These are unedited transcripts and may contain errors.
Plenary Session
4:00 p.m. - Monday, 13 May, 2013.
CHAIR: We are about to start, so, please make yourself comfortable. And we are starting the second session of the Plenary, and I would like to invite the first speaker, who is Michele McCann, with Teraco, on building the keys for Africa, where she is talk about how to reduce the [](presentation) how to reduce the barriers and how to bring Africa to the Internet pool. Please...
MICHELE McCANN: Hi everyone. Thank you first of all for welcoming Africa to RIPE. I mean, this is quite an event for us. We are not normally participate so this is quite new for us. So, today I am pretty much going to take you through pretty much a very broad level overview of Africa. Where we're at the multihome and...
So, from an Africa perspective, if you currently see the top corner there, it actually highlights it for the top hubs within Africa. So from a business case point of view, those are definitely those areas you need to start concentrating on, any of you guys keen to coming into Africa. You will see this is a famous map that we use locally within Africa and pretty much it represents the geographical size of Africa which shows the issues we have faced over many years besides the political ones. So pretty much population at the moment, I am going to read some interesting stats to you guy I say. At the moment we have got over one billion people, which is about 15.6 percent of the world's population. And of that, only 7 percent are actually Internet users. So, you'll see there is quite a low amount of Internet users. Penetration on the Internet side is about 13 percent and rising. And another interesting fact is Africa is the second largest mobile market globally after Asia. So you'll see quite a bit in terms of our GDP growth and the way we are going from a GDP perspective.
So, as much as there is a lot of fear around Africa and where we're going, there is a hell of a lot of business opportunities that we are starting to see, and that's probably because we are starting on pretty much on a base of zero, and getting ?? and trying to grow quite quickly. So infrastructure projects is a big focus for us at the moment. Financial services and a lot of international financial companies coming into the physical environment, as well as the local guys starting to expand throughout the continents such as Standard Bank, First National Bank, etc..
Government spend: So, there's been a lot of increase in terms of Government spend, and some of the projects you guys might be aware of is one of the like the SKA project, where we are putting down satellite hubs etc., throughout Africa.
This is also quite an interesting map from an Africa point of view, we use this one all the time. And if you have a look there at the one that sort of ?? the smallest line on that map, which is actually sat 3, so interesting enough for the last kind of ?? besides the last two years, we've pretty much been running off less than a terabit worth of capacity coming into the Continent itself. So I mean you can imagine I just heard now that this conference itself is running off a gig worth of capacity, which is kind of an unheard of thing and we haven't had that kind of pleasure in terms of capacity until recently.
So now from a capacity point of view, we have currently grown that we are able to connect to four continents, globally. SA itself is the 15th largest telecoms market in the world, as well as we've currently got over 25.8 terabits of submarine cable coming into the physical continent, growing to, once the bricks cable lands, growing to at least about 16 terabits of capacity coming into the physical continent. And on top of that, we have seen a huge price reduction in international capacity from about a year ago we were where sitting at about $100 a meg to now under $30 a meg. So this is actually created quite an exciting opportunity for us, as well as quite a challenge for us. As I listened to some of the previous speakers where you guys are going on about IPv6 and scaleability etc., etc., we are still trying to figure out IPv4, so obviously from an educational point of view that's an interesting challenge we need to overcome.
Quite an interesting site. This is pretty much from an SA point of view. We have had major amount of changes that have happened and as you see the time period has been quick. The biggest changes that 2008 Altech ruling where, basically, we had a monopoly of telecom which basically was the only operator you could buy services from and as of 2008, Altech went to our Government and said, hang on, we need a court ruling where now there needs to be multiple networks who can provide services so now we have got over 300 networks actually building out on the rolling out fiber networks throughout our country and our continent from an Africa point of view.
Another major change that's actually happened is you'll see the WACK and EASSY, this is how much capacity is actually coming into the continent now which obviously shows a huge amount of opportunity for us.
So, from a South Africa point of view, is essentially we have been pretty much bench marked, together with our brothers in Kenya, Egypt and Nigeria to be one of the key pinnacle points from an economical growth side of things. So, I'm going to read you some stats here which is actually also quite interesting.
So 2013 outlook, so, looking ahead to 2013 report, revealed that 71% of local businesses expect and increase in Revenues in their business over the next 12 months and forecast throw from 2013 to 2017 at 4.1%, compared to the previous queer of 2 .5 percent growth.
So the reason why I'm showing that is that shows that we have got businesses and start?ups coming alive and a huge amount of, which represents once again on a GDP perspective.
So, key market growths as from a South Africa point of view, is tourism, and I'm sure a lot of you would want to go to Cape Town so, that's normally why that's top of the list there. We have got great wine, more than welcome. And I did bring some bottles if anyone would like so share a glass. Mining and minerals so, our guys guys around this kind of thing is people would you know like an Anglo American for example, that's pretty much where South Africa kind of started, is with your gold and diamond mining and that still continues to grow, despite what you might see on the news from an international perspective, it's still massive growth portion for us.
And a big portion for us is the ICT side of things so that's why I'm here today. Because previously this didn't seem to be such a growth side for us.
So, here is a little bit of an overview from an Africa ICT landscape. And so you'll see that 99 percent of the content consumed by users in Africa is imported. So, this has been a huge debate from an African point of view and it's been one of my only personal debates is where essentially we constantly go into the likes of the Googles, the Microsofts, which currently a lot of us, are actually hosted out of the country, specifically let's sake Microsoft for example, where we need to use that under sea capacity and a lot of us are trying to sort of say guys, please bring something into the country so that we can locally connect on to that to improve latencies and performance of applications.
But, personal argument that I have with my African colleagues is that we should be producing content and distributing out to people like yourselves from an international perspective, where you guys can actually start utilising African content and distributing that out. So you'll see that obviously mentions roundabout a lot around the trade and deficit or Internet transit deficit that we mentioned.
So there is that 13 percent and rising and an interesting one, which from a Facebook perspective, is at the moment, we are doing, from an African side of things we have got 51 million users at a 4.8 penetration rate, so we clearly love Facebook and, once again, largest mobile market. So key thing for any distribution that you need to put into place is making sure that you can distribute across a mobile application, when you want to come into Africa.
