Connect with RIPE 60 : Facebook Twitter RSS linkedin

These are unedited transcripts and may contain errors.



Notice: Use of undefined constant steno - assumed 'steno' in /var/www/html/ripe-60/steno-transcripts.php on line 24

The closing plenary commenced on the 7th of May, 2010, at 11am:

ROB BLOKZIJL: Good morning. Welcome to the last session slot of this 60th RIPE meeting. I see we have quite a few survivors of last night's dinner here, still more are coming in. If you could all find a seat. Thank you. As usual, this last session is a collection of a couple of small items but we start with the last presentation from the EUF and I would like to call Anand bud he have to give his presentation of the 32?bit ASN in K?root service network.

Anand: Good morning, I am Anand bud he have and I work for the RIPE NCC and manage DNS services at the NCC and I am here to give you a presentation about our experiences with a 32?bit ASN. So, first a little bit of background:

The RIPE NCC manages several DNS zones, forward zones including ripe.net and its related zones. We also run the parent zones of all the reverse†?? all the reverse DNS zones for the address space that we allocate. We also operate the ENUM zone and we provide secondary DNS services for lots of other zones, including the reverse DNS zones for the other RIRs.

At the moment, we have very modest infrastructure for this. We have a pair of servers which we call NS Pry and NS Sec and they are both based in Amsterdam. NS Pry has all the reserve zones of the RIPE NCC plus forward zones and NS Sec is the zones where we host zones from APNIC, LACNIC and AfriNIC. And these are just single servers; there is no clustering, redundancy or anything like that at the moment and they are unicast so they have this single addresses from the RIPE NCC address range.



So we are planning some changes to this infrastructure. The motivation came a little bit from ICANN, this is soon going to redel late arpa zone and we decided this would be good time to change all our infrastructure and make it better and more redoesn't and highly available so we are going for multi?server architecture where we have multiple serves behind the routing using OSFP load balancing. We plan to do Anycast with this so make it highly available. And as I mentioned we are going to also serve the in?addr.arpa zone from these serves along with all the other zones. And to make things easier, we chose to have a separate autonomous system number for this. This is so that we can have an independent routing policy, independent of the RIPE NCC network.

So, we applied for an AS number like all the other LIRs do; our request, of course, has to be approved by the arbiters and, after some questioning and back and forth, they approved it, and we got assigned our 32?bit ASN, 197000. So a little bit about interoperability of AS numbers:

Older BGP speakers which don't understand 32?bit ASNs can stick talk to routers which do have this support and what happens is that they that I they are talking to AS 2, 3, 4,5, 6, and what the new router does or the 32?bit capable router does is that it puts the AS path in this attribute called AS 4 path, it's a transitive attribute which means the older BGP speakers just pass it along. And end routers that receive prefixes can unravel the correct AS path using this attribute. These days, most of the major vendors, router vendors such as Cisco, Juniper, they all have support for 32?bit ASNs in their routers and their operating systems.

So, some details about how we are connecting this stuff:

We have a router at the Amsterdam Internet Exchange. It's a Juniper router running the latest stable JunOS. We peer with 6 providers who /TRO provide us transit for IPv4 and IPv6 and all of these support 32?bit ASNs so this made it very easy to configure sessions for them and for them to configure sessions with us. There was no special configuration to do.

To keep things simple, we are also peering with the Amsterdam Internet Exchange route servers, they have a pair of them running a recent open BGPD and this has support for 32?bit ASNs as well. And we checked with them, we asked them and they informed us that of the 333 peering sessions that they had, 111 of them were capable of speaking 32?bit ASNs, and the others not. So exactly one?third at the time.

So what we did was we also sent a message out on to the Amsterdam Internet Exchange mailing list, asking†?? inviting all these peers to actually to looping and craze routes to our new servers and report back to us if they had any issues to report. I am very pleased to say that nobody reported any problems at all, and several people were able to successfully reach our networks.

I had a look at route servers around the world and here is a snapshot of the South†African Internet Exchange route server showing reachability of our prefix, the IPv4 prefix. I have highlighted the AS path in bold and you can see that our ASN appears as 23456, so obviously there is a router along the way and it's not able to understand 32?bit ASNs, it sees us as 23456, however it does see us, there is no problem. Here is a similar snapshot from a different route server, open transit, and it's running a fairly more than Juniper OS system and this one understands 32?bit ASNs and it sees our ASN just fine and I have highlighted it again on the slide.

