Till 57, you think? Next one.

WILFRIED WOEBER: Yes, till Dubai.

CHAIR: I just wanted to make sure that people have some, develop some thoughts about after completing these welldefined tasks.

WILFRIED WOEBER: I can assure you we do have that on our minds. Just another minor comment, what I like to you do or us as a community, is sort of to raise hands or hum or whatever, show the consensus that what this task force is actually proposing, because there was the word, these things have been approved. The task force is not in a position to approve anything. It's the community. So I would really ask you, or this community, sort of to give us the feedback after reading the document, that this is the way we want to go forward. Thank you.

CHAIR: Yes, meaning, that these documents are almost all ready for publication, if I remember well. They will be published with a onemonth period for comment and that's what you are referring to. Please read them and comment on them. Not only when you have critical remarks, but also when you think it's a great job well done. Okay. Thank you.

Jim you are next. Am you have a second chance which automatically means to well known undocumented rule, that you have half your speaker's time.

JIM REID: Thanks very much. I just want to give awe very quick update on what's been happening with the DNSSEC trust anchor repository task force has been doing.

This came out of the work that was done in the RIPE meeting in Tallinn and there was a need to find somewhere for storing DNS trek trust anchor keys and the working group was split down the middle about this and therefore a task force was set up to try and deal with it.

So the task force charter was essentially came up a set of requirements which could be used any kind of trust [unclear] the repository and of course that was not going to be necessarily be the NCC, it could have been IANA or any other trusted neutral third party, or whatever we came up. Then the idea would be the task force would report back to the working group on what the next steps would be.

So, the task force met a couple of times. There was discussion on the mailing list. And we came up with a set of requirements. We did some liaison with the NCC and with IANA keeping track of the IANA efforts to set up a task force sorry, a trust anchor repository and Daniel did a great deal of work behind the scenes, because we didn't want to do anything that would be seen to be undermining the IANA effort because that would have all sort of problems and we didn't want to try and exacerbate that issue by suggesting that perhaps the NCC would be in a position to do it and IANA shouldn't do it.

So, we have kind of been overtaken by events and I think we have got to the situation that this task force can declare success or victory. The ICANN board has approved the resolution to set up a trust anchor repository and the task force feels therefore we should support and encourage the IANA effort and not do anything ourselves and after all, IANA is the best place to do this kind of stuff. So what we feel is that the background set of requirements we have produced could be submit today IANA just for their information at purposes and that would be it.

So, we produced a letter of support. First of all we got unanimous consensus for that on the working group, on the task force mailing list. We then had consent at the DNS working group meeting and we have also had consent on the mailing list, as far as nobody has objected on the mailing list. So we think this letter is good to go.

So we have now got the endorsement from the working group to send this letter. Now, unlike the last time we sent a letter to IANA, we asked for support from the RIPE meeting. This time we don't feel this is appropriate because the letter says things that we have a commitment to help IANA produce documents and requirements in additional tools and things like that, that requirement is really from the working group rather than from the RIPE community as a whole. So I don't think we need to do that here.

So, really what we are going to do now is send this letter and essentially put the task force to bed because the IANA [unclear] should be coming on certainly in the next few months.

And at that, I am done. I would like to say thanks to all the members of the task force, especially to Daniel and of course to Richard Lambe who is actually making this thing a reality at IANA.

Any questions?

CHAIR: No questions. Then, I think on behalf of all of us, I thank the task force for a job well done. Not many people get IANA or ICANN to do anything nowadays, so...

(Applause)

CHAIRMAN: Thank you. Next on the agenda is a report from the Enhanced Cooperation Task Force.

MALCOLM HUTTY: Good morning everyone. I am going to talk to you about politics and I know that a lot of you get thoroughly board by politics and the rest of you probably hate T I have eleven slides and only ten minutes to deliver them in and at the end of it I am going to make a proposal that we don't go ahead with something just yet that is not supported by the RIPE Chair. Hands up who thinks I am going to achieve all this?

Here goes: First of all what is this all about? There is increased interest over a period of many years now from governments in the Internet institutions including the RIRs. Some governments are actively campaigning against the whole RIR model and much else about the way the Internet is run at the moment. There are some friendly governments and there are also governments in the middle that are kind of neutral on this and open to persuasion and will decide whether or not to support the continued model based on the how well we are seen to perform by them according to their needs. So we need to provide support to those that are friendly and to persuade those that are open to persuasion. Otherwise, if that doesn't happen, then maybe not today and maybe not tomorrow, but sometime you'll be ending up having address policy in the control of the ITU and your national regulatory authorities.

That's my pitch for your attention.

So, we need to demonstrate that the resilience of our system. The fact that it works. The fact that it isn't about to break overnight. That there isn't something that could go wrong at the NCC that would suddenly mean that the Internet would stop working. We need to show governance projects and [unclear] of our bottom-up model to urge governments to support it and accept that there is continuing work from the NCC, outreach work, to work with governments to support all those things and that's all been properly mandated through the NCC's own governance processes through the activity plan and the NCC board. But there hadn't been a history of detailed or of much overview of that by the RIPE community. And so that's something that we think that could be improved upon. And we also need to demonstrate, [unclear] a way that is visible to governments, community support, the RIPE community support forth NCC's position and all the RIRs position in the general Internet governance debates. That's the need and that's why the task force set up in the first place to address this need. That's why I am here.

So, here is our mandate. It's a limited role. We are a task force, we were asked to write a [unclear] it's not our job to conduct the activities with governments. And we certainly don't have authority to take decisions or to change the NCC's activity plan. Our report is to you, the Plenary. But it actually has a wide appeal. This report has the opportunity to be a good explanation from our community of who we are, why we are, and why we work the way we do and that it is working well. That can be important and valuable outreach tool to those governments and help us carry this argument.