Terrestrial fiber. This is another big myth that I'm hopefully going to solve today. That we do have a lot of terrestrial fiber already in place and growing. So you'll see a lot of lower part of the continent is actually covered from a fiber perspective, and over ?? and the interesting fact is that over 3 million people are within a reach of an actual fiber node. So before, and I'll say before about three years ago pretty much, was majority of connectivity could only be accessed via a satellite, back hall kind of environment which as you know is that you are sitting on a 600 millisecond second latency issue which means certain applications wouldn't work. Now, we actually do have fiber available and it's continue continuously increasing to roll out, which solves a lot of our cross?border issues, like you'll see between South Africa, Botswana, etc..
This is probably the more exciting part, so from a peering point of view, we have been very ?? we don't have a links or an AMS?IX kind of environment, we are very new off the starting block when it comes to peering, so you'll see we have actually only throughout the entire continent got 26 active another Internet exchanges, which does prove to be a challenge for us in terms trying to promote content coming into the country etc. And there is a lot of learning that we need to do around our African exchange environments. I mean even down to commercial models, how do you pass it, how do you technically distribute, things like IPv6, etc.. so, we have had to overcome a lot of those and hence why only those 26 Internet exchanges are actually live and active.
From another big issue that we have got in Africa is that from a physical facility point of view, so you guys would have something like a Telehouse or an Equinix available. From an Africa side, the only really neutral data centre that we have that serves the entire continent is Teraco, which I'm from obviously. So, there is a lot of growth around that and as you can imagine to build one of these facilities and for a Teraco to try and roll out to the rest of Africa, it's a huge amount of capital investment. So that's another challenge that we obviously face. But slowly overcoming.
Then, I think the really exciting part about, it and you'll see the lady there doing the thumbs up, the free Facebook, is with peering now actually brings new product opportunities. And some interesting guys have actually rolled that out is orange, who I think are here somewhere. They have actually rolled out that you can start getting free Facebook via the mobile environment as well as obviously the browsing side of things. So there is some interesting products that have actually come out.
Then, the last block is also quite a political debate for us is obviously reduced IP transit fees. So, because your side from a European perspective is actually very mature, there is a lot of previous companies earned a lot of revenue out of physical IP transit, and now with peering becoming a reality, there is a lot of changes that need to happen to various SIP business models and that is actually happening quicker than we expect, which is once again fantastic for content owners coming into the country.
So, I have kept it really short and sweet because I'd like to actually give you a bit of a chance to ask me questions, specifically around Africa and any ideas. I have stayed away from the technical side of things as well, rather focused on the commercial so that you can have a little bit of an opportunity...
Any questions?
AUDIENCE SPEAKER: Hi, I am Vesna Manojlovic from RIPE NCC community builder for RIPE Atlas among other measurement tools. I am not really asking a questions, but suggesting a cooperation of a sort. It would be good from our point of view, if you would deploy right Atlas probes throughout Africa to actually measure over the years this growth that you are expecting. So, I I am inviting you to find me later and to talk about it.
MICHELE McCANN: Absolutely. That would be fantastic. I mean from an Africa point of view once again we are going through a massive learning curve, and we have to try and really jump this kind of digital divide between the two continents so anyone else that has got suggestions around that, that would be great. So thank you.
RANDY BUSH: Vesna, instead of giving you 25 to bring to AFNOG next month, give me 50.
To see all that wet fiber around Africa is just stung and wonderful, if you know the history. It was horrible.
MICHELE McCANN: Your history you rolled out the first digicode net.
RANDY BUSH: It's true, sadly enough. The terrestrial fiber, could you discuss regulatory issues crossing the borders, which is I know in east Africa from Tanzania to Kenya, crossing that border, there is fiber going mmm and they are looking at each other, etc., so, that's worth discussing because that's a little different than we're used to in the States or Japan or here in Europe.
MICHELE McCANN: Absolutely. So from a terrestrial fiber perspective, there is multiple issues so there is the physical, which you see we are kind of getting over in terms of actually trenching and rolling out the fiber but we do have a hell of a lot of issues in terms of the, from each Government actually start to go work together like, as a crazy example, is if you go to the Mozambique border, you'll be sitting on the South African side and you get great Internet and literally you cross less than a kilometre and suddenly you have got no Internet. You can see these two physical nodes sitting next to each other and that is something that Africa, as a whole, need to actually overcome. We are starting to see it happen between the static regions, so, from a South Africa to, let's say, a Zimbabwe, sets, so those kind of collaborative measures are starting to happen but up north, it is still a huge challenge and generally, people ?? a satellite is still being used from a north perspective. Does that kind of... it's a long debate, we could speak about that for hours.
AUDIENCE SPEAKER: Alex. You mentioned about lack of local resources in the Africa. So, just small remark. I think it should be operators initiative to deploy some caches like Google cache, or route DNS which doesn't need a lot of resources, it is I have been deploying some of these things and it is very, very easy, you don't need some strong resources, and... yeah... just mostly only ??
MICHELE McCANN: And I do agree luckily from our side we have had a huge amount of support from Google, as well as from Akamai. What we do need is we do need other content as well, so other content meaning let's say, for instance, and I know lot of you might not like this, so Microsoft, so Microsoft for example has launched a huge 365 initiative into Africa, and the biggest issue we have got around that is, let's say, for instance, a C com cable goes down, suddenly all the businesses utilise ago 365 will be on down so, there is those challenges and I mean, local DNSs etc., we do need to start bringing more of that content into the country and I am hoping by us having more under sea cable availability and more terrestrial availability, as well as facilities like a Teraco starting to be available, that it provides a low risk entry point for large content providers to be able to invest within the country.
AUDIENCE SPEAKER: Fergal Cunningham from the RIPE NCC. I have a question on chat from David from IP6 Net Limited. Do you suffer with security issues on infrastructure cabling etc.?