So, we thought a little bit about this and asked ourselves what the success criteria should be for this project. So we came up with two: One was visibility in Netsense, which is the RIPE NCC's monitoring system, it uses data from our RIS project and tells you lots and lots about prefixes. And the other important one was visibility from DNSMON. DNSMON is a distributed DNS monitoring system with over 100 probes around the world and these all send DNS queries and we thought that this would be a good criterion because the experience of DNSMON would be similar to the experience of resolvers around the world. And the question was, yes, will clients be able to reach us, and we wanted to know what it would look like in DNSMON. So here is a snapshot from the Netsense dashboard showing one of our prefixes. As you might be able to see, the visibility is 100 percent across all the continents except for Africa; there is just no data from Africa at the moment, but all the others can see our prefixes just fine.

Then we looked at DNSMON. This is a snapshot of IPv4 reachability, as you can see it's green all over, all the TTM probes can reach this prefix just fine. And this is a similar graph for IPv6, so all the 36 IPv6 capable DNSMON probes are also able to reach our IPv6 prefix without any problems.

So, it just works. I have nothing negative to report, no negative experiences. We haven't had any trouble with any of our peers, so I just want to say to all of you out there, go and use 32?bit ASNs, it just works, we have shown it. Questions?

(Applause)

ROB BLOKZIJL: Are there any questions?

Rudiger: Dutch telecom, Anand I am very happy to see OK, nobody has been hurt so far, and except for a few glitches, essential Leo /KURG early last year, I think that is kind of the overall observation in introducing the ASN, the long AS numbers. Nevertheless, the introduction of full support is obviously very slow, and well, OK, just seeing your /ST?Tistic that only a third of the, fairly sophisticated membership of the AMS?IX actually are up?to?date, is kind of a nice (statistic) nice statistic to show that. And I would point out, even those that are capable in their routers are not necessarily fully equipped to do the thing because all the additional, say, policy configuration, and so on, may be lagging behind like, well, OK, I know a fairly large network that essentially has all the routers at the moment capable, but waiting incompatible software change with the next release expected stays back and says OK we will not officially do the damn thing before we are beyond that. So, that is a comment, well OK, I hope I am not boring the audience too much. Particular question: Did you actually change the registered route to the right origin AS, because I somewhat think you might have dropped off our map if you had done that.

Anand: We have registered this route correctly in the RIPE database. We had a little glitch early oranges I think some people noticed it. That was an administrative error where we had the wrong maintainer, but besides that, this†?? the prefixes are in the RIPE database with the correct import/export lines.

Rudiger: The route object, does have the right AS number?

Anand: Yes.

Alex: From the RIPE NCC, I have got a question from Jabber. This is Jan Prins who ask, bet†?? better being in the Netherlands what is the time frame for this change and when will K?root move to this new set?up?

Anand: Well, this is not for K?root. K?root already has an ASN 25152. This is an entirely new ASN and it's for our other DNS services, so there is going to be no changes to K?root. OK, thank you for listening, everyone.

(Applause)

ROB BLOKZIJL: Thank you, Anand. The next presentation is by Miriam Kuhne, who will do another RIPE labs show.

MIRJAM KUHNE: Good morning. This is going to be very short, I don't want to bore you again, those who have been at the NCC Services Working Group have seen a more detailed presentation. This is just a quick summary and maybe a request for participation.

Just to summarise for those who have not familiar with RIPE labs, it's a website labs dot right.net but it's a community tool and a community platform, and I would really like to stress that, that it's not just the NCC who is contributing content and publishing tools and measurements, prototypes, analysis, any cool ideas you might have, but also you, as members of the community, can contribute to RIPE labs, so first of all, if you have any feedback about what is currently on it or what the RIPE NCC is working on, this is a place for you to discuss that and to provide the feedback for us. But also if you are working on anything interesting, there were a number of interesting tools or ideas presented, for instance, during this week, if there is anything you think this might be of interest to others in the community, you would like to announce it or promote it some more, I would be happy to work with you to put that on RIPE labs and you do have do all this yourself, just contact me and we can work together and write a short description together and I can put it on RIPE labs, you don't have to learn how all the editor works on RIPE labs and things like that. So it's not difficult and it might provide you with, you yes, a good discussion forum with others of the community who you can then works with on these ideas you might have or prototype you are working on.