So, getting started: It was a slow start, I have to admit this. We were authorised to start this task force at RIPE 54. But we didn't actually get through the stage of getting members, volunteering and in place until RIPE 55, the last RIPE meeting. It was a bit of a struggle. Because this work, this political stuff, is outside the usual run of what we do here in the RIPE community, and we were kind of learning how to make this happen in a way that was, that could be seen to be supported, because it can't just be about me saying we need to do this and going on and doing it. It needs to have the whole community support and to be shown to be so. So we need today generate the interest and the involvement and we need today balance the leadership, the idea of those that have been working in this field a lot and have been tracking interest field closely with the need to demonstrate that there is the whole communities involvements and transparency in this process. Because otherwise it's not worth anything.

That took a little time. But gathering pace. First question: I had to face as chair. Open or close task force membership. A very small number dedicated experts or everybody that wants to take part? There was strong arguments in favour of both. Essentially, a compromise, a political compromise. A drafting group, those that are most active and involved in the detailed experts in RIPE processes and, but open membership for the [unclear] task force, everyone can have a say. Our first outline draft was achieved in December 2007, that set out the framework for what are the report was going to look like. Drafting group discussions, then carried on over the course of this year and our first public draft has now been published, we had a full meeting of the task force on Monday, that I am reporting on now and our first public draft is available for download from the task force web page.

Now, you want to know what are we saying? What's this report going to say? Here is our core argument. The Internet community need drives the functional requirements. So the need to avoid address space conflict drives the requirements for coordination and then the policies and processes to ensure coordination to avoid address space conflicts follow from that. And then, the systems and the institutions are just there to support and enforce and make that happen. That is our argument as to how our community works. And it sounds, I would suggest to us, very obvious. We know that this is how, this is what the need is. We know that we are here, what we are here to do. But here is the cool bit: This is inherently a bottom-up model that is being described. Form follows function follows need. This is the opposite of a governmental style approach, the top-down model where you have an institution that takes powers and then looks around to assess the need to action and where to use those powers and where not to do so. Even a corporate model is a top-down model. You start off with a [unclear] you look at what its assets and skills are and then you look around in the market, where can we sell to make a profit in that area. So the top-down model is quite the opposite of our bottom up approach and this explains why it is so hard to explain our model in the governmental area. They are just not familiar with how this works. It also explains why they don't understand why we don't do things. Because we have powers, they are saying there is a need, why don't we exercise them. But we don't know that we haven't got the powers because our form follows our function, not the other way round.

Now, we believe that this core argument is a strong justification for our role and how we operate from first principles that if we can communicate this to governments, will make a big step forward in gaining their support and understanding for our position in the global Internet debate. Within the task force, we have achieved a strong consensus of support for this approach of this argument and I'd commend that to you.

But we were talking about enhanced cooperation. Where is all the enhancing bits? Well, enhanced cooperation is a term with a lot of political baggage in it from the world summit on the information society. Very briefly: The idea that Internet resources and the Internet as a whole is critical, a critical service for nations, creates demands from governments for proof of the resilience of all those systems at all levels. And an assessment of the social value of the Internet means that governments starts to demand provably [unclear] models from those that are involved in produce [unclear] and so those want to know what we are doing about that. How are we addressing the need to demonstrate, to prove the resilience that we have and how do we prove that or governance model is adequate to support a global resource. There is lots that's going on already. The RIPE NCC has governmental bilaterals, roundtable events, participation in multilateral fora and so forth. There is lots that's being done. But that's need to be described and shown.

So, much of the specifics in that area that we are documenting is best known to the NCC, that best discharges those functions rather than us in the community, that is best able to state the position and why it is the way that it is. Our report is not entirely finished yet. This is what's still to be done. A bit more qualitative assessments of the things that are, the outreach things that are being done by the NCC. The recommendations on those activities and how they may develop in the future isn't yet complete. We also came up with a draft recommendation for a working group that when we looked at it in the task force meeting decided it was simply fall formed for some serious political reasons we were strongly consensus formed that we should not use the term enhanced cooperation in a title of a working group and that we should not use the charter that we had put out as a straw man in the draft that you can review, download from our web page. We are not quite complete yet but we have got through the core arguments and most specifics and what's left is largely, or a lot of that is about getting details from the NCC on specifics that can be then further assessed.

So, the key recommendations from the task force to you are going to be:

First that the report should be used both to explain and to demonstrate the good governance of the RIPE community processes in action. We believe that this report, when it is finished, should become a RIPE document. So that it can be there as a statement of the community, and then that will mean that it will need a maintainer which probably means a working group. A forum will be needed on an ongoing basis, because a task force has a finite duty, to produce this report, but as Rob was just referring to with regard to the Data Protection Task Force, the forum will be needed on an ongoing basis to look at these matters to link the community's review of what the NCC is doing in the cooperation activities to ensure that that has continuing consultation with the community. So, a working group could do both those functions, be the maintainer of that document and that forum for review.

So we do expect to be recommending a working group soon. But not yet. Not until we have a charter, not until we have a title and a finished report. Since the task force meeting one of our members, Patrick, kindly came up with a proposal for an altered charter and an altered title but we haven't had a chance to discuss this properly yet. We haven't had a chance to speak to the suggested co-chairs there who aren't present at this meeting.

Finally, the reason why I believe that it would be premature to agree to that proposal today, is because there is incredibly strong value to the outside world in the statement, the RIPE community accepted the report of the task force and followed its recommendation to establish a working group. That sounds to governments like good governance. That up sounds like you have thought about this, you have followed a process with proper consultation, reached a result, reached a conclusion and taken action. If it happens in that order, that will be a very strong statement that will, in itself, be extremely persuasive, if we go ahead and create a working group today before the task force is ready to finalise its reports and against what I must report to you is the consensus of the task force, that they should not be created today, that will, I believe, very much undermine our ability to say that.