MICHELE McCANN: Physical security? So, in terms of security issues, I'll remark more on the physical side of things. So, from a copper theft perspective, we have had huge issues around copper theft where the guys actually dig up the copper and actually start making all these fancy baskets that you buy over here. Next time you buy a basket, just remember you are probably cutting off a business somewhere. So, that is an issue that we have faced. Interestingly enough, we have had faced the same kinds of difficulty, not to such an extent on fiber, because it's the guys kind of dig down, take the fiber and then they realise, hang on, I can't do anything with this but then the fiber is already broken. And then the last issue that we actually face is probably a competitive one and it's actually a crazy thing that happens around fiber. Specifically, I'll use Kenya as a great example and one of the operators was complaining about this this Kenya, is the guys will essentially go in, they'll lay the fiber and do, you know, normal fiber to the home and sell their services, and then in the evening, their competitor will come in, physically dig up that fiber, so they sabotage the fiber, cover it up and then lay their fiber next to it and then onward sell their services, so, you kind of have that kind of competitive nature happening in certain areas within Africa. I think from a South Africa side of thing it's stabled a lot more, I think it's just generally weird that we're the country sitting at, versus, let's say a Kenyan, etc..
AUDIENCE SPEAKER: Sandy Murphy, SPARTA. I was interested if you could talk about the sorts of arrangements at the different exchange points you talked about 26 different ones. Is it a place where people go to do buy roll peering agreements? Is it open peering, route servers, anywhere anything like that?
MICHELE McCANN: So from that perspective, it depends on which country once again. But, I'll use what we have done actually and worked quite closely with the LONAP guys etc. To kind of educate us going forward, so we have launched in South Africa, a Nap Africa, which is called Nap Africa, which is an open peering environment and this is the first scene in Africa where we have multiple options where you can do anything from multilateral peering to bilateral, depending on what your organisation needs, and we do have route servers in place. There is a number of the other exchanges who not necessarily as mature in terms of that, who are still sitting with a bilateral kind of arrangement and no additional value adds, so we do as exchanges, we have actually got an exchange forum called [Afex], and we actually collaborate together to actually start discussing what each exchange is doing and which one is doing right and wrong, but typically there is a lot of growth that needs to happen around the exchanges side of things.
AUDIENCE SPEAKER: Warren Kumari, ex?South African.
MICHELE McCANN: Why are you ex?
AUDIENCE SPEAKER: This was really interesting, but what can this community do to sort of help? What is the ask here?
MICHELE McCANN: So I think step one as an ask, is, from an educational point of view, I think that that's always where everything starts. And what I'm doing here is to obviously ask for you guys to actually start monitoring Africa and seeing how you can add knowledge and just basic education around these environments. So we have a lot of initiatives happening with, let's take ISOC as an example, where a lot of the guys are going around and focusing on peering exchanges and getting new peering exchanges up and running. I mean, the smallest of countries in Africa called the Gambia; for example, I think, step one, education is key. Step two is from a continent perspective, so, making content available, and step three which we are seeing already, is a lot of global carries coming to invest in Africa. And Gaza have done it really really well as an example, there is orange, who literally have come in with the international knowledge and been able to pretty much take over from an Africa perspective to become one of the leading operators throughout the whole of Africa. So, I think there is three key things which will take a long time from your perspective and from ours, and from an Africa side of things as we very open to actually start listening to your guys ideas, which I think previously we were kind of closed?minded.
CHAIR: The last question, please.
AUDIENCE SPEAKER: Blake Lillis, I have my neo telecoms hat on today. From ?? you mentioned that you were the first multi?country carrier neutral provider in Africa. Why do you think that is. Is more a regulatory thing or an impediment to the business environment or a culmination
MICHELE McCANN: It's a maturity. You saw in my last block I said IP transit, I mentioned that quite hectically and basically previously a service provider would build a data centre to put their own bandwidth etc. Within that environment and then oh by the way you can co?locate with me. And then they would charge transit fees. An unbelievable transit fees just to be able to interconnect into different cabinets so. If I'm going from cabinet A to B in the same data centre I'd be charged transit fees. So, that's how business models were previously. Then what actually happened, funny enough, was, Teraco's born out of international investment, so Guy [Wilnay] from Equinix, etc., where they decided that let's take the risk and it was quite a risk from a capital point of view, to say okay, guys, it's about time Africa has a carrier neutral, vendor neutral facility available, and luckily for their perspective for shareholders, obviously the timing was right, because all the undersea capacity was there, the landscape was now open through to Altech's ruling so they were able to fill up the data centre. So I do hope that we do find some more, as business development it's always fun to have some competition. So I hope we do see some more neutral environments opening up but we have to see what time tells.
AUDIENCE SPEAKER: Which other markets are you looking at besides South Africa, if you are allowed to disclose that information?
MICHELE McCANN: That's fine, not a problem. We have got three facilities, just to give you an overview: Jo'burg, Cape Town and Durban, Jo'burg being the biggest one. We current are looking at Kenya, we did have a look at Nigeria but due to power and infrastructure constraints, it didn't make sense to actually physically build a facility there. So, I think the next probably key from an investor point of view, would be Kenya.
CHAIR: Thank you, Michele, for a very interesting presentation about the changing landscape, IT landscape and Internet landscape in Africa. I wish you luck and success.
(Applause)
CHAIR: Next I would like to invite Jan Zorz from the Internet Society who will talk about best current operational practices and share some ideas about them.
JAN ZORZ: Thank you. Hello, my name is Jan Zorz. I come from Internet Society. And I'm here to share with you some thoughts about the best ARIN operational practices and repository and interesting of building.
I will not go through my history, many of you know me and I know you guys. I will just tell you about the change.
I joined the Internet Society deploy 360 programme in December, and my goals or tasks are IPv6, best current operational practices and to increase, to try to increase the operational feedback to the IETF.
So a few words about the deployed 360. This is deployment oriented port role. We are trying to get as much as possible of deployment?oriented documents with examples and how to build stuff, how to deploy new.
Technologies: Why? Because, there is lots of documents around the Internet, but then you need to understand which documents are credible, which documents are valid and which documents are not. So we are trying to build a repository of deployment or noted documents on IPv6, DNSSEC and routing security and resilience.
So, the next areas of work, two these areas are work are related, but we are approaching them as two separate areas of work. So, we have this operator to the IETF improvements in mind that we would like to suggest and also the next thing is operator to operator, where we are exploring the idea of creating a global repository of best current operational practices, this is called PCOP.
So process number one. What is the IETF? Many of you know what the IETF is. Simply, sanitises the protocols and services and implement and operators, you guys, are using those proposals columns and those tools to build the networks and to run the networks and to move packets around.
So the goal of the IETF is to make the Internet work better definitely.
