These are unedited transcripts and may contain errors.

---

The plenary session commenced on the 6th of October, 2009, at 11 a.m. as follows:

MIRJAM KUHNE: Hello everybody. I hope most of you are back from the break. Can you hear me? OK, good. My name is Mirjam Kuhne, I just started working at the RIPE NCC in September. , however, some of you might know it's not really fair to say that I am new to this, I have been standing here before in a different role or similar role, if you will, a few years ago. Well the RIPE NCC has developed a new project or programme, it's called RIPE labs, it's an initiative developed by the RIPE NCC but is meant to be a community tool, a tool from the RIPE community for the RIPE community, so I wanted to highlight a few of the goals and benefits and ideas behind RIPE labs, and then after I have finished, Robert Kisteleki from the RIPE NCC will show you some examples and after that Franz Schwarzinger will show you a tool and at the end of the session we will hopefully have plenty of time for discussion and questions.

Well, let's go back in history a bit, because first actually was RIPE, there was RIPE community, as, you know, some people just came together and coordinated the activities and they met and organised themselves in early IP days in Europe and also I mean, RIPE as a community like the RIPE ?? the participants in RIPE organised the activities and someone said oh why don't I go off and right a contact database so we know who everybody is and somebody else offered to host mailing lists and we had somebody else offering to host the meetings, and so it really was a community effort and only after a while after the RIPE community was growing they decided to it would be quite handy to have like a service centre, or a coordination centre that could help with all these services, and that is how the RIPE NCC was born.

And then RIPE NCC took on a lot of these activities and they maintained what is now known as the RIPE database and organised these meetings and many other services that are still around today and still are used by many of you today, and only after a while, after the RIPE NCC was known, was a trusted service centre and a neutral place for ISPs and operators and others to coordinate the activity, somebody thought why don't we ask the RIPE NCC to also hand out IP addresses in Europe (IPs) because at that time that was still done centrally from the US and only then took on the task of RIR, which is now probably the most known and prominent activity of the RIPE NCC but it's not how it started really and what it was initially meant to do, or at least it wasn't the only activity, so in a way RIPE labs is an attempt to go back to the roots, a right of RIPE and RIPE NCC.

So what is RIPE labs? At the moment, it's pretty much just a website and I will show you some of it later and but it's hopefully becoming a platform for discussion and for pleasanting new ideas and tools and research results and giving feedback to each other on these ?? on the content that is on RIPE labs. So what can you do on the site? Well you can test and look at tools, test them, see how they work, if you find them useful. You can also contribute your own research results, your own prototypes, software. You can participate in discussion forums, you can also maybe create your own blog, look back to all these details later. And basically, you can influence what is happening, what is happening in the NCC and also in the community, and participate from an early stage on the developments. At this point I would like to mention Robert and who was the one who came up with this idea initially a few months ago and has put a lot of effort into getting this started and also then has put a lot of effort into getting this off the ground and helping to set up the initial slide and content and has been very supportive of this project.

So why RIPE labs? Why do we think that is necessary? As I said, it really is a mechanism for you to, well for us a feedback loop, to see, first, be the first to see what is going on at the RIPE NCC and in the community so the RIPE NCC will use this platform to present and write new ideas we have, new research or, you know, ideas and tools that come out of the science group and we can publish maybe potential new services or new features of existing services and the main point here is that so far, I think the RIPE NCC is often maybe tried to assume what the community wants and so we went back and developed something and then we thought it was a great new tool and we published it and only then we ask, is that really useful to you and to the membership and to the RIPE community? With RIPE labs we are trying to come back to the community earlier on and earlier ?? well, when we have like just an idea or prototype or better versions will be published there and then you can give us feedback and say if you find this useful or not and based on your feedback we can ?? either develop it further or not or make it more blue or green, depending on what you ?? on what your comments and your feedback.

But on the other hand, also you can use this as I said to publish your own ideas and get feedback from the community, maybe you have some cool research results, you know, you don't know how to publish them further or share with your colleagues in the community or maybe you have a cool tool that you developed and want to put on the RIPE labs site and share with others and maybe somebody else wants to pick it up and develop it further. It's open to any, well any ideas that you might have.

So, that is pretty much summarises what I have just explained. It's a feedback, it's a faster feedback loop for you to connect with the RIPE NCC and also with each other and, should you care about this, you know, maybe because it gives you a mechanism or a way to influence directions of some of developments first from an early start and it's really meant to be a community tool and hence it's also called RIPE labs and not the RIPE NCC labs, for instance, RIPE NCC will use it but also you, hopefully, will use it.