So, what are the next steps? I strongly believe that we can achieve a final version of this report well before RIPE 57. Probably in the next month or two. You can expect a fully formed recommendation for new working group from the task force, that we will, do expect to achieve a good consensus for well before the next plenary. In fact, in discussions with I don't know, maybe a dozen of the working group chairs last night, there was actually support for the idea of having a mini RIPE meeting, plenary to adopt this sometime between now and Dubai, but apparently for, there were other reasons why that isn't something that could go ahead but nonetheless, we believe that we can have ready forward you at the beginning of the next RIPE meeting a proposal to sign off on the creation of a new working group. So, please schedule we would like you to schedule time for a new working group at the next RIPE meeting so that we don't have any delay, but it should be not formally approved until the task force has reported to you.

Thank you and questions.

CHAIR: Any questions? I have a small question. What do you expect to be well before RIPE 57, end of next week?

MALCOLM HUTTY: Working group proposal or the report?

CHAIR: The report.

MALCOLM HUTTY: No. End of next month, yes. End of the month after, certainly. End of next week, that's a bit tricky. We need further help from the NCC as well. That will probably involve a physical meeting with the NCC. There is some practicalities there, but this is far advanced. There is a public draft available, that in itself demonstrates progress. From a governmental point of view, I'd suggest that actually having a final outcome isn't the ultimate need here. In order to demonstrate that there is work being done here, what they need to be aware of, be certain of is that we are following good governance process and if we show we are having a proper open consultation on this, that is something that they will buy.

CHAIR: Right, I have, as Marco mention indicate his introduction, a slight problem here. I think in the first place, this looks all very good. I think we can, I can speak on behalf of all of you that we are very pleased to see this well thought of report in the making.

(Applause)

I have a slight problem with the timing, and I know I am not alone. I would have preferred to see a decision at this RIPE meeting on the creation of this proposed task force. I have seen a draft chapter drawn up by some of you with a short but very clear working plan in it with proposed chairs. The advantage of creating a working group where it is very clear that this report has not been finished yet, but it is leading to the recommendation to create a working group to do it now is that between now and the next RIPE meeting we have at least something to show to the outside world that we take it seriously. The proposed name of the working group is not on enhanced cooperation, but it's something like the RIPE collaboration working group. So I would like to get input from the community. Approval of accepting this document, we do usually, not necessarily, at a plenary session at a RIPE meeting because not everybody potential potentially interested can always make it. The RIPE mailing list is our usual way of taking decisions and that can take place any time between now and RIPE 57. That's why I asked, next week or so.

Input please:

AUDIENCE: I seem to remember that one of the mechanisms we have used for transitioning the introduction of a new working group in the past, is to create a BoF when the need becomes clear about all of the piece of the Lego are not yet ready, and at the following meeting, transform that happen BoF into a working group and I am wondering whether that might not be a way out of the different perceptions of the timing and readiness that seem to have been expressed?

MALCOLM HUTTY: Sounds good to me.

CHAIR: Not to me, because a BoF usually is a once in a lifetime rather informal hoc meeting, usually in spare time during a RIPE meeting. We are further ahead here. We have a task a BoF usually ends up in a task force which might end up in a working group. We have a task force already which is about to finalise its work so a [unclear] would be a step back I would think in this chain. We can think of something maybe along your lines and have a working group information, a working group being informed. Okay, I see a couple of other people.

AUDIENCE: I would agree with Rob on this. I think, you know, saying that we have to wait for the results of the task force, otherwise the Government will think that we are not following a process because it looks good. I think that's I don't think that statement makes any sense. I think this issue is too serious and I think we know where the result of the task force are leading and I think the establishment of a working group at this stage is incredibly critical. I have been involved in regulation, Government regulation in my region extensively, and you know, if you don't do things in time, you will find that governments have already established preconceptions and to then later on and come in and try and make those preconceptions and try and support the bottom up approach is going to be extremely difficult. I am not saying it's impossible but it's going to be extremely difficult. Why wait? Why wait until RIPE 57 to make this decision when we can make this decision now? I say let's make the decision now we need a work group.

AUDIENCE: Andy Davidson. I respect your opinions Rob, but I feel more aligned with what Malcolm was saying, if I'm honest, because I think that clarity of our message is very, very, very important if we want governmental departments to actually take, listen to what we are saying. I agree with Malcolm's views that doing things in the right order is respected by the people who are the intended audience. And the other point that Rob was making that we need something to show governments that in programme we have the report. We don't need to show them that a working group is being formed in advance of the publication of the report. We have got the report moving through its draft stages to there is stuff we can show and

CHAIR: A few comments. A working group is a rather permanent feature in our environment; that shows seriousness to governments. A working group usually is in charge of maintenance and creation of crucial documents of which we have one here, which is almost finalised, but the task force proposes that the maintenance of this document is being handed over to the new working group. And proper procedure, it has been mentioned several times. Let me remind you: We have wellestablished procedures in how we internal RIPE governance and many them of have not been documented but have somehow grown over the years and usually working groups are created when the chair of RIPE feels that there is a need in the community that there are enough people who want to be actively involved in participating, that there is a draft charter and proposed chairs and I think and then the Chairman puts it up to the community as represented here, and the community decides. And I think all those conditions, at least I see them as being fulfilled. Patrick you are the next one.