But, during all these years, we observed an opportunity for an increase operators feedback to the IETF process. There are operators coming to the IETF but not enough. We would like to see more operators coming to the IETF and give more input. Because the standards could be better designed if the operators feedback from the people that are actually using this would give back the ideas how to improve the standards and how to make them better.
So, as I told you, we have observed an opportunity for improfit communication.
We firmly believe ?? being an operator for more than 20 years, I'm not coming from the academic part of the world, I am coming from the operators part of the world, and I understand that the operators need to tell the IETF how to design the protocols. So, we have been thinking about why, not so much operators are coming to the IETF meeting. Of course, it's a lack of manpower and time, the daily issues, you guys are probably more happy being in a NOC than going to the IETF to fight about the how the protocol should be designed. And specifically, we observe the reliance on the vendors, you guys, you pay money to the vendors, you say your wishes. How the Internet should be run, what you need and vendors go to the IETF and try to standardise what you wish for.
Well, it's a reliance, it would be better if you guys would going to the IETF and tell what's wrong with the standards if you are not satisfied with them of course.
We need to understand that the operators have a longstanding group process and the IETF has a longstanding standards making process, we need to respect both of these processes and try to interact more.
So, I have my personal analogy for this is, a Formula 1 car, right, was developed ?? there was a case where a Formula 1 car was developed in the wind tunnel, still is, and there was a bunch of engineers that were measuring in flows and air efficiency on the car, and they find out that if they rise a nose for a bit the flows go better, the car is more air owe dynamic and the pressures are right, so they started to rise the nose, and when they reached the point when the air dynamics was perfect, so, they put together the engineering solution and made the whole package perfect by their standards, and then the driver came by and sat into the car and said, guys, I can't see the road from this car, I can't drive it because it's useless, the nose is too high.
If the driver, the operator would be involved in the process earlier, this could be prevented, they had to redesign the car of course. But, it was luck that that guy sat into the car and said, hey, I can't drive it, because otherwise they would ship it out and they would discover it just before the race.
So, the driver needs to be involved in an engineering process. Here, when we are building the Internet, it's the same. You guys are the drivers. You guys are driving the Internet. You are operating the Internet. You should be involved in the standards making process.
And process number 2. This is about creating a global repository of best current operational practices. And this is actually my first talk at the RIPE Meeting that I'm not talking only about IPv6.
So, today the best current operational practices are shared between you guys in many different ways. We have a hallway conversations, we have operator group presentations, how people deployed IPv6, DNSSEC, RPKI, whatever. We have different means to share this best current operational practices. But, there is no no central repository that is believed and that is agreed among the operators that one, that needs the current operational practice on how to run network X would go down the document, read the document, and and implement it.
So, there is too much clutter on the idea landscape, so you can't easily find a reliable vetted and user friendly documentation.
So, there may be a need for a globally documented repository, that documents how you are running your networks today, and how you are implementing the new protocols and new tools like IPv6, DNSSEC, RPKI, routing security or whatever you want to put there.
I'm hearing a question about, so this presentation is about the idea. I'm here to ask you for as much feedback as possible if you guys think this is needed at the end or not, how would you like to run it, where would you like to have it, because this is a process that should be owned by you guys. I will talk about that later, but I'm here in questioning mode, I want to understand how to build a proposal over the summer and say, okay, this is what we heard from you. Here is our proposal. How are we going to do it? So, these documents should be written by experienced operators. I usually get the question: What is BCOP than BCP? BCP can be written by an operator, because it's operational practice. BCP can be written by basically anyone that tries to understand how to use the protocol.
We are committed to assisting in the creation and the promotion of the global repository, and also we are committed to help with the regional, to start the process, to find the right people to drive it and then it form the whole idea.
So, what my proposal ?? well what our proposal is that let's create a discussion group or a Working Groups or whatever you want to call it, or a track, for example, in NANOG, NANOG created a track inside, so the whole track is called BCOP and there is the regional best operational practices effort, and creates an ARIN that are the main drivers in this are here in the room so I hope that they will come to the BoF ?? Chris is there ?? so, let's create an effort where operator comes together and start talking about putting these things to the documents. And the same group could, that's RFC 2119 could, be used for operators to communicate their needs and wants from the standards process. Probably in the discussion groups or Working Groups, the people working from the IETF are taking a role in standardisation process would probably show up and try to understand from the operators what are their needs. But this can be added later, but this BCOP process should be owned by you. We can facilitate the whole thing, we can help with the global repository, but at the end, it should be owned by you guys.
So, discussion topics: Presentations from operators that deployed things and that they are running networks. So, usually we see this presentations from the operators saying oh I implemented IPv6 in this way, I implemented DNSSEC in that way. And they come to the stage, they say that, they go off the stage and that's it.
In this effort, the presentation like this would be until the serialisation of the document. When a guy comes up and stage and says I implemented this, this is actually a call for help in the room to say listen, I got this experience, if there is anyone in the room that has the same experience that I did, please come to me, talk to me, we can co?author the document but this is basically, I am presenting here because I want to start a document about documenting the best practice. We can also have an open discussion of what the operators need to use the deployment of new technologies.
And later, an additional purpose may again, RFC 2119 be added in order to improve operator feedback to the IETF. We could use these Working Groups to communicate the needs and observations back to people from the IETF, but we need to understand that the process in the IETF is, even we we get the consensus from 100 operators. If one guy shows up and says I have the consensus from 100 operators, this would be counted as a personal opinion of one man.
So ?? but, we are trying to create the ways to interact also outside of the IETF.
And this is the proposed schema. This is you guys, operators and members of the NOG communities around the world and we could create a regional BCOP effort in NANOG, RIPE, MENOG, AfNOG, LACNIC, other NOGs, and try to produce some material, some documents, how things are run today, how things are implemented. And then we can create a global BCOP repository and a global BCOP committee appointed or elected members from regional BCOPs so, the guys that are running these regional BCOPs should meet once a week ?? once a month, and discuss what is done in different regions, and then select and propose the best documents to be promoted to a global level.
Where can we help? We can facilitate this one and we are committed to get engaged with starting the effort inside the NOGs around the world. We have been travelling around the world to present this to different NOG communities. We have been to NANOG, to Rocky Mountain IPv6 Summit, we have been to LACNIC, then we went to Barbados for CaribNOG, I'm here now, then I'm flying to Saint Petersburg for ENOG, then I am going to Zambia for AfriNOG and we'll try and get as much feedback as possible what the operators think and we will prepare a proposal at the end based on what we heard from the community.