Just quickly a snapshot of some things that are currently on RIPE labs. All of these have kindly be contributed by the RIPE NCC. You see there on the left, a new IPI that the database people in the RIPE NCC have developed and Paul and Denis, they have presented this in Working Groups during this week. On the right top there is a graph that comes out of an article describing the change or the growth in demographic changes of the RIPE NCC membership over the years, together with allocation statistics, and that has been published a while ago, it might be interesting for you. On the left bottom there is the RIPE NCC†?? the IPv6 RIPEness statistics that have been presented earlier this week, I would chose†?? well, IPv6 readiness or RIPEness of LIRs in all countries in our service region pretty much and that might be quite interest to look at for your country and for your particular LIR. And then on the right bottom is also some numbers related to v6 measurements that Emile of the science group has been doing related to capability of resolvers and clients, IPv6 capability that he has been measuring.

Just to give you some ideas, I just thought that there is so much on RIPE labs, almost something for everybody there.

Just at the end, some next steps that will be coming on RIPE labs. We actually trying to make it easier for you to contribute and to leave comments. We are going to change the website and the layout and make it a bit more modern and easier to provide comments and feedback and maybe do†?? have some rating mechanisms where you can say if you like something or not. We will introduce also like a more†?? a topics cloud, a tag cloud where you can find content easier, also going to cluster or structure topics a bit better so you can†?? so it's easier to find things and to follow a project also, as it evolves. And we will certainly present you with the new website and the new structure at the next RIPE meeting. It will probably be launched early in June.

And yes, that is all I wanted to mention here. We are pleased†?? please take a look, participate. If you have any comments or if you need help or you can't find anything, something that you are looking for, yes, please don't hesitate to contact me or send mail to labs@ripe.net, and at the end, I wanted to, what we learned from George Michaelson that movies are cool, so I wanted to show you this as a movie. That is the IPv6 RIPEness statistics Emile and and some others in the RIPE NCC have been (Emile)†?? you can see how the number of allocations and any other criteria change over time in those countries so it's just a quick movie and you can see it on ripe.labs.net, see all the time how IPv6 readiness gross in the NCC service region which is a nice thing to follow and we will certainly do more of these kind of measurements. Well, thank you. That was all I wanted to say during the closing plenary.

ROB BLOKZIJL: Thank you, Miriam.

(Applause)

Are there questions? Comments? No. Next we will get a report from the RIPE NCC technical support team about the network which we have been using this week. And I think Eric is going to give this presentation.

Eric: Good morning. I work for the RIPE NCC. I am going to do the technical summary of the meeting. A quick introduction: Our team consists of over ten people at this meeting. We had James unfortunately not able to come, he helped us with the set?up, we had Milan Karrenberg who accepted up with the set?up and cabling this time, the son of Daniel, two new team members, and we had around Jim Maarten from the IETF meetings who helped us provide an insight into how they run their meetings.

So, to cover quickly what our team actually does. Basically anything that has wiring, except for beamers and audio, so that includes DHCP and this time v6 as well, all the networking, power blocks, services centre, presentations.

So this is a picture from our room. We got three new switches this time, Juniper EX 2200. This is the room, all the cabling that is red is ours, it's a bit easy but only for a week. (Messy).

So this is is a summary of our network, this time we had two different up?links from two different ASs as well, dial telecom and Sesnet. We have a few VLANS here for administrative purpose, Juniper†?? on this picture you can see the up?link, it's right there.

(Applause)

And it's a cable across the road to the building on the other side because apparently it was not possible to connect it here.

We built a simple triangle this time, two switches in pets rooms one hidden in the registration desk to provide failover in case any fail. We had a number of new set?ups at this meeting, we started doing flash streaming, the old was Windows Media and Quicktime and the flash one is very nice, makes our set?up a lot simpler, works on anything except iPhones but maybe next meeting we will have iPhone stream as well. And also our web team has managed to integrate it nicely with transcripts and the IRC so that it's easier to participate.