AUDIENCE: Patrick F�ltstr�m. I think first of all I would like to say that I think the summary that Malcolm was putting forward is correct and also read the minutes of the meeting and I think that also expresses the concerns of the task force. That said, I don't see personally any reason whatsoever to do these things, creation of the working group and the publication of this document in serialised. I think we need both the report and the working group about seven months before today. So we already long behind. Yes, I wrote a proposed charter for the working group, and of course there can be some changes in the text itself. When I talk to people and also got only unfortunately two responses, or maybe only one response in the task force mail list which I think also matches the consensus in the document. I think the document that the task force produced is so close to be finished that the working group very well can have that as an item to have that created and finalised. I don't see now risk from a Government perspective to misunderstand what we are doing these kind of things and I completely agree with our friend from Bahrain that delay in either of the two action items would be a great risk, begin the serious meetings that are coming up this year.

The last thing regarding the proposed working group chairs. I think that's a separate question than creation of the working group, because a charter can be updated and the chairs can be replaced. The important step and at the significance that I would like to see at this Plenary is the decision on the creating of the working group because I don't want to wait until the next RIPE meeting. I have been in contact with both proposed working group chairs. I know that neither of them have got commitment on the time and also heard some concerns from people in this room that support my idea but they have concerns with the people themselves. I think that can also be resolved by cooperation between the working group and the RIPE Chair. So I am strongly in support of creation of this working group.

AUDIENCE: I actually believe that we have, as it is put on the screen already, we have a strong statement here and this is almost close to actually establishing a working group here right now, but it just postpones it to the next RIPE meeting and maybe even to something in between. So, I guess the decision making, are we going to have a working group but not right now as of today. It gets clear from the statement from my point of view that there will be a working group and this message can be sent out to the community and even to governments and all sorts of other bodies, but I am with Malcolm here, as in it should not be established right now as in the process from the task force to a working group is not entirely finished yet. So...

AUDIENCE: Jim Reid. I think the question here is initially one of timing. I don't believe anybody here at all disagrees with the idea that working groups should be created, the yes is do we do it now or at some point in the future. I understand that so. Arguments being made by Patrick and ma had about there is a need for this thing to be done as a matter of of some surge see and there is Government concerns and perhaps they should have been done sometime in the past. But, I have heard no reason of nothing so far that would indicate you know why would another month's delay make a difference. If someone could convince me we have to create this working group today and there is an overwhelming compelling reason for it then that's something would I listen to. But so far I have not heard that argument made in a way that makes sense.

CHAIR: I think that had you want to reply to this.

AUDIENCE: Actually do I want to reply to this. I am really sorry [unclear] in. I'll tell you why another month will make a difference. Because we said we want to give the RIPE message to the Government. Is this message that another month would make a difference? Is that the kind of message you want to give the governments, that we don't really think a month is significant? I think a month is very significant and if I think it, you can get your bottom dollars governments are going to think one month is significant.

MALCOLM HUTTY: Governments are going this think that following an open and transparent process is the most important thing, to have confidence in our model.

CHAIR: That is what we are doing right now. This is our governance model. To have an open and free discussion and come to a practical conclusion.

AUDIENCE: A really quick comment on trying to convince Jim. Okay. One of the main task of this working group from my point of view is to give a permanent mailing list specifically and also meetings at the RIPE meetings where the RIPE NCC can verify with the RIPE community that the statements that they are making on direct questions they get from governments are actually in line with what the RIPE community believes. Okay. At the moment, there is no such venue and the more we delay the creation of the mailing list or whatever is needed for RIPE NCC, the higher the risk that the RIPE NCC are making statements that are not in line with what RIPE community believes, which means that we risk that RIPE NCC is not following what is outlined in the document. That's my main reason.

AUDIENCE: Remco, speaking as a concerned citizen this time. I have got a problem that the problem is that I simple sympathise with both points of view. I think thinking about T I think we have a timing problem and I think the timing problem is actually evolving from the point that for some reason, perhaps it has been sneaking in over the last few years, apparently we as a Plenary, can only discuss and come to consensus twice a year, namely at a venue like this one and I think that's actually the core of the problem that we are facing here. So, and that just occurred to me during coffee is why isn't the Plenary in the same organised in the same way as the working groups are. So why isn't there a Plenary mailing list for discussions or if there is one, why aren't reusing it?

CHAIR: Okay. Thank you. Hans pet err, newer anney, Daniel.

AUDIENCE: It was mentioned here that we should follow the right process. It's a question to me whether that means following the RIPE process or the governmental process. It seems to me that we are trying to satisfy both. I think in the past we have established working groups with less of a well defined charter than this. So, trying to compromise here. My proposal would be that we form the working group conditional to the finalising of the charter and conditional to commitment from the chairs. As these are the two things. Then the principal decision will be taken to this meeting, but it will also satisfy your requirement that yes, we need to finalise some of these documents. So that means that we will make the principal decision now with some tasks to be finished before it's sort of formally there. That means that we can go on create a mailing list and go on and start the work and go on and get commitments from bosses and so on and maybe it's a chicken and egg problem. Maybe it's not possible to get commitments from the bosses before we form the working group. That would be my pragmatic proposal.

CHAIR: Thank you. Nurani?

AUDIENCE SPEAKER: As a member of the task force as well, I think my feeling of the task force was that we wasn't that we were going to wait with getting any work done. I think that's important to point out. That awful us felt that we were willing to commit the time and effort and that it was important that we didn't just let this die out or let this lie until the next meeting. I think the discussion about the working group, the there were proposals about forming a working group but there were concerns about performing a working group without having a charter or a chair. I think whatever way we go for, whether we continue within a task force or whether we form a working group, I think everyone agreed that we want to get work done between now and the next meeting. Weigh want to make sure we share the drafts of these documents. We get everyone's input so this is not a closed process and that we have some clear results to the next meeting.

CHAIR: Daniel. Then Neil and then I think we will try to come to a conclusion.