So, what are the advantages? Of course, common understanding of best operational practices amongst operators worldwide.
This could improve the Internet vastly, because not all the operators are travelling a lot and are exchanging their experience, but could use these documents to improve their networks to see, okay, thee guys are really successful are running the networks in this way, why shouldn't we dot same? I don't know how to say this politically correct, but I will just not do it.
So, you heard from Wicket; this week in Geneva another meeting is going about ITU and all this political stuff. And the governments and the regulators are trying to get their role in the Internet. And they will keep coming back, they will keep coming back until eventually at one point in time, they will gain some right to regulate something. This is a bad thing. But if this happens, if this happens, that point in time we have just two options. First option is, if the operators agree and document how currently Internet is best run, and how to implement new technologies, we can point this regulators and the governments to that document series and say, listen, if you really must regulate something, then mandate this, this is the best operational practice for today. For example, BCP 38, use filtering. Internet would be really a better place if that would be implemented everywhere. So, if at one point in time, if the Government would come and say oh, we need to mandate something, but we don't know what. I would happily say well if you must, I'm not saying that's a good thing, well, BCP 38, right? And I'll saying that's not mandating something is not a good thing. But we might come to this point.
And the other option that we have is not a good one, if we don't put this documents together we have nothing in hand. So, the governments may invent their own rules and mandate their own rules, how to run networks, how to operate networks and so on. And this is not the best way of preserving the openness and the Internet as we know it today.
And of course there is BCOP effort improves the communication between the operators glowly.
So, I have a few questions. Who in this room is willing to participate in BCOP process and actually do some work? Show of hands...
Good. I put this question in just to make you understand that you can be part of this process. And at six o'clock, I don't know in which room, there is a BoF, we will be talking about this, we will get together, the interested people will get together in a room somewhere here and try to discuss how to start this community effort here in this region. You are more than welcome to join, and share some thoughts with us.
And the next question is: Currently, we have no preference where the global BCOP repository should reside. Should it be neutral? Free format, that we decide and form on the fly, learning as we go? We could start it under the eyes of the ISOC 360 and move it; remember, this process should be owned by you guys. We can move it elsewhere if you decide it so. Or should it be under the IETF as a new RFC series, or under the IETF as a free format series, or something else? We need ideas, need feedback.
So I would like again to invite you to the BoF at six o'clock, and I'm done. Are there any questions?
CHAIR: How would you like to go ?? do you want some feedback on those questions or will you tackle this at the BoF or do you want general feedback?
JAN ZORZ: General feedback and the feedback on the questions.
AUDIENCE SPEAKER:
MARCO HOGEWONING: Not to be impolite or rude, but the thing that sprang to mind is one size fits nobody and I love the initiative to try and fit something that tests global best current operational practices, but how far are you willing or did you think about regionalising some of this, it's really hard to come up with one best current practice that fits globally. Markets are different, economies are different, Internet penetrations are different and it feeds up but I can still imagine that that local groups can have their own repository set and that's my recommendation. This forum is one of those examples that we have got four RIRs especially for this. I admire the global thing. Please contract regional differences in development.
JAN ZORZ: Sure, of course, because the operational practice from, I don't know, LACNIC is probably not the same as in this region. So, yeah...
ROB BLOKZIJL: I would ?? yeah... I am fully supportive of this initiative and I would just like to give a word of warning. I think it was the last question where you should try to convince the IETF to start a new publication series or something like that. I think one of the reasons you are starting this effort is you are not happy with the way the IETF works. So, I would not, at the beginning, spend all your effort in trying to reform or educate the IETF. I think first do some real work, come up with some documents and then start walking around and asking people, do we have any ideas where we can have a repository for this?
AUDIENCE SPEAKER: Leslie Carr, I sort of want to disagree with the previous statement that a lot of regions have a lot of differences. I feel that, of course, you know, getting IP allocations might be different in Latin America, but how to best set up a filter for BGP peers should be the same everywhere. There is a lot of ?? there is a lot of commonality that we all have in our heads and it would be great to have it written down.
MARCO HOGEWONING: A quick response to Leslie. An example that actually ?? that was thrown at me last week was for instance, yes, we all run similar practices but for instance, you see in some countries in the Middle East that do require content filtering, multihoming, a simple task of BGP multihoming, of course, becomes quite more interesting if you are required to do content filtering and actually keep sessions together. A symmetric routing is all, of a sudden, a problem. Yes, we can still use a lot of same practices but sometimes you have to look into localities and tiny little problems that can throw you off and we can still fix, it but it still is a clear case of not one size fits not everybody.
LESLIE CARR: Wouldn't we just be able to have a content filtering best practice, as icky as that sounds, that could apply anywhere where people hate Internet freedom.
MARCO HOGEWONING: Like the concept of multihoming can be different based on regional differences and you could try and fit it into one document but maybe you need two documents and maybe you need a separate set for specifics. That was where I'm trying to go, get separate publication point or a separate set for regional differences.
CHAIR: I hate to cut this, but we are sliding into BoF slowly and the BoF is at six o'clock?
JAN ZORZ: Yes, somewhere around here. Brian tells me it's in this room, so after the break, come back here. We have a big room. Cool...
CHAIR: If there are no more questions, then we invite everyone to the BoF. Thank you very much.
Applause)
CHAIR: Okay. That ends the normal presentation phase of today's Plenary. And now we are going to start our lightning talks. So ?? actually, the first one that I have is Patrik. This is, Patrik Falstrom is going to be talking to us about fun with certs.
PATRIK FALSTROM: Okay. So, what I will talk about is not so much the actual work that we did, but the process, because I think that is what you should be aware of. First of all, a quick summary.
Internal server names. Domain names used in certificates for TLD and domain names that doesn't really exist in the DNS, so we had a look at this, got a certificate, said, oh, by the way cannot use on the Internet. Click this box, go on, continue, down you get a certificate, you set up a fake root delegate the site route yourself. Set up web server. Yes, everything works, done.
So, this was pretty bad. So, that was what we did and more about that at the DNS meeting, Working Group meeting. But what you should be aware of is that what we did in this Security and Stability Advisory Committee in ICANN, where I'm the Chair, is that, when we discovered this we found out that this was so bad, so we could not share it, not even with other parts of the ICANN community.