Also we have our new presentation system. This is actually the first time I present on a presentation system instead of my own laptop because we can go and see that everything works so we are very happy, it makes everyone's life a lot simpler and easier. Then we made some changes to the wireless in the plenary this time. Our standard set?up for this room: 6 access points, you can see the three in the back. Based on the input that Jim Martin gave us, we now put 4 extra ones, so what used to be a normal conference seat is now a wireless bay station mount. So, we had a very small†?? some small issues we run into. The webcast did not have IPv6 during the first opening plenary. The issue with that was that we†?? the system preferences of the Mac route server was sure that it was configured but on the interface it didn't appear. It took a while to figure out the host itself had a firewall, that problem was not in our router, that took about one?and?a?half hours and since then we have had IPv6. We suffered some hardware damage, this is one of our servers that we ship, the other side is even worse, a piece ripped off. We also lost two Foundry power supplies, the connector is not fixed in any more. We were lucky because all the hard disks survived so that worked very well and we had spare switches so no operational effects.

So some of our statistics: We installed as usual a TTM box in the hotel so this is the GPS antenna on the roof and used it for one way delay measurements in a mesh of 90 other boxes and we find, again, a box which actually has more stable and lower latency IPv6 than IPv4, the upper graph is IPv4.

(Applause)

This is actually a 25 percent difference. We also see the same of the box of ISC in the US, that is slightly less difference but still IPv6 at lower latency.

This is the stats that Emile has been making, of measurements, how many people come to ripe.net have native v6 and the right part of this graph covers when the RIPE meeting starts, so we actually almost triple the native v6 traffic to ripe.net. But even if we exclude the entire RIPE meeting network we still see a bump of native v6 connectivity so apparently more people with v6 come to our website during the RIPE meeting.

We have also kept stats of the webcast users. This graph starts on the left on Monday afternoon and the green line is v6, blue is the total so we can see there is between one?third and half of the clients using IPv6 when watching our webcast but also, that more people with IPv6 leave the webcast running at night. This also does not include the audio cast which is from the side room on Thursday so maybe that is why Thursday is is a bit lower. As I said this time we had two up?links, our peak was at 40 megabits in 27 megabits out, the average was ten in and almost three out. The average inbound is actually double that of our previous RIPE meeting which may be because of the improved wireless; our outbound has halved compared to the previous RIPE meeting, probably the outbound is lower because we no longer do streaming from here, it now goes all to our office and from there, streamed.

For IPv6, the numbers are as usual, lower. A big peak of 6 in one out and average under 1 megabit, this means 6.2 percent of the traffic is IPv6 and this has decreased from Lisbon, in Lisbon 7 percent so it's slightly lower than last time. For the wireless we deployed 18 bay stations, we used to do 11 and our peak was at 340 associations, that is 30 percent higher than the last RIPE meeting. You can also see every colour has its own axis point, before we used to have up to 100 clients per axis points, now our maximum is 40 with the average being about 20.

So that was it. Do you have any questions?

SANDER STEFFANN: Can it be that the IPv6 traffic is lower because you are now not streaming from here, because it looks like a big percentage came from the streams?

Eric: That is a possibility yes.

SANDER STEFFANN: So it might actually be better than last time?

Eric: Yes.

AUDIENCE SPEAKER: I would like to thank you for enabling the DNS validation on the resolvers here.

ROB BLOKZIJL: OK. Thank you, and your team.

(Applause)

This was the end of the presentations which we had for this morning. The next, what we do is some items that came out of some of the Working Group meetings and have to be reported here and first I call on Roland Perry, Cooperation Working Group, one of the items discussed there had to do with developments within the ITU, International Telecommunications Union, and some of their wild ideas on how to get involved with IPv6. Roland will explain and will ask you a question which you may answer.

Roland Perry: Thank you, Rob. During, I suppose, the last five to ten years, there has been some disquiet expressed inside the ITU that there is some inadequacy of our current Internet address management system. This disquiet is expressed by resolutions put before ITU means of various kinds. There are several streams of these meetings and currently, we are concentrating on a new that the ITU have set up called the IPv6 group. We attended that meeting with all the other colleagues in March and, as a result of that, they have set up some mailing lists which we would like to be able to contribute to in order to better identify what the concerns are within the ITU because they are their current thought process is to say, well actually, lots of people have been saying there is something inadequate about the current address system but nobody actually puts their hand up and says exactly what it is that they think is inadequate. And therefore, the RIR community does not have a†?? anything to answer. There is no answer we can give back to say well actually, that concern is not an important concern as you think, we think we can embrace that concern within our existing system. So we had a discussion about this in the Cooperation Working Group and the result of that is a community statement which you can all see on the screen, I won't read this out to you and I would like to ask you for your approval for this community statement.