AUDIENCE: Daniel, Internet citizen. I'd like to support ance bet aproposal an refine it somewhat. I think we should actually make the principal decision here. I don't think I didn't hear any disagreement about the fact that we need this task sorry this working group, a permanent thing, we need a permanent forum to discuss these things and as Patrick has said quite clearly, to you know, align the points that we talk to governance about and be sure that you know, we speak consistently and the RIPE NCC speaks consistently for us when they do this work. I think so, to me there seems to be consensus to do it.

I think we can do it at some point in the future to sort of project due process and the way I would want to modify Hans's proposal is to say, to establish consensus in the Plenary that we actually authorise Rob to do this, because Rob's always done this, you know, the chair has always created these things. At that point in time, when he thinks it's appropriate, the work has been done, the charter has been really hammered out and the commitments from the chairs have been verified, and if we do that and we can do it whenever Malcolm and the task force are ready, and Rob can just do it at the appropriate time before we meet again in Dubai. (Applause)

CHAIR: Thank you Daniel. Neil.

AUDIENCE: A quick comment on Hans's suggestion and Daniel's elaboration that. I think that's a much more thought out idea than the one that I suggested a few minutes ago. I'd like to add a more historical note but I seem to remember being at some stage a number years ago, the chair of a new working group whose charter had not yet been completed and to which was the finalisation of the charter was one of the first [unclear], so we don't have to wait until every single one of the ducking are in a row. They are coming together. We can move.

CHAIR: So, I think I will take up Hans's, Daniel's and Neil's suggestion and put the following question to you:

Is there agreement that we need a working group on collaboration?

(Applause)

That's one strong message, I would say. The rest is a matter of implementation, and I would like to take up Daniel's suggestion, and I will explain how I interpret this.

There is a draft charter, which not many of you have seen, so that has to be published. The mailing list is the RIPE list, that is the most wide mailing list we have. And this is a RIPE wide discussion. Next week the draft charter will be published, which also contains suggestions about chairs. Also, next week, Malcolm will send an email to that list pointing out to the existence of this document and where to find it. The community will be invited to submit comments, both on the work group issue, not the existence of the working group, because that we have decided now. But the wording of the charter, the contents of the charter, and the proposed chairs.

There will be, I think, in all fairness, a four week comment period, and then the chair, me, will, after consultation with, for instance, the authors of the charter document, conclude whether we have a working charter and if that is the case, the coming into operation of the working group will be announced with, in the meantime, the RIPE NCC will be asked to set up the usual backoffice support mechanisms.

This document, I completely fully agree with Malcolm, it should have the widest possible circulation and the widest possible participation in not finalising it, but supporting it. Why don't you follow a similar track: Publish whatever you have now with annotations of "This is still being this part is under development and this, that and the other" or just leave bits out. Ask for comments and support, and put time line for yourself when version 1, as an official document, will be published. Hopefully then by the working group.

Is this acceptable?

MALCOLM HUTTY: I have a question: If the working group is going to be are you going to wait until the task force has produced that version 1 document to establish the working group or if not,

CHAIR: No.

MALCOLM HUTTY: Is the task force going to be dissolved at the point when you create the working group.

CHAIR: No.

MALCOLM HUTTY: So is it going to remain the responsibility of the task force to work on this

CHAIR: To finalise this document.

MALCOLM HUTTY: To bring it to version 1.0.

CHAIR: Unless the task force decide, as task force we all join the working group and we finalise it there. This is, for me, totally irrelevant under what heading you do this. More important is that the work on this document is finalised. Hopefully in a timely manner, meaning four weeks or so.

MALCOLM HUTTY: Okay. The task force will undertake that work then.

CHAIR: Nobody is proposing to interrupt a work in progress, that seems to be going very well. So...

If we do it like this, then we have succeeded in concluding two things. One, we have created a working group which will become operational in four weeks from next Monday, or something like that. After no, four weeks after publication of the relevant documents. Fine. Thank you all and thank you Malcolm and thanks to the task force.

(Applause)

CHAIR: You may have gathered that we are running a little bit late. I know. So we, without further ado, continue with the next unscheduled item on the agenda. Daniel:

DANIEL KARRENBERG: I know I am between you and the only really entertaining part of this. So it's just to note that the anti-spoofing task force we wound up at the last meeting and we promised within the following few weeks to actually produce the documents or publish the documents for it. I just want to say that we have finally published those documents. They are RIPE 431 and RIPE 432, one which one about how to and one about how you convince your manager so I just wanted to make the point that we have finally now finished this. The documents are out there. So if you want to do something about reducing sort of the address spoofing, you now have read documents you can point to and I understand they are not in the list of documents yet, but they are there if you know how to get to them and the list of documents will probably be updated in the next hour or so. So, thanks again to the people who worked on them. We have the documents.

CHAIR: Thank you Daniel.

Next on the agenda is, it's not yet the entertainment part Daniel was alluding to, though some of us think Google is a great game. Especially if they try their hands on IPv6. Lorenzo will explain some aspects of Google's efforts in the IPv6 deployment.

LORENZO COLITTI: I am standing between you and lunch, flights, whatever it is, wherever it is you need to be. So I'll try to make it very quick.

So, why am I here? To tell you about how we did it and to tell you about why we did it and to tell you about the lessons we learned while we did it. And where do we go from here.

So, we see IPv6 as critical. Our mission is to organise world information, make it accessible, accessible to everyone. So we are now users in the IPv6, we must be ready to support IPv6. And IPv6 is critical to the growth of the Internet. V4 is running out yes, mobile devices need to talk to each other. Appliances need to talk to each other. It might be a good time to reclaim end to end. That doesn't work, breaks, anything that's not simple client server, development of new applications because people have to work around T and it's generally a bad thing.

So, what have we done? We organised an internal IPv6 conference and we invited a few speakers. The most important thing, apart from participation, is that we made the server change of our IPv6. I don't know if there are other ones. I think not yet, but hopefully someone else will too. And we are work on IPv6 and we hope to offer more.