So what we did is that we created a working party that looked at the various things, is this really true? Is this really as bad as we thought? So we did a study. Just like we do all things in [ASAC], we formed a working party etc. And we searched through the observatory and we found that at least 157 CAs had handed out this kind of things and that's a lower bound because these are only the certificates that are actually visible on the public Internet.
So, we talked with the CA browser forum and found out that they were aware of the issue and they had already decided to stop issuing the certificates and revoke the ones that are issued three years after the first new DTLD contract was signed. We thought that was about three years too long time.
So this started to become really messy and our conclusion was that ICANN has to do something immediately. So, what do we do? And this is the key thing of this talk of mine. What we decided to do in [ASAC] is ?? was to, on the fly, invent a new process in ICANN which touches upon operation. So, what we did was contact the ICANN security team, ask them to take the lead, and first of all develop a plan for a coordinated vulnerability disclosure that all of us in the operational community, we know what that is. Everyone in the organisation has that, people working with certs are used to these kind of things but ICANN didn't have one. So the ICANN security team developed this process and contacted the Chair of the CA Browser Forum on January 23, and briefed the CA Browser Forum on February 5. This was sort of pure luck, sometimes you have to be lucky because the CA Browser Forum happened to have their yearly meeting in February, okay. So this was just pure coincidence.
They passed a ballot on February 26 this year which said that the certificates ?? the CA should stop issue certificates 30 days after the contract was signed with the new TLDs and 120 days after signed contract the existing certificates should be revoked. And given this was done in February, we in ASAC decided on March 15, we decided to publish our advisory, so our advisory was ready in December, and then we published it not until March when we saw that ICANN actually had done all of these be activities and we saw that things were happening.
So, of course, there are a couple of issues with this. First of all, that not all CAs are members of the CA Browser Forum, although every CA seems to follow their advice. At least the trustworthy CAs and those are the ones that we use, right?
It's also the case that we have a question in the discussion whether revocation of certificates for is actually effective or not given the big discussion about CRLs and the use of OSCP and other kind of protocols because people think that doing revocation checking in the browsers is actually slowing things down from the user experience point of view. And it's also the case that many of the certificates ?? sorry, a lot of these verification is happening on air gapped computers and networks not connected to the net. So, this story is not over and we are continuing in [FSAC ]to work on these issues and work with the new TLDs. But the interesting thing for this community is that what we did in [FSAC] when we created this report was work with ICANN security team to do a little bit more operational process than before, and come up with a security disclosure process that might, hopefully, don't have to use it again but we now used it once so we'll see how it goes.
Thank you.
Questions...
CHAIR: No questions from the audience?
(Applause)
CHAIR: So this next speaker probably needs no introduction, but I'll introduce him anyway. Randy Bush, from IIJ, will be talking about some measurements they did about ping and using that to measure latency.
PHIL RUSHTON: Put on your diving helmets, we are going below layer 4, we are actually going to talk about the dirty stuff. So...
We were doing some experiments with RIPE Atlas trying to measure and we found that we were getting really high dispersion of delay times, RTTs, and we were ?? this was really spoiling our measurements, and so we said, hey, we better back off and calibrate our tools, and we had done the calibration of RIPE Atlas previously, so we are interested in calibrating ping itself, okay. And so, the first place we found something ugly was ?? I'm working with, as you notice, Stefano and Luca, these are from Rome ?? so, they just grabbed a server in Roma Tre and ran a tracer out to Ashbourne, and a ping to Ashbourne and found wide variance and so they ran a tracer out, a Paris tracer out series. Let's make sure we know what Paris traceroute is. It uses the flow ID to explore the hashed equal cost multipath or LAG, which is link aggregation groups, which you take multiple circuits and you treat it as a single layer 3 circuit. And multipath is where there are multiple ways to get there, okay, and you treat them as equal weight, equal cost multipath and so the packets by the router are hashed over the alternatives.
And the way this is done by the routers is the dark stuff, the source port and the destination port in the UDP header are used in the hash, and when it's ICMP it's type, the code and the check sum, type code and check sum in the port unreachable response. So these effectively are hashed and decide which of the multiple paths are taken. This allows us to take a circuit which is a bundle, or equal cost multipath, and measure the latency on each of the hashing alternatives.
And, boy, were we surprised.
This is Roma Tre to Ashbourne, and what we are seeing here is these colours that are essentially vertical with a long tail here and a little bit of a tail here, are the different latencies shown by different hashes. The red is what you get from ping, this purple, pardon me ?? is the black, is if we took these and we did the floor of them, we truncated them down, and this says, oh, this is very near to ping. So ping is kind of the ?? the normal ping gives you the average of this. But this isn't too bad. We are seeing, you know, what's this? 106 and change, 107 and change. So this is 1%. So big deal.
Okay...
But we found out why. This is the path essentially when we're a large SIP, who I won't name, crossing the Atlantic, this is Paris, Marseilles, Paris, New York, Dallas, Washington DC, Washington DC, etc.. so here you are seeing them crossing the Atlantic using multipath because this is ?? SIP is run by a friend of the community, and, as he says, his current routers won't do LAG very well. So, he uses multipath and you get traceroutes that just look like a mess.
So then we said what's it like on a simple path? Here is a single layer three between two routers, one router in Dallas, one router in Ashbourne, and ?? so we have a server here, a server there ?? this is all within Dallas, and this is the one that got us. This is a 9, 10 gigabit LAG. So, 90 gigabits.
Here you go, the variants in delay is between 32 and 45 milliseconds. This means the ?? so, this is approximately 13 milliseconds out of 30 is the min. So this means the measured ping dispersion is 50%, right. You can't measure normally in this circumstance.
Now, again, okay, the black is actual ping without the hashing. So, you'll notice the distribution of that is about the sum of these and the red is if we take the sum of these and you'll notice it pretty well models the real ping. Where it differs is interesting, in that probably the probability in the hashing of hitting each of these is not equal. The red assumes they are equal so. Black says ping very little and found slightly different hashes.
And by the way, a trap that I forgot to mention over here is these great big stair case. This is Linux. Linux rounds the ping and says three digits and that's all you get. So it's kind of ugly to use for this kind of experiment, so all the rest of this stuff is FreeBSD and slow [Larus].