ROB BLOKZIJL: Yes, the statement you see on the screens here basically says two things: A) we don't think there is a problem; and B) if you think there is a problem, speak to the RIPE NCC. They represent us in this. So, the question is, are there any objections that we work along that line using this statement? Let's phrase this more positive: Do we all agree that, as far as we are concerned, that we are happy that the RIPE NCC goes to all these boring meetings?

(Applause)

I think we have consensus here, and though it is not mentioned in the statement, I think it goes without saying that the RIPE NCC will report back whatever new developments take place.

Roland Perry: We will, yes.

(Applause)

ROB BLOKZIJL: We don't go through all the Working Groups, but there are a few items coming out that†?? that came out of Working Group sessions this week that we felt needed at least reporting to the plenary session. And one has to do with the IPv6 Working Group and Marco will say a few words there.

MARCO: Yes, I just asked for two minutes and this is just to let, you know, we actually rechartered the IPv6 Working Group, it's up there, basically everything that has to do with IPv6 deployment or the interaction between IPv4 and IPv6 is in scope. We also got two new co?chairs and one of them, David Cassens is still around although he left the building already, and Shane, and Shane had prior engagements before he was elected to co?chair. We got a brand new contact which is IPv6 ? WG ? chair@ripe.net, so any concerns about IPv6 Working Group and there are still famous mailing list. We have quite a good session, some bring a storming to do with the new charter and saw proposals pop up so this was a successful meeting as far as I think.

ROB BLOKZIJL: So new charter and new leadership.

(Applause)

Thank you, Marco. A similar situation exists in what used to be the test traffic measurement and I think we will have a similar report as the previous one.

Ian: We don't have an updated website yet so you will have to put up with looking at me. We completed the rechartering effort we began in RIPE 59 in Lisbon. We have also now changed the name from the test traffic measurement group to the measurement analysis and tools group. This reflects the broader focus of the†?? the broader approach of the new group with a move away from the focus solely on the RIPE NCC test traffic measurement network.

As well as the change of the charter and the change of name we have also had a change of chairs; Henk of the RIPE NCC has stood down after more than 12 years and we thank him for his efforts over the years.

(Applause)

That meant that we needed to get somebody else to take over the chair role alongside me and we got two new chairs, we had an election, it was more sort of election, it was more conclusive than the British general election, quicker than the NCC GM election as well and I would like to welcome Richard Barns and Christian Kaufmann as the new co?chairs.

(Applause)

And that was it. See you all in Rome. Thank you.

ROB BLOKZIJL: Thank you. This, as far as I know, concludes the short reporting from Working Group developments. Did I forget anything? No. Right. I am looking for Olaf. If you are a newcomer I would recommend to fasten your seatbelt. And without any introduction, I will just sit here.

OLAF: For the newcomers over here, again I was not only kindly asked by Rob to appear here but there were some other people as well that asked me to come and I want to start this with some poetry which is not projected to the screen. If arguments†?? you have to picture me with a big latex robe, I don't know if you want to go there but just close your eyes.

If arguments remain after option processing and neither the minus CE or the minus S option has been supplied, the first argument is assumed to be the name of a file containing shell commands.

So this short and futuristic poem is an introduction to the secret Working Group

(Applause)

Most of you that are not newcomers also know that pluggy at and stealing is something we usually do without any attribution so I am not going to say this was a text that came straight from the batch man file. Now, I did expect something because I†?? I had†?? order of the slides arranged a little bit differentry. Normally, when you are on strange presenting, something happens, you know, and it has been happening more often and more often during the last RIPE meetings. So I had expected that I had†?? would be interested by at least five people with gigantic cameras flashing it in my face and I had prepared for that, except it didn't happen this time. I think that is†?? people are a little bit afraid of the secret Working Group and don't do that. But, we prepared for them anyway.



This was actually opening slide because I expected to be interrupted but it didn't happen.

The next slide, the next poem is called digital nostalgia, we sometimes long back to old times.

The RIPE NCC board election, broke new ground with a web?based selection. But the paper votes too for the few without clue, were retained out of misplaced affection.

(Applause)

Nigel Titley: RIPE meetings are tightly directed, precision is clearly expected. The direction is clear, there is nothing to fear, it's only your mind that is affected.