You have probably seen T it has this nifty animated logo. We had to do that. It was leaked during the Plenary of the ITF, it was leaked at some improbable hour in the morning, we don't know how, but it has not yet been announced, we should announce T what is it? It's a first step but sometimes a first step is the hardest. It has imitations. It's only search for the moment but there are creative ways around this. If you search things you will find them. You can get gmail, you can get calendar. It all work. It only works on IPv4 sites. It's okay for now.

It doesn't work perfectly on a v6 only connection. Some little things break. But hey you can't get to the search results anyway, so it's not very useful on a v6 only connection.

It's on a separate host name. Why? Well, because if we put a quad A for a google.com then we break our users and we do not break our users, ever.

User response: People generally happy. A few a few blog post and interestingly, some people were saying, hey my v6 is faster than my v4, other people said this is how you hack it to reach email. This is how you use it to each v6. Can I have reader this and that over IPv6. Well we are working on things.

What did we learn? This is mostly operational lessons. The routers that we used, the device that is we used, they work. They work pretty well but we don't have feature parity yet. We can't filter on extension headers, lots of hardware at least, lots of hardware out there will look at the first header only. That means if you have a TCP packet behind a routing header or behind a fragment header then your router sees it as a fragmented pack the and cannot for example know if it's distend to PCP port AP, it can't look at T what do you do, if you want to do any fire walling at [unclear] you have drop it at the edge. So, that's quite a problem.

So, that needs to be fixed and that's probably a hard problem because it might [unclear].

Proving on we tried to get NAT PT to work. We gave up, but we now know it's possible and that's great. But the implementations are still rather temperamental. It was apparently very hard to get the RIPE meeting configuration to work, so kudos to the people that manage today do that. A lot of implementation don't have support for 6 to 4 Teredo. If we want to use them, we are going to need hardware support for them.

Load balancer support isn't mature yet. VRP isn't supported. On [unclear] load path MTU didn't work, this was a showstopper when we first deployed IPv6 at the internal conference. We manage today hack around it, but this is a very bad problem.

But, you know, it's not rocket science and it works. This is enough for our initial deployment. We can live without this today, because traffic is not comparable to v4, we can. It works and it's supportable. But, if we need to scale to high volume, then we need these features.

Reliability: Okay. So our load balance has leaked memory. If you have well, I won't describe the bug, but so, on the eve of the launch, three routers in two continents on a two centres crashed within minutes of each other. The vendor response was in certain rare conditions, a certain type of router X may crash when finding the best match for a specified prefix" three at a time is rare this crash is more likely to happen over v6 because your prefixes are [unclear]. Well, what do you know? Yeah, they are longer, they are bigger, but whatever. If you want to deploy something you expect it to stay up, you might want to consider using dedicated v6 divisions initially. A word to the wise. Maybe your gear works perfectly.

Internet working. If you do, since your hardware can't filter on anything that contains an extension header you have to drop packets with extension headers at the edge. Well, so fragmentation doesn't work any more. That causes empty black holes [unclear] sent big packets there is to way to get them through the edge filter. You can get [unclear] to the [unclear] minimum of 1280 and you can get it done.

Routing is patchy at the moment. We get indiscriminate transit, we get IS paths that go all the way around the world. This slows convergence because BG needs to explore everything. It increases latency and we don't like high latency. People like to see Google come up fast. So, we don't like that. Oh, block holding. On the day of the IETF launch our /32 wasn't visible through the ITF network, so of course when people are in a room and talk to each other, they fix it very quickly, but this is something this is something to be aware of. There are tier 1 networks out there that say we don't peer with anyone. But oh by the way we can't get to a lot of v6 Internet because we don't peer with anyone. So they have incomplete routing tables. I have seen people with 700 routes in their routing tables and by the way they don't have hours and when they are customers complain to them, I am forced to say I have no sympathy.

Inter domain performance is unknown. We have seen reports of stuff that's better within v4. We have seen people, there is Hurricane Electric has been doing some measurements on this and it doesn't look too bad. But we need data on this. We need to find out how the Internet really works.

Tunnels, tunnels are bad, they increase late tee, they make debugging very hard. We should avoid them. Please try to avoid them. Particularly if you are main traffic. Because BGP is extreme re crude and it doesn't get you where you want to be as fast as you can. It's priority is just policy. So...

6 to 4 Teredo. Personally I don't like these. They are performance can be suboptimal because if you do it he employ relays in your network you can out optimise the outgoing path but the incoming path you are still at the mercy of whatever 6 to 4 Anycast relay the user is going to use and I think my colleague has, is in California and he use as 6 to 4 relay in Italy and that can't be right. Also, they don't provide stable addresses. I think personally one of the biggest advantages of IPv6 is global reachability with stable addressing. So you know where to find whoever you want to talk to. And these methods to do not provide stable addresses. Teredo in particular your IPv6 address depends on the port of the NAT that you are talking through. And it will change as soon as that port changes when the NAT device is rebooted. Even if you have a static IP address.

Operations: This is important. We must dispel the notion that IPv6 is experimental or testing. No. It's a production service. It must be monitored. It must be supported. It must be designed to the same qualities as IPv4. It doesn't have to scale as much as IPv4 but it has to be the same quality. So how do you do it? You talk to your knock and you make them aware of IPv6. You scale it down but you don't skimp on reliability and one thing I found to be true [unclear] need to design as close to the IPv4 network as possible, so that people won't be surprised and they'll just look at the network and they'll not what to do already. So design it similar to IPv4 as possible. Make the principle much least surprise work for you.