So in presence of this you can't use ping to actually measure latency reasonably.
So, we were looking ?? we are still working for why; it's very reproduceable, but the circuit loading in all circumstances were less than 50%. We tried different times a day. We tried different probe spacings and timings, and, by the way, that was 100 pings followed by 100 hashed pings, followed by ?? ten hours to run one measurement. So this wasn't just throw it out and see if it worked.
So ?? and we tried UDP and ICMP. The big thing we noticed was, where the operator was hashing at below layer 3. In other words, they hashed on MPLS, okay. Or they hashed on anything below layer 3 and in sum routers you can control what you are hashing on, then it had a very large effect.
So, here is one that is only a layer 3 hash. So, that's pretty good. It's Linux again so, you get the step function. But that's not bad.
So, somebody asked, what happened if something was in the middle of this? An event happened. Wouldn't that cause the dispersion? No. This is what an event looks like. Because you got those things that were before the event and those things that were after the event, and that was the length of the event.
And finally, this is kind of a table to kind of give you a feel for what we saw. This was the Roma Tre to Ashbourne, less than 5 percent dispersion, and ECMP and LAG were both involved. Here from Rome to ?? I forget where BXL is ?? Brussels ?? Rome to Brussels, 5 percent, Rome to London, 5 percent ?? all reasonable, both ECMP and LAG were used but in fact most of this was layer 3 hashing.
Tokyo to San Jose ??
Dallas to London, oh oh...
Paris to Ashbourne: Clean.
Dallas to Ashbourne: That was the one we saw earlier.
Dallas to Seattle, also ugly, so on and so forth.
So the lesson is that for actually trying to do measure of latency, ping is not a reasonable tool. Okay. And by the way, if you care that we have a detailed paper just submitted to IMC on this, so all this is there with tonnes more.
So, what I would like to ask is, is Robert Kisteleki in the room. Can I beg please, we will give you this ?? the code for this ping. Since the Atlas probe is now six times bigger than it used to be, can you slip this in there please?
RANDY BUSH: You don't have to give a commitment.
AUDIENCE SPEAKER: We will try.
CHAIR: Randy, before you step down. So, I guess I'm not a hundred percent sure. Is it actually ?? are the latencies varying or is it ??
RANDY BUSH: Yes. The RTT is varying. You are looking at RTTs. You are looking at ?? I mean, let's take the ugly one ?? what this is saying ?? we're okay for time by the way... you are the one, you're burning it ?? so what this says is, when you tuned the parameters that are hashed to select this particular hash, okay, you got an RTT distribution that's set pretty much smack on 32. And when you tuned it for that one, you got 46.
CHAIR: I guess my question is: is it really ping that's broken or the network?
RANDY BUSH: Oh, it's the network. It's the network. But it's not ?? but the problem is normal ping is not seeing that, right. When it's not the network is broken, this was ??
CHAIR: Yeah...
RANDY BUSH: But this is the ?? the black is the CDF you'll see with ping. [Ren], you can't wait.
AUDIENCE SPEAKER: I agree the network is broken and we'd like to fix, it but unfortunately we do not have a reliable way to monitor that and we have been asking vendors to tell us, for example, at LACP around three times over SMCP so I can see that and get that from aggregate because a maintenance in the fiber vendor provider, as it took a ?? lick down and bring it up and it's is completely different and I don't know about that.
RANDY BUSH: I can't even ?? on neither Cisco nor Juniper, I can't even find the length of those separate queues, right. Because one of my hypothesis was, well, what if YouTube happens to be hashing all to compete on this queue? You can't even find the length of the queue going into the line card, you can't do it. Okay. But... by the way the tools available, open source, and all that stuff, and so if you want to see if your LAGs are exhibiting this, just send me an e?mail or crystal or whatever, and we'll give you the stuff.
CHAIR: I'm going to cut line off here because I let Randy go on for really long because I was interested. So... not that I'm not interested in all the presentations. But...
AUDIENCE SPEAKER: Is there something you could learn from looking at the Xen one?way delay metrics instead of ping?
RANDY BUSH: I can't ?? one way delay means a lot of infrastructure. And I can't easily deploy that infrastructure where I want to measure. So, for instance, when I wanted to look at this, I'd have had to set up both boxes in both places with realtime, you know, something half decent better than NTP, etc., so the setup is ridiculous.
AUDIENCE SPEAKER: But the question is have you thought about ?? there is something ??
RANDY BUSH: Yeah, let me show you the paper, when you send ?? you can control the response. When it's I C ?? when it's UDP, you can control what the hash of the response is going to be. When it's ICMP, cannot select what it will be but it's predictable. So if you look in the paper, we actually separate the outgoing timing from the response timing to make sure we're not seeing something that's only happening in one direction. And it's happening in both.
AUDIENCE SPEAKER: One last question, or a comment? Sebastian Wiesinger, Norris Network. For me, when I look at this, I don't know if I'm missing the point completely, but my first instinct was to say it looks like the variance comes from the modern routing crushings in a lot of queueing in a lot of ways with a lot of QoS that's not really visible ??
RANDY BUSH: These aren't QoS kind of people. These are backbone operators, they move packets. QoS is deciding whose packets I drop, I don't get paid for dropping packets.
AUDIENCE SPEAKER: You know, where the packets are queued and which queues are serviced first.
RANDY BUSH: It could be. But the thing we could do there is, we couldn't go in the router and find the queues of the different ports. The only thing we could do is in all these cases the actual load was less than 50%.
AUDIENCE SPEAKER: We know it's there but we don't know exactly why. Okay.
RANDY BUSH: I have some suspicions. I suspect here ?? well, the next ?? the one that looks absolutely horrible ?? what if they are aggregating circuits that are on multiple line cards?
AUDIENCE SPEAKER: So it's ??
RANDY BUSH: You could do that. You can do that. No, it won't pass through two line cards, but separate hashing... don't know... we're still working. We're still working but if the Atlas probes can do this, give us this pingability, we can get serious measurements on this. Maybe we just happened to walk into a room that was ugly and the whole rest of the Internet is perfectly fine.
CHAIR: I'm sure that's the case.
RANDY BUSH: It could be.
CHAIR: Thank you Randy. Very interesting.