And there is the URL, you can look at it now. I don't know how it will act, actually. We tried to foresee like what happens, many people will use it, but we don't know what happens if the whole RIPE meeting will jump on it at the same time. Just in case,

DANIEL KARRENBERG: It's doing fine.

MIRJAM KUHNE: Thank you. You will see there, it doesn't show everything now, it's pretty much just an introduction and you can see there are some cat goreees, news and tools and also category with forums and blogs. Just play around on the site and see how you like it and if you have any comments, let us know.

Before you start, you know, playing on the site, I just wanted to highlight some RIPE labs rules. First and foremost, it really is open to anybody. All the content can be read by anybody, you don't have to register or to read, anything can be read. Obviously, we hope you have some civilised discussion following etiquette rules. If you want to contribute to the site and want to leave comments or post some comments or leave your own ideas, I would like to ask to you register before you do that, it will ask you for user name and we also ask you for a motivation, like why do you want to participate in RIPE labs and ask you if you have LAR get an idea who the participants are and help us late tore classify the feedback we get, but we treat all this information confidentially and hopefully motivate what we do with it and why we ask you these questions. We will have to see as we go along but this will at some point also help us to identify what the members want, what the RIPE community want and evaluate that feedback. And the other important rule is we really cannot give, well a lot of service guarantees to what is on the site. Until now, the RIPE NCC was always very ?? trying very hard to provide service, full services to the community and guarantee that these services are up and running and look pretty all the time and everything. Now, this will be different here because a lot of the tools and a lot of the software and also some statistics and articles will be just ideas or prototypes and they will not be fully fleshed out and ready. And also, some of it might be submitted by a third party so the RIPE NCC is into the control of that. We will hopefully be able to guarantee that the site is up and running and be there for you to participate, but not necessarily everything that is on the site.

However, based on your feedback or based on how things develop and get picked up on the site, some of the things that will be posted there first might become full RIPE NCC services later, or, you know, we will take them off if there is no interest or some other one might pick up an idea that you might find interesting for your own operations. So it really is very much based on community feedback.

Well finally, just a list to illustrate what is currently on RIPE labs, some you probably already see it in front of you. There is quite a lot of stuff on there already. Most of it developed inside the RIPE NCC for now, just to give you some teasers and some ideas what is possible. There is something called resource explainer that you will hear more about that later. Introduction to the Internet number resource database, you can see what happened to your IP resources in the past and Robert will show you more about that later. There is a calendaring tool that Hank put up, you can plug in your RIPE meeting appointments and it's explained how to use that. And there is some cool eye cannedies and movies, an article on 16?bit ASN number and finally there is something called netsense which is our next generation information services and Franz will give you an idea about what that is and what it entails. So if you have any questions now, we can maybe have some until Robert is ready to come up. Otherwise we will have time later. I also want to illustrate that there is a demo stand outside later that I will be there after this presentation and we also get some cool T?shirts like this, and if you have any questions or please come by, we will be there and can play on the site and explain you more how it works. Are there any questions now? Otherwise, we will keep them until the end if Robert is ready. You are ready? Sure.

DANIEL KARRENBERG: 50 people on and still doing fine.

ROBERT KISTELEKI: So, what I am going to do is, what I am not going to do, I am not going to go through all the bits and pieces of content that we put up there but I think I am going to show you a couple of nice and interesting things about what is already on the site and then you can go on and explore on your own, and see if you like it. So as a start: You may have heard about INRDB, one of our projects that is the name is for Internet number resource database. This is essentially a huge database of Internet number resource related data that the RIPE NCC collected over the years and we extended it with other databases as well from CAIDA and GOIP services and so on and the reason being is that if we want to do historical researches such as the YouTube that some you might remember, and similars, we need a good data store that helps us actually dig into information and tap into that information. So the RIPE NCC has RIS and TTM and all those beautiful services, we also have the RIPE database about registrations and all that so if we could hold that information in the same place, that enables us to do beautiful things with it.

So currently, this database contains all of the RIS table dumps over the last ten years, so if you imagine that is immerse amount of data and we are able to query it in realtime speeds so that is good. We have IANA assignments just because it helps news some research, RIR stats files, and some other external data sets.

So actually, using this enables us to come up with new and cool services, which ?? some of which we want to show to you just now.

INRDB is accessible through labs already, this is an application I am not going to show, there are example queries you can dig into the RIR stats files and I think maybe to RIS itself over the past ten years of history. So you can go on and explore it on your own.