Apologies for talking at such a high bawd rate, but I was asked to compress my presentation. Where do we go from here? This is more speculation on my part. I think one thing that's important is increase richness of connectivity, peering, peering, peering. Thereto a lot of networks that don't have very good connectivity and there are some networks that do have some very good connectivity and at the moment, one of the possibly ways forward is for those networks who do have good connectivity to collaborate with each other. Peering maybe and then perhaps even exchanging services on a bilateral basis. Personally, I think NAT PT is essential, but others may disagree. NAT PT will give us ways to ease the address crunch, let's face it in 2012, I don't think we'll have universal IPv6 deployment. I wish we did, but it doesn't look that likely.

One of the advantages of NAT PT is that you decouple clients from servers. You can establish IPv6only networks, client networks and the content side doesn't need to be involved yet. So, you break the chicken and egg argument of edge network saying hey there is no content and content saying hey, there is no users. So you can deploy users first and then content can move to IPv6 at their own pace and when quad As are published recollect guess what, people will start using IPv6. When this happens, NAT just goes away. This requires mature NAT PT implementations or at least [unclear] ones. [unclear] settle for functioning ones, which we do have.

So, what do we need? We need MPLS traffic engineering. 6 PE is not a good solution because when your tunnels go down your traffic it black hold. That sounds bad to me. I would rather have native connectivity. On the other hand, if you don't use 6 P you don't have traffic engineering because you are presumably doing it on shortest path. So you might need that. We need extension header filtering and hardware. MPU blackholing is bad. The data centre, we need VRRP, because neighbour unreachability section works but it's a few seconds slow and for production quality we need faster than that. We need a if you hundred milliseconds.

What else do we need? On the user [unclear] we need NAT PT that works. It's deprecated. Well, maybe we can agree on something. What I think we need is a barebones NAT PT standard, not all singing, not all dancing. It doesn't need to solve all the problems in the world. All it needs to do is to be functioning equivalent to IPv4 NAT. The Internet works with NAT and it seems to work at the moment. Since NAT is broken anyway we can have NAT PT that's the same level of brokenness and still work.

Undeprecate our RFC 2766. [unclear] get shot down. People are going to implement it anyway. For users we need heart err support. Employing these will make a dig difference to connectivity for people who go through these.

The real change that faces us is how do we adopt IPv6 while maintaining a quality service? Because we try to make sure that the website is available for everyone. And works quickly for everyone. So putting a quad A on the website is not the solution today. It would completely break a few of low percentage of users and it would increase, it would increase latency for lots of users and we all like our Google searches to come down fast. It would mean lower reliability for a lot of users. For us it's unacceptable to break users. We just can't. So we can't publish quad As and this is the reason why we launched ipv6.google.com.

We need to try to measure the IPv6 Internet, see how big is it, how fast is it in how many users have broken connectivity. I think this is very important. I think that the work that Geoff Huston has been doing is a step in the right direction. And also, I think that if we want to accelerate IPv6 deployment T might be a good idea for people who are actually really reemploying it seriously as opposed to people who have lab networks to talk to each other, peer with each other and try to exchange services on IPv6 selectively. Because when two networks directly connect, problems can be fixed. Quality of services can be guaranteed and both networks gain by the experience by actually getting operational experience on real traffic.

So if someone wants to do that. Let us know. We are here or you can write to us. You have my email address and we can collaborate.

And yes, that's it.

CHAIR: And we have time for two questions. Daniel first. Okay.

AUDIENCE: It's not a question. I have a suggestion. Fredy Kuenzler, by the way [unclear] for real fast v6 deployment or if Google would claim or actually do a higher page length for v6 enabled websites, I mean, everyone would want to have v6.

LORENZO COLITTI: Page rank is what our users want, yeah. We don't touch that. We never have. So...

AUDIENCE: Thank you Lorenzo. I would suggest to everyone, and this was you know, fairly good and detailed and everyone who has seen this, go back to the slides and I am quite sure Lorenzo will talk to you at length to explain some these things that may not be immediately obvious. But we have to be I would want to warn those who have seen this now and to whom it might look very complicated and look like there is very many stumbling blocks on the way of doing this, you shouldn't take away this message. This is just an honest thing and has lots of details in there. But none of them are really show stopers, and that's something

LORENZO COLITTI: Absolutely.

AUDIENCE: I think that's something that needs to be said. One other misconception I'd like to warn against, is where you say you make richer v6 connectivity. That shouldn't mean make more tunnels. I mean you said non tunnels and things like that. I think that's meant there, and it is in some of points is make it as solid as v4, make it follow [unclear] as much as possible and take it seriously, and not sort of make, you build a sort of lab or overlay network of v6 but make it you know the same quality as v4 wherever you can and the. So whenever you read make richer v6 connectivity, you should read it as make it production quality, please.

LORENZO COLITTI: Absolutely. Maybe it wasn't that clear. It was mainly operational lessons we learned, but, yes, the take home message is here is what Daniel said. We did it. It's not rocket science and it works. Let's all do T that's the message, one of the messages and yes, don't do tunnels. I mean, do it properly. It's not that hard and it really isn't.

CHAIR: Okay. Thank you very much.

(Applause)

We are nearing the end but between you and your lunch that is always this interruption from the secret working group.

OLAF KOLKMANN: I was wondering, how many people know what's coming? So that means that we have some explanation to do about the exception that exists within the RIPE community and the RIPE processes.

There is a group, and this is the group, without a charter, without a chair, without bottom of process, without practical conclusions, without open discussion, with links to governments, industry and the intelligence community. This group does not work with volunteers, does not meet its milestones it does meet its milestones because there is a presentation all the time, it's a group where it's acceptable to break users and definitely it does not compress its presentations. You will be here for a while. And now I am going to check if the sound works. That's good because we need that further on.

So, for the people who do not know what this group is. This is the secret working group and I was very surprised to see that, as always, the secret working group adapts its methods of getting the material to us, to the location we are and material was thrown over a wall. And remember, please don't ask questions during the presentation. RIPE meetings are on quite headed. When policy things are defeated. But if you feel stuck for time, then just trot out this line "I am not finished yet, Jesus be seated."