(Applause) Olaf: For the next presentation, I need to connect something and while there is plenty of†?? can somebody bring me some cat 5. Our next model is called Asperger geek, he wears jeans, bare feet in test labs and a T?shirt that not only outlines his empty†?? repellant properties. The jeans carry a great many utilities such as the obligatory cell phones, yes multi task, at least 13 UDP 45 plugs and they cheat card. The instead of the human relationships with accessory which consists of the tight knot which symbolises the attention to detail unconnected from the rest of the physical world.

(Applause)

Now, Daniel ask if we were still friends. I think this is a little bit challenging.

Anyway, for those that still have their eyes closed and think of me in a latex suit, you should open them now, because there are more memories of yesterday. It's called shadow porn.

You can close your eyes again now. While a bass owe voice came through the door. And the ladies wore less through some more, the table all lifted, and gazes all shifted, as cleaners with mops mopped the floor.

(Applause) obviously, this was a show sponsored by a shop just around the corner...

You must notice, I have noticed this, you know, leads they are everywhere, we are fond, which colour do we like best, yes, blue.

Nigel Titley: RIPE is forming it's very own church and it's taking a little research, we have got new LE Ds and our candles and these, will illuminate Rob on his perch.

(Applause)

The secret Working Group has observed a trend, a trend, and the trend consists out of this: Labs. They are everywhere. NL net Labs, RIPE labs, cz.nic labs, forget them, we have the secret Working Group labs.

Now, the secret Working Group labs doesn't have a charter but what we do have is a chemical lab and what is better, a time machine. So what do we do? We subjected this little IPv4 cheat card to our time machine, and what you are about to see are pictures that are taken over the next couple of months in our time machine. And you will see what happens to this magic card, we believe that the RIPE NCC has printed these cards with some sort of time vanishing ink.

(Applause)

We have stared a long time at the other side of the card and nothing happened there.

Nigel Titley: The remotes on the stage were such fun, as a strange little ritual was done, the first one was dead, number two jumped ahead, and the third was a phaser on stun.

(Applause)

Olaf: So we are now at the final poem of the day, the final piece of the report, and we are looking forward remembering ourselves that bottom up policy is a very valuable asset for this industry.

Now the next meeting will be in Rome and we have to keep this bottom up policy thing in the back of our head while there.

Nigel Titleyly: When in Rome do as Romans all do. Policy has its followers who, are a strange sort of bunch but there's never free lunch as the losers may find to be true.

Olaf: I know you are puzzled but...

(Applause)

ROB BLOKZIJL: We don't do questions after this report. Though, you might be inspired to contribute at the next RIPE meeting.

Right. The thing left before we close this meeting, is a couple of, for me personally, very nice items. We run competitions as you know. In order to stimulate you to do early registrations, we have the winners of the registration competition. The first three people who registered for this meeting and are still in this room, will get a little prize. So, if your badge in the upper right corner says RIPE 60 ? 2, then you are Andreas, and you are the first of our lucky winners. If you are new here, you might wonder why do we start with 2, it's very simple: Mine says one, not because I am such a quick action boy, but the 1 comes out because it's hard coded in the registration software. If your badge says number 3 then you are Jeffrey and you are our second lucky winner today. And if your badge says 4, then you are Sergia.

(Applause)

And you are the third lucky winner of this competition.

We had another†?? not a competition; a lottery, we had the RIPE NCC services centre operating this week and everybody who dropped in with a question or a problem got a lottery ticket, and Andreas seem ma is not here and he is going†?? nice to see that NCC services team is more than one person. And the winner is number 28.

(Applause)

And I hope your problem had been solved. Our good friends from cz.nic also had a competition running, on on?line game competition and who from†?? oh, yes†??

SPEAKER: Thank you, thank you very much, so the nibbles competition, you made a good job and work hard, all the meetings, through the lunches, we are very lucky that we always closed the competition at 6:00 so could you go for some social so and the winner is Tobias.

ROB BLOKZIJL: Thank you. This is the end of the competition part of this RIPE meeting. And what is left now is something that gives me always extreme pleasure in doing and that is thanking all the people and organisations that have made this RIPE meeting a very pleasant and successful one. I think it goes without saying that we should start with the most important one organisation for this RIPE meeting and that is cz.nic. They have been our local host and they have been giving us tremendous support in organising this meeting and organising support in organising the socials and we would like to personally thank some of our friends from cz.nic and I understood that Camilla will handle this part of the programme.