Instead, I am going to start slowly and show you something that is relatively simple to get out of INRDB but it's really, let's just say it's an eye?catcher, I think. What I am going to show you is resource allocations as they have been going on over the last couple of years, but I am going to show you this in a fashion that I don't think you have ever seen before. This is going to be a movie about race and the racing cars are actually RIRs here, so what you are going to see, let me zoom in, what you are going to see is RIPE NCC ARIN APNIC LACNIC and AfriNIC. On this side you will see the actual laps that they did. The movie starts end of 2003, when the RIR started to publish the same format of stats files so that was easy for to us grab. This is how many laps they have done until this time. This is where we join in. Each lap is one /8, so if you see the cars going faster or slower speed of allocations. Tried to express it in /16s but it's an arbitrary number because as you will see the numbers flying by that does make sense, we didn't want to do/32. So in any case, at this stage, AfriNIC has not even started. We are in 2003. And ARIN is way ahead of everyone else. So I guess ?? I was just ?? I will just start the movie and maybe I will just highlight some interesting points here. Now we need some music. Maybe ?? don't you just love it. (Applause)

This is supposed to be a stable browser. Still the music is on so that is good. OK. This is why it's on laps.

So now we are on. Race has started, AfriNIC is staying there, waiting for the others. They are going to start in early 2005. So, oh, what is that? ARIN is sometimes going backwards, can you imagine that? They are giving back address space to everybody else. But everyone is on. Go AfriNIC, go AfriNIC. They are going to start soon but, mind you, we are going to transfer them a whole lot of resources at the start. Almost immediately did one full lap. That is good. Everyone else is just going around and around, that is good. Crazy. I will give you the explanations later on. So as the time goes by, you can see all these cars are just going back and forward. It's crazy. OK we are still in 2006. Mind you, one line here is the/10, just to give you a form of reference. AfriNIC is trying to make the first lap, they are almost there, they are almost there. You can see in the meantime that in the laps, APNIC is catching up. They are really, really coming up. So if we can just go, go, AfriNIC, they will do the first lap soon. Maybe I would have ?? yeah. Need some push from the other RIRs maybe. 2008, the movie runs until mid?September 2009. And I can assure you that the final will be really amazing. APNIC is really catching on. At the start, the difference was like 20 /8 between NCC and APNIC and now it's only three. So that is a fact. And now look at the NCC, look at the NCC, yes. It's kind of in a limbo, going back and forth. So that is what we had.

(Applause)

And thank you. The reason why cars go backwards after now and then, you can watch the movie yourself as many time as you wanton labs, so the reason why the cars every now and then go back is because we are just looking at the stats files as they are published daily and if just Geoff was standing up here he would have said we are using the linear first order derivative, trying to measure the speed of allocations and if an RIR gives back address space to IANA or to another RIR that means that that space is gone from their stats files, at some point appears in another stats files so ARIN in this case, went back a number of times because of ERX, right. Some ERX space was taken out from their stats files and then put into someone else's. Also the RIPE NCC sped up at the very end, because of 51 /8 I believe, it appeared in 2008/2009ish. There are reasons for this. I am not claiming the movie is perfectly accurate but pretty close to reality.

Moving on. I want to have another movie, actually. Let's see which one. This will not have sound. NCC and routers.

So what you are going to see here is the address space as managed by the RIPE NCC. The blue bars ?? so each bar is is a /8, on the left?hand side you can see which it is, on the right?hand side you can see how full it is already. This movie also starts in 2003 and you can see that the green bars are the ones that IANA allocated or assigned to the NCC already so as time goes by we will see more and more green bars popping up. The blue ones are the allocations or assignments that the NCC did and the red ones are the routing entries, routing table entries so coming back to INRDB, we run through the whole ten years of history of routing and tried to apply that on to such a movie to each frame of this so the size of the chunks that you can see and the position of them actually represent the position and size of that prefix.

Let's just ?? this will not have sound. So we start in 2003 and you can immediately see that address space is just flying off. We made, for example, here, an allocation ?? OK, let me pause here for a second. What you see is roughly in 2 ?? May 2004 we received four new /8s and immediately there are chunks of address space out of it advertising the routing system. That is, of course, the deboneising effort, this is a pattern you will see. As soon as a new green bar pops up the debogonising effort kicks in and we start to advertise some part of it and as soon as we start allocating from that /8 or a couple of days before that it goes away. So you can see that pattern. And you can also see our reserve was like down to less than two /8s so we got a new allocation from IANA which is immediately tagged again so that is all nice and good. And this is going on, this is going on at a scary rate actually because if you see the days are going by and flying by. So in the interest of ?? actually, this is interesting, a whole/9 was gone in early 2006. This is, I believe, France Telecom and you can also see they are not advertising much of it, not yet anyway, but as time goes by it will grow and grow. This movie does not contain two /8s that we just recently had, two /8 and 46 /8 so those are missing. Let me just fastforward here a minute. 2008, 2009. So you can see that the free space that we have, which is again the green stuff is going away rapidly. The movie stops here, mid?September and you can see that we have roughly two, three, you can get the exact number if you want. This is where we receive the new /8s from IANA. That was enough about movies.