SPEAKER: We are told that the pool of v4 is more shallow than ever before. When you try a long dive, and you want to stay live, better hope there is a shovel next door.

SPEAKER: This is not funny yet, but this is all about policy process. The policy process we have now for pour is in need of a shovel.

SPEAKER: My honey she called me and told, a statement bought [unclear] she was bold, no matter the size, I should realise, no others than hers I should hold.

SPEAKER: So, as you know, the association, yesterday, started their election for new board members. Since that room was needed there, they are now in the seller.

We are now in our 10,000 enter round. Proxy votes it seems are not bound. With 5 out of 4 It remind us once more That IP engineers cannot count.

The commercial break from Zattoo

As welcome as warm hippo poo So uninstall flash Make their PowerPoint crash And then go for a distinct at the zoo.

SPEAKER: Talking about hippo poo...

One day at the Berliner Zoo A hippo, while smelling her poo Said "Look right" oh dam There is a perv with a cam Let us stop the things that we do."

A gentleman came from BT And was slightly astonished to see That if he didn't pay Then I'm sorry to say That his LIR might one day cease to be.

Berlin's Grand Palace opened wide its doors To geeks and nerds who came from far and wide to meet and talk and fill its many floors and gorge themselves at food from every side. One evening we were getting in the mood, by gazing at the lumbering beasts all around as bellowing they drank and took on food Whatever and wherever it was found. The lights grew dim as the twilight fell and looking through the glassy side of pool, we watched the beasts disport themselves so well, relaxing in the [unclear] an evening cool. And we, the hippos, civilised and fair, smiled out at the professionals out there.

SPEAKER: Imagine you are in the test traffic working group. We now have everything on the screen so you can start talking.

(VIDEO: "Master we have mail for you.")

Have you run out of new things to do. RIPE TT needs a project or two. But in blackness of night Comes a voice, warm and bright Saying "Master, new mail for you"

And that was that. I see you in Dubai.

(Applause)

CHAIR: Right, we are almost done. One of the last additional items on our agenda is it's prize time. For those of you who are new. If you register yourself very early for a RIPE meeting and you are among the first three, you get a prize. And the prizes this year go to number 1, that is me, and I am always skipping because I never register seems to be a feature of the software. So, whenever I step down, it needs a complete rewrite of a line.

Number 2 this year, that had, you have been extremely active, not only you got yourself elected on the board, but you got yourself a prize.

The next one is Leo Vegoda. No, these are people who pay great attention to RIPE NCC announcements of RIPE meetings and take immediate action. We like such people. Leo.

The third on the list is Remco. I know personally that he has been struggling for years to reach this.

REMCO VAN MOOK: This is a highlight in my career. Thank you.

CHAIR: A few more prizes. I think the RIPE NCC was running a business centre this time for those of you who are not newcomers, you will not recognise it, because it used to be called the Hostmaster centre for many, many years. It's all the same but with new branding. I did something that results in a prize.

Thank you to everyone for coming. I hope we manage today answer all your questions, and I hope you'll come again, see us in RIPE 57. Number 8. Are you here? You should have a blue ticket. Number 9, are you here?

(Applause)

CHAIR: Thank you. And now my list becomes more and more mysterious. We have a v6 hour winner. I cannot imagine what that [unclear].

We disabled the v4 wireless network. During that session, we had about 120 users with the v6 test. Almost every attendee in the session applied in the survey and we have one winner for that draw. And I think you have the name actually Rob.

CHAIR: It says we have one v6 hour winner. Those are my instructions, sir.

NICK HYRKA: We do have the name I apologise, we will post it immediately. Bradley Freeman.

CHAIR: What did he do to win this?

NICK HYRKA: He just filled in a survey and his name was drawn at random.

CHAIR: Some of you may have noticed that, or may think there is still one presentation left on the announced agenda. You are right and you are wrong. Because very oningcaption issuesly is not only a very good host to a RIPE meeting, but understood that we are a little bit pressed for time. So he kindly offered to give his presentation at the next RIPE meeting in Dubai, where it will not be a presentation on the proposed new configuration, but it will be a description of the new configuration in operation, is that right?

Once again, we want to thank DECIX and for being excellent hosts and that is part of the success of this meeting. And we have a little token of our appreciation. Thank you very much.

(Applause)

Last thing that remains to be said is we are coming to the close of this meeting and we have the next meeting in Dubai on from the 26th to the 30th October. Dubai is a lovely place in a lovely part of the world. I think if we had more time, I could have asked that had to say a few words there. But I hope and expect to see all of you in Dubai. Some of you might think it is too exotic to go to. For those you, rest assured, have a look at the RIPE website, the information pages for the next RIPE meeting are up an running already. You will find and through the coming weeks and months you will find more and more information there about visa requirements, about hotels, about everything that is of relevance to entice you to come to the next RIPE meeting.

And there is great shopping.

Judging from this RIPE meeting, I am sure it will be again a hectic RIPE meeting full of discussions because a lot of things we have discussed today are not been concluded and I am sure between now and for or five months a lot of new items will come up. So do come to Dubai.

Remember, however, we always start RIPE meetings on a Monday and finish on a Friday, leaving the weekend free. Weekends in the Middle East fall on different days than we are used to so the RIPE meeting starts on a Sunday, and finishes on a Thursday. Keep that in mind when you are planning your trip. The RIPE week is slightly shifted.

Nevertheless, it's not rocket science. I hope to see you all at the next RIPE meeting and unless somebody has a very burning question now, I say thank you for coming. Have a nice time if you are staying on in Berlin. Have all a safe trip home and until next time thank you for coming.