Camilla: Thank you. I want to thank cz.nic for their support and some people especially and that is Ondrej Filip, Tomas and the girls, Suzanne and others. So we have some tokens of our appreciation.

(Applause)

SPEAKER: So, thank you very much that you enjoyed the meeting and I have for you very nice gift also, because right now start good sun weather.

(Applause)

ROB BLOKZIJL: But we all enjoyed, apart from very nice meeting, of course, also very nice socials and there are a lot of people involved in organising these socials and there are a lot of organisations that have been supporting us and I have a slide here which says thanks to our sponsors, we have an ever?increasing list of sponsors, which is healthy, I think, and I would like to have a round of applause for interxion, NetNod, cz.nic, Brocade Google, Dial Telecom, Cesnet, thank you all very much. Thank you very much for your support in making this meeting so successful.

We had elections, the RIPE NCC association had elections for two board seats last Wednesday evening, three excellent candidates and two seats, so there were two winners, there were no losers, you don't lose that, especially if you are Janish Sacho who has been serving on the board for so many years and I think the chairman would like to say a few words.

Nigel Titley: Thanks, Rob. Janos and I were elected to the board I think for the first time at the same time. Janos has the inestimatable advantage that he is much more handsome than I am, that his financial acumen is far better than mine and his English is much better than me, much to my embarrassment. I will miss him enormously on the board, as will the other board members, and I hope he will continue to advise us, at least as we continue forward, and I think we have something to present as a token of appreciation. Janos?

(Applause)

I hope you can fit this in your luggage and thank you very much.

SPEAKER: Thank you very much, everybody. Thank you very much everybody for your kind words and I was†?? it was a very big pleasure for me to be in this community and serve this community and I hope I will be able to continue to attend these meetings and see you all. Thank you very much.

(Applause)

ROB BLOKZIJL: Last but absolutely not least on my list of people I would also personally like to thank very much is Joao, for those who don't know it is the person who carries the plenary programme. There are a lot of people involved there, but Joao takes the lead and Joao takes the responsibility of for many, many meetings already to put an excellent programme together and, from time to time, I think it is nice to remind people that these programmes do not drop out of the air but there is a lot of hard work going on in between meetings to get a good programme together. So you can do it now†?? you have to do two things; you have to think about whether you can commit yourself already, well, next week at the latest, to start working on nice presentations, and secondly, join me in a round of applause for thanking Joao for all his good work but he is not leaving.

(Applause)

It has been said several times already that this was a good meeting, it was a very special one because we broke a record: We had 427 people present at this meeting, not counting RIPE NCC support staff and we never had so many people in a RIPE meeting. So, I would like to thank all of you for coming. We also had an impressive amount of newcomers here, about one quarter of you are newcomers and now I sympathise with Olaf which one of these two am I supposed to fire off.

AUDIENCE SPEAKER: The third one.

ROB BLOKZIJL: The taser gun. A. This is a pie chart which, if you are a newcomer you will recognise because this is one of the first pie charts you saw on Monday morning, where are you all coming from, organisational?wise, no big surprises here: About half is what we call the industry, ISPs, but it is always very healthy to observe that other sectors of society are also represented among you, governments, educational institutions, other organisations, other RIRs and other others. This is taken from the registration information. I sometimes say that the other others, which is nearly 20 percent, you did not tick which type of organisation you are working for, so either you are not working, I don't think so, or you are working very hard but have no clue for whom you are working. Right. You are coming from countries, this of course is also very interesting; it is not surprising that at the RIPE meeting in a particular country we attract a lot of attendees from that country. It is surprising that one of the largest national groups at a RIPE meeting is the United States of America, but as you can see, you come from many, many, many different countries, which is a good thing.

So, without further ado, I would like to thank our hosts once more and I close this meeting, I wish you all good trips home, if you are staying on in Prague, it's a very beautiful city, and as our friends from cz.nic said, they switched on the good weather mode. Have a nice time in Prague and see you all at the next RIPE meeting in Rome. Thank you for coming.

(Applause)

LIVE CAPTIONING BY AOIFE DOWNES RPR

DOYLE COURT REPORTERS LTD, DUBLIN IRELAND.

WWW.DCR.IE