There are a couple of other movies on the website as well, so please feel free to explore. Coming back to our original schedule.

The other bigger project that I really want to show you and we are getting a bit more serious here is the so?called resource explainer. You can read that the article on labs but I am going to give you a different presentation about it; I am going to show you what it's capable of so you can explore on your own at some point.

Now, could you go ahead and a start hammering our application service now, but if you could wait until half an hour until I finish, it would be really nice. So what is it? The resource explainer is is a new service that we are thinking about and we think that it's going to be really useful for all the community, so therefore, we started to explore this and therefore, it is a very good candidate for RIPE labs because we are not yet sure where is this going to go but we are sure the idea itself is good. So the idea itself is based on community needs. We saw a lot of e?mails flying by mostly NANOG saying, look at this resource, and wouldn't it be nice if I knew everything about it before I received it from ARIN or maybe they knew everything about it before they gave it to me because it's just dirty, we have been thinking about a service that actually explains these resources and gives you the whole history of it: Where has it been used, was it allocated, how was it routed, is it in some black list, Spam list and so on. So this idea came from our colleague and we just explored it and we think that we are pretty OK with it. There are a number of reasons why you would want to do this. One of the expectations is that transfers are going to happen at some point ?? maybe/maybe not, I am not going to go into that debate. But if I was a LIR who is willing to buy or rent or get some resources from another, I would like to know if that resource is actually clean enough for me because if it's in all of the black lists and Spam lists, then I am not going to be happy LIR. So that is one of the reasons.

You may just want to dig into history and see how it specifically was used over time.

Again the idea we want to make one service that gives you everything you need to know about those resources. It's a one?stop shop. Let's try to do that again live. I have a couple of examples that I collected for myself and maybe I just explain this. First, a guided tour about the RIPE NCC's own/21. I chose this because we know this, we can be kind of familiar with it, so it gives us a pretty good introduction of what we are going to see. What you see here is a number of tabs which show you the different aspects of the resource that we can tell you, you can see for yourself who is responsible LIR, who we think the current holder ?? or sorry, I shouldn't use the word "current" ?? who we think the holder is, how it was routed, what do we know about DNS, reverse DNS, was it was active on the network, is it in any kind of black lists or Spam lists and where is it located in terms of location. For all of these tabs, we have two sections, the most recent information and the historical information. So you can see that the greenish part of it is actually the current state as it is now or however you define "now," you can specify your own time frame if you want, and the blue section at the bottom tells you the history of it.

If we start up with the RIPE NCC prefix, what we can see is as of two days ago, which is pretty recent, as of two days ago an exact match was found in the routing system originated by AS 333. So the RIPE NCC, that is good, no more specifics, no less, that is OK. For all the pages you can read the methodology, so if you are unsure what you are seeing, you can always click here, expand it and read it out to check, well, what is this, really. But you don't have to, you can just go ahead and jump into the details.

So during the history, which is pretty stable, I can say, what we see is AS 3333 has been announcing this forever. Mind you, the timescale is ten years here and you could actually see and you will see a number of times that the actual loading time. Page was like a second or two maybe. Interestingly enough, we started with a /22 and well, whenever we have the data for and around this time we actually switched over to our own/21. Now, there is this interesting thing here. 193 /8 announced by AS 3352, so who are they? Currently, we don't have the AS type reports and IPv6 type of reports so, that is common, but of course, we can find out who these guys are. So AS 3352 is Telefonica Spain. Are they here? No, they are not.

DANIEL KARRENBERG: They don't come to Portugal.

ROBERT KISTELEKI: Interestingly enough, at some point they announced the whole /8. If that makes them happy, fine. Let's explore the other tabs. Who is responsible RIR for NCC's own prefix. That is a good question. So current state, responsible RIR as seen in the IANA, so the IANA says it's ours, that is good to know. Responsible RIRs as seen from the RIPE NCC's own measurement, the measurements that I showed you the previous meeting in Amsterdam, we are abusing that to actually show to ourselves and to the community where the resources are, according to these measurements. What you see here is the history according to IANA and to us. The gaps here are just artifacts because we just don't have the data unfortunately.

I must say thanks to Geoff who was giving us a hand for collecting historical IANA data and Leo as well but we still miss pieces of IANA assignment pages, so if anyone has them, let me know. Besides that, assume there is data here, it's pretty constantly with RIPE NCC and our own measurements also say the same. Ours start in mid?2004 so we don't have anything before that so that is why the gap is there.

Moving on. Who is the resource holder for this resource? Well, let's look at that. This, as it is today, data coming out of the RIPE database, it's slightly interpreted form so we tried to make some sense for the data because we don't have to be able to recognise and interpret who is data pieces; we intend to expand on this in graphical form so what you see here is really a prototype, we will work on it more. More interestingly, how it was historically? So what we show here is the maintainers of the different objects, or different size prefixes in the RIPE database and you can pretty much see that it has been stable during the last time, so there have been times when we had different maintainers, so that is all good. I am giving you the tour. DNS and reverse DNS comes from CAIDA originally. What it really is, is CAIDA is doing constant measurements, what they do is they fire trace routes all over the place on basically semi?run ? and record each hop along the way and translate to DNS names so they have the IP address, they look it up, so they have names for it. And we put all that information behind the thing, what you see here is a pretty good overview of what kind of hosts have been seen throughout history, so this part is exactly that. So how much of ?? how many hosts and was covered over time, how much .nets and .orgs have been in this space. Interestingly enough this is RIPE NCC's own space. We have in expendable form the history of these objects as they were in the RIPE database. So for example, the /8, 93 /8, covering stuff for RIPE NCC's own stuff, there was ?? there was a change in 2008. Maybe it's more interesting.

So originally, the subject was created by someone who is called DK58 and it has changed to AP 110. Who they are is left as an exercise for the reader. You may have good guesses about that.

Moving on, network activity. This is coming from the same CAIDA data set. This shows if they have seen active hosts in those particular /24s so now we are looking at/21, which is from 0 until 7.255 and in these/4s they have seen at least one host answering to the traceroute, so it was measured. The .2 is missing from here and that is ?? I think that is because it's behind the firewall so that is just fine.

You can see that ?? well, this is whole history, I will zoom in at some point, you can see how it changed over time. So how much of this space was seen active again in the CAIDA measurements? We intend to expand with other active measurement sets but for now we have CAIDA which is pretty OK.

Black lists. So current is own space has a host in it which is black listed in the UCE protect database, something for us to look at it, but it's already information for us because I don't know if we would have spotted it anyway. He can also look at the history and check that, it was added in the last day so it's not that bad.

Finally, geolocation, it has the same structure, what is true now, so the most recent stuff is 100 percent of this prefix is in Holland, that is good to know, and actually it is in Amsterdam. No magic here, Google maps, everyone knows that. And how it changed over time. This is also interesting. The goip information comes from MaxMind, they have available data set, so we are putting that on and we intend to get more historical data for that. It hasn't changed at all, has always been in Amsterdam. That's good. This is the basics. We have a printable version, so if you want to look at all the pages you want to print because you want to hand them over to your boss, you can do that. OK, that was easy.

The RIPE NCC's bigger ?? this is the whole NCC block which also covers K?root and RIPE meetings. So what we start with is routing history of it, and again, this is ten years of routing history. Let me show this to you again. It was three seconds, something like that. Ten years of routing history. What we can see here is is a lot. Just to highlight some of these. We can see that this, I believe this is brown stuff 8 /22 which is the RIPE NCC meeting prefix ?? well, somewhere until 2003ish, it has been announced by different AS numbers, which is perfectly fine because we have been to different places. So who is AS 766? Again, this will be an integral part of the report but for now I can just look it up, and that is Spain, interestingly enough at that time the RIPE meeting was in Barcelona, surprise, surprise. Coming back we can see it has changed ASs but after we actually just changed our behaviour and we have our own AS number, which announces the RIPE NCC meeting prefix, you can also see that the actual RIPE NCC meeting prefix just got bigger at some point, so from a /22 it switched over to /21 here. That is great. So what is this? There is a prefix here which stopped but coincidentally, a number of other prefixes popped up. Well, you will actually recognise the numbers there better than I will and this is the K?root instance. So for a while it was stably announced from one single AS and we switched over to our K?root AS, this is LINX. We switched over to our own ASs and we changed the behaviour. Well, this was the time where we introduced Anycast, so it's kind of clear how it goes over time. AS 3333, it has been stable, rock solid and it started to announce more and more prefixes as time went by. So what this illustrates is that you can get a lot of data between the couple of clicks and then you can make sense out of it, hopefully. I am not going to go into these stats, not interesting. Geolocation, that must be fun. So where are these guys?

So we can see almost all of it is in Holland, as of now, that is the current history. Some of it is in London, that is LINX, still. If someone can tell me where root in Holland is, that would be nice.

DANIEL KARRENBERG: That is a secret.

ROBERT KISTELEKI: Secret. This says this is root in Holland. Actually, I didn't know about this ?? ignore it. What we can see is this has been true forever, ever since we know at this stage. Let's look at more interesting prefixes. I.rootservers.net. Since we finished this tool a week ago ?? finished ?? we reached a stage where we can show it to you ? Daniel jumped around like a kid, I can find this and that. For example, iroot is one of them. You can see the history of them, that it has changed providers from here to here and we can actually look up who they were. I am not going to go into all of these details now because that would be just too much time. But what you can see here is ?? oh, those guys again AS 3352, so they have been here as well. I think that this actually shows us that the blue one, the green one is /23 and the blue one is /24 so they started to use Anycast at this point. That is cool.

K.rootservers.net. OK. Ten years of routing history again. This is pretty clear, again. On the right?hand side we have some measurement artifacts, don't mind them. You can see these two prefixes have been announced in parallel so this is again Anycast in action. And the Spanish guys are here, they are everywhere. OK. A.rootservers, this is Verisign. I am ?? and I am only looking at the routing page, I didn't go into all of the details. You can do that. This is something if someone can explain what is going on here, I would be happy to hear that because both of these ASs are Verisigns as far as I know, they seem to be jumping around from AS to AS and back and forth. The Spanish are here. OK. Finally, m.rootservers, the Spanish are there as well. Can you imagine that? This is actually in Tokyo, I believe, because this is ?? but they are everywhere.

Google.com. Let's see what we know about them. They actually received a shiny new prefix at some point. At this point, somewhere 2006, yes, and they started to announce that on the Internet. That is fine. The Spanish are here. Before that, it was used by someone else so this huge gap actually shows us for a couple of years this prefix was not visible on the Internet. So it has been reallocated, OK, that is cool.

Moving on. More interesting things, DNS neighbourhood, I will ask the question, this is a quiz: Who recognises this prefix? No one does. I can tell you, well, actually you will be able to figure it out yourself if I show you what domains are in there. Source forge, thinkgeek. Different company. Same DNS neighbourhood.

Moving on. A very Dutch ISP. I think this will be more recognisable to some of you. Yes. So they have been active, it seems. So this is the current state, again. If we look at history, oh these are the cities they have been active in. Right, fair enough, nothing magical here.

DNS, routing, black list,. That is going to be interesting. This is access for all. And they have 34 resources in black list currently and this is how it went over time. They really had some mad customers, I guess. So this is some information that they could use as well. You can actually monitor your own prefixes and I imagine the first case sincerely everybody will be looking up their own prefixes, that is fine.

There was a mail, just a couple of examples, to NANOG saying we obtained this shiny new blog but it is ?? it is in black list. Well that just happens so these guys had to clean it up and just to show what happened there: This is their effort of cleaning it up. So they have been trying to do that. That is fine.

Unfortunately, we don't have enough data to do this. This was one of the initiative messages, the very early ones that we actually looked at and we thought about this service, so there is this ISP somewhere in the ARIN region who received a shiny new block from ARIN and they started to see weird things happen: For example, that is code from the google.com comes up in Spanish, sends traffic to somewhere else and none of the credit card transactions are accepted. Now, that is a bummer if you really want to run a website where you want to sell stuff and ?? so apparently this was registered in Chile, the point here if the hostmasters in ARIN did have access to such a service they could have checked the state of this resource before they gave it out.

And finally, just to close this off, we received two new /8s from IANA so we immediately popped up, what do we know about it, so let's see what we know about them. Obviously the main question is are they used or have they been used. Currently, RIPE NCC RIS number is used which is good because that is our standard way of doing it, but ten years of history shows that even though this block was not allocated, there are bits and pieces popping up everywhere. And I am pretty sure that the Spanish are here as well. Yes. It's them. Well, it just happens. So if we zoom in to the last month or so, then we see debogonise, so this is how it's supposed to be. We start announcing some prefixes. And just to close this off, are the other /8 that we received, you can click through all of these yourself, that is fine. That is a bit cleaner but we can see bits and pieces popping up in recent times.

So that is it. I encourage you to go ahead and explore this tool, check your own prefixes if you want or someone else's, for that matter. You can dig up a lot of interesting history here.

Further plans for this: We do want to add AS number and IPv6 support. The problem with those are we have much, much less useful data for v6 than for v4, but we will do whatever we can to give ?? to make this useful. We want want to specialise to different user groups, our own IPRAs can start using it this to check prefixes. An ISP is interested in a different aspect or law enforcement is interested so we think that we will just to different user needs, I am an ISP and want to route this customer, give me what is important for me. I am law enforcement, I want to know how to shut them down, so that is the focus. But the data is essentially the same but the presentation can be a bit different.

We do want to add more aspects and more details to existing ones. So for example resource certification is coming, that explains the certification history of this specific prefix. And more importantly, we want to hear your feedback. Do you find this useful? Does it work for you? Does it help you? And please start exploring it and put your comments back into the forums. And that is it for you. Any questions?

MIRJAM KUHNE: Thank you. Thanks, Robert.

(Applause)

Franz, where are you? Do you want to come up and go on. I think we will keep questions until the end, otherwise I am afraid we won't have time for Franz's presentation.

FRANZ SCHWARZINGER: Hello. I am Franz Schwarzinger. I am going to show you something new that we made also as part of RIPE labs which is called RIPE NCC Netsense. Netsense is all about measurements that we do at the RIPE NCC be it active, passive, measurements, and we really, really want to make the data that is generated more easily accessible for you.

Before I am going to show you this exciting new thing, a little bit of history, how did it all start and begin: Quite a long time ago, almost exactly 19 years ago, the first Hostcount was published. This was basically just a count of DNS hosts in Europe and it looked like that, very simple. But what is interesting about this, this has actually not been done by the RIPE NCC but by RIPE ?? by the RIPE community at the time, which was very young. A few years later RIPE NCC was founded and already as part of the RIPE NCC the routing information service has been born. Routing information service also known as RIS is massive collection, we have 16 collectors around the globe and a lot of data that you saw in Robert's presentation comes from that measurement service.
And today, altogether we have more than terabyte of data collected.

Then in 2000, a year later, we launched test traffic measurement service which is more of an active measurement tool and we do highly accurate measurements around the globe and we can do that because we have specialised hardware quipped with GPS clocks which allow us to do one?way delay measurements which is the only service on the planet that can do that.

And in 2004, we launched DNS monitoring service. This was born out of the route servers who wanted their servers to be monitored, but these days we also monitor top level domains.

More recent history: In 2006, Information Services was founded. This was an attempt of the RIPE NCC to bring all those measurement services into one place, in one team that can work really on this and focus only on those measurement services, and we did quite a few imprompts over the years. One thing we simplified TTM pricing which was very well received and we got a lot of new subscribers. We put in a lot of work into making the back?end faster and more resilient. We developed a few new tools to access the data such as RIS dashboard and new Hostcount interface and expanded our network to span the globe. And I am very happy to say, today, that we are now spanning six continents, the only one missing is Antartica, and that is very nice.

But all together, we are still a little bit missing the wow factor behind all those things, so last year, we started to get some help from outside and we found a very nice company based in Amsterdam, just very close by to our offices, called SunnyDay, one of the first things we did, started to catalogue our entire portfolio which turned out to be more than 30 tools and services and tried to make sense out of them.

And this is what we came up with: It's basically two axis, on the top we have ?? we categorised everything from operational to strategic, that is basically our user needs.

On the operational side you want to monitor things, you want network ?? as network engineer, my network is not working any more, do something with it. After something like that happen, you want to diagnose it, why did it happen, where did the problem start, how can I avoid it in the future, and more on the strategic side then we have forecasts so how fast is IPv6 growing, and things like that.

On the other axis, we have the things that we offer you, and on the very bottom we have data sets, those are very, very critical for our operation, without data sets we can't offer you anything really. On the next level we have tools which allow you to access this data and on the third level we have analysis services where we would come, look at this, look at some of the data and provide a report about it.

Again, on the bottom the data set is very important, fundamental basis, on the top we find out the tools are very important for the monitor and diagnose part and forecast and diagnose part is very important for the analysis services.

But those two axis are very nice and good for us to know but it's really important for you is that there is a third axis and the third axis actually are you guys, our users.

So we did a little bit of analysis of our stakeholders and we came up with seven groups of our most important stakeholders, our most important users and we put them again into this diagram, and what we found out ?? well, the data sets again very fundamental, there are a few user groups that want to have direct access to the data sets but really where most of our users are clustered are here and here, the top part being more governments and journalists and the bottom part more technical people, people from ISPs.

And for our first version of Netsense what I want to know you today, we decided to focus on this middle part. We do that because we saw that most of our very important user groups, and especially you guys who come to the RIPE meetings, are really in those two boxes.

So what else did we do with stakeholders? We did interviews with you at the last RIPE meeting, we interviewed about 20, I think, people from different companies, different industries and tried to figure out what ?? how they like our current measurement efforts, how we can improve them. And we got amazingly good feedback from you, and this is very nice and the four biggest parts, we need more user?friendly tools, we need to present our information better, we need more global overviews, that was especially something that came from the government, and people wanted more like drilldown approaches. So you can see a very high level overview but then get into the details like clicking around. And yes, really I would like to thank you all, a lot, who participated in that and gave us this wonderful feedback.

Another we saw from our RIS dashboards that we have at the moment are very well accepted, people like them, but a bit of problem with them. They have a lot of information, as you can see here, there is a lot of different things where you can click on and go to and you have to read a lot and once you get to some data it's a little bit cluttered and you have a lot of stuff and you don't really know where it comes from, it's a little bit too much and it's not very pretty, it's not so sexy, it's not wow. The first thing we want today do was make the user experience much better and a very, very first thing that we did draft of this whole thing, was this ?? again, it's not pretty, but it gives you kind of the idea of where we started with. On the top left, we have an overview of prefixes. Very simplest showing you which prefixes are announced by RIS. Naturally going from the left to right we show information about specific prefix, once you click on one you will see information. On the bottom we have space reserved for information that is about the whole AS and you could see consistency check. As I said already, it's not really very pretty but we thought this user flow would work very well and we had the feeling by doing it that way we could easily extend it in the future and add new things to it. So, but yes, the really nice thing is coming now, Netsense, how it looks today actually after months and months of hard work and you are the first people to see it right now so behold, this is Netsense.

(Applause)

Thank you. So you see it's ?? the approach is similar to what I showed you in the picture before but now it's actually pretty and I would like to show it to you live. I hope it works. Firefox. I was going to use Safari. We actually start off with Overview, how is the Internet doing today? We are taking the data from our TTM network. We have ?? we select boxes from all of the continents, now that we have this nice coverage we can actually do that, and look at the relative loss between nodes on those continents and you can see the Internet is doing pretty fine today, it's all green, all 100 percent, which means hardly any loss happened in the last 24 hours. What you can then do if you are a network operator, you can go to this box up here and enter your AS or prefix. I am just going to enter the RIPE NCC prefix for now and click the sense button and see what happens. So as I showed you before, on the left side we have a list of prefixes that are announced by the RIPE NCC and in this picture you see at the moment, it shows up here, you see the visibility of the AS that you entered. Whatever you enter you will always on the map first see what you actually wanted to see, and in this case it's visibility and you can also click on the map and zoom in and see where all this data really comes from. So we have this drilldown approach that you people asked us to do. And also, you might want then to look more into what is actually happening with those prefixes. Click on it, the map refreshes and we see, oh, it's perfectly visible everywhere. If we look at the other ones the IPv6 prefixes not as visible as others. So that is where the less visibility from our global overview came from.

If we scroll down a bit, we have the part about the whole AS and for our first version we put here a new and improved routing consistency check. And what it does basically, it looks at your routing information in the RIPE database and it looks at the BGP routing tables and compares them, here you can say it looks pretty OK. Only the IPv6 one doesn't work here. And you can sort those tables, you can filter here on the left?hand side and switch between your prefixes. This is pretty much it for now. We built this user interface in a way to be very extensible so we are going to add new features in the future. For now, it's a very nice first version, I think.

So let me go back to this. This was the demo. RIPE labs, how does it tie in? Well, we are launching this Netsense this week via RIPE labs and it will be available to all of you sometime tomorrow. At the moment, you can watch it and look through all the features that are there and you will hear me explaining a little bit more about it. And Netsense also really, really, really needs your feedback. We want to improve it. We showed you a very, very first version at the moment, but we want to improve it further and we need your feedback to tell us what kind of information are you really interested in. So try it out, watch the screencast, participate in the RIPE labs forum, there is a post on RIPE labs showing ?? explaining Netsense in a bit more detail and you have a forum attached to it where you can participate in discussion. And you can also send us an e?mail to Netsense@ripe.net and if you want to get a little bit more into the technical details behind Netsense I would invite you to come to the test traffic Working Group tomorrow and we are going to show you a bit more about this.

This was it for me. If you have any questions now or maybe Mirjam wants to keep them to the end.

MIRJAM KUHNE: This is the end.

(Applause)

MIRJAM KUHNE: Thank you. Thank you, Franz and thank you Robert. Well, we have some time for questions now, if you do have questions either about RIPE labs as a whole or what some of the stuff that is on there at the moment, I see some of you already registered and participated and left comments on the site. That is very good. And I will get back to that later on the site itself. No questions? No comments? It was all crystal clear? I doubt. You are all stunned and overwhelmed ?? or hungry, maybe. You want to go for lunch. Oh, yes, there will be T?shirts out there. There is a demo stand out there, I will be at during the lunch break but also doing every ?? during every morning break throughout the week and I will be there and some of my colleagues will be there from time to time we can show you a bit more what is on RIPE labs. If you come by after the session, there will be some brochures and also a T?shirt if you want to pick up a T?shirt there. I can't believe there is not a single question. I am not sure if I should be happy about this or disappointed. OK. All right. Well, thank you very much then for your attention and enjoy your lunch break.

(Applause)