(Applause)
CHAIR: The last lightning talk for today actually comes to us from an initiative that RIPE has started called RACI, the RIPE Academic Cooporation Initiative, and this is an effort to get students from the academic community to be more involved with RIPE. So there are actually five students who kind of were selected for this one. Marcos Dias, who is going to be giving a presentation; also Nikolay Melnikov; Vaibhav Bajpai; Jurgen [Schonevelder], and Runxin Wang. So, these are the students, so if you want to come and talk to them, welcome them into the community and find out what they have been work on and that kind of stuff, it will be really good. We'd like to give a special thanks to Penny Storey and Brian Nisbet, who have been a lot of help for this particular thing and getting this initiative going.
So, with that, I guess I'll let you take it away.
MARCOS DIAS: I am afraid my talk is going to be more for shallow dive, fairly non?technical, and we are going to be talking about policy and social potential.
Thanks to RACI, [IBES] and Trinity College for their support. Information is the raw material of contemporary society, says [Minoch Astels] [phonetic] in the rise of the network society. Countries such as France, Greece, Spain and Finland have acknowledged Internet access as a legal right for citizens.
The UN human rights Council says that without Internet access marginalised groups in developing states remain travelled in a disadvantaged situation. This affects the world's poorest countries but also those living in suburban, rural and remote areas of developed countries. While the digital divide materialises through spacious segregation, it also emerges through technological segregation. Where those unable to pay for premium plants are condemned to patchy coverage, redundant infrastructures, lower speeds and reduced download caps, sometimes through artificial scarcity.
Lack of strong Government policies contributes to perpetuating the cycle. However, policies alone do not materialise into social benefits. If digital citizenship is become a basic rights, strong policies must be combined with access through an infrastructure that provides a minimum universal standard of service and skills necessary to benefit from access, where, in other words, digital literact.
Finally, we must account for the unforeseen potential of future applications, platforms and users.
The National Broadband Network, or NBN, is a major infrastructure project that aims to provide high?speed broadband access to all Australians. It is a main component of Australia's national digital economy strategy and its aim of becoming one of the world's top five digital economies by 2020, which, in turn, are resold by service provides to end consumers. 93 percent of the population will be served by a fiber optic to the premises, or FTTP, also known as FTTH, network of symmetrical download and upload feeds of 100 megabits per second. Speed and ?? key for services such as video conferencing, tele health and virtual education. It is a long?term solution that can be upgradeable to 1 gigabit per second and beyond.
The remaining 7 percent in revery moat location will be severed by wireless and satellite access. The Government's opposition the coalition are the front runners of the upcoming Australian election and oppose the use of FTTP on a cost basis arguing instead for a fiber?optic to the note or FTTM that, leaves the last mile to be dealt with Australia's aging corporate network. They have stated the NBN is not necessary at a time when Australia's road, railways and ports are completely clogged and recently argued that a download speed of 25 megabits per second is more enough to download four movies at the same time. The coalition's short?term vision puts cost above future social and economic benefits. It fails to acknowledge that Internet infrastructure can actually reduce transportation needs, for example, tele working and suggests that the Internet is no more than a glorified media consumption platform. A true model of digital citizenship involves the user as producer and active participants in the network society and, therefore, uploads speeds matter.
Their insistence that FTTN can be easily upgraded in the future to FTTP and their lack of clarity on upload speeds... copper have led their plan to be ironically dubbed the fraud band.
If Australia's opposition has any doubts about FTTP, they might want to check out Google fibre's initiative. A project to build fiber?optic to the premises networks in selected US cities offers 1 big GIT per second upload and download speeds. It is similar to the model being implemented by the NBN, except it is only being implemented in selected urban areas. Google fiber started in the twin cities of Kansas City and Google states that they will build fiber where people are most excited about it. Unfortunately, there are a lot of the excited folks in rural, regional suburban areas of the United States that are very unlikely to get it any time soon. In contrast, it is expected that 75 percent of regional Australia will have access to FTTP and if the NBN's original aims are implemented. While Google states its interest in sporting digital inclusion its roll?out is tied to a profitable business model that attaches continues to the rollout of the network. Each neighbourhood has to reach a target number of pre?registrations to guarantee Google fiber service. There are currently three plans, two of those are premium plans and offer the full gigabit of speed. The third plan is a so?called free plan. For $300 fee to cover the construction fee costs, the user can enjoy download and upload speeds of 5 and 1 megabits certificate second respectively. That's less than 1% of the speed of the premium plan.
In the meantime, the United States national broadband plan or NBP aims for 100 million use, homes with 100 big bits per second connection by 2020. However there are no specific policies or guidelines on how to achieve this. Even if Google decided as goodwill gesture to increase the basic plan speed, it is unlikely they will advocate for the millions living outside main Metropolitan areas.
So, can we measure the social benefits of the Internet? The economic benefits are clear. The first waive of broadband allowed platforms that were previously unmanageable such as YouTube to become part of our everyday life. A study by the world bank states that an increase in 10% of broadband penetration accounts for an increase in 1.2% of GDP per capital a. Social benefits are harder to measure especially when they depend on future applications that will only materialise when faster reliable and symmetrical download and upload services are widely available and accessible.
The benefits brought by an international, universal access to high speed broadband from both an economical and social perspective are hard to ignore. For example, a 10% increase in employees networking in Australia could save 120 million litres of fuel per year and reduce traffic at peak period by 5 percent.
In this brief presentation, I have discussed the importance of strong Government policy towards unlocking the social benefits of the Internet. This involves a combination of policy, infrastructure, digital literacy and a long term vision.
Projects such as the NBN and Google fiber show the balance that should be achieved between digital citizenship as a basic right and the economic value of information. The slightest extended lightning talk is based on an open access article that is available, if you Google the key words, "Dias first Monday" that are on the screen. Thanks for listening and I appreciate any us suggestions, comments and feedback. Thank you.
(Applause)
CHAIR: Paul, did you want to say a word or two? We have a question at the microphone. Go ahead.
AUDIENCE SPEAKER: Hi, Marcus, you guess by my accent where I'm from. My name is Mathew, I am here representing Amazon but obviously I'm Australian. What are your feelings about ?? given the election is coming up later on this year in Australia and the alternate party has a very different view on the value and how you account for the costs of building an FTTP, what do you think are the risks for Australia on having a policy that is not certain, because one of the big problems Australia has and a number of other countries have in my experience is a lack of certainty around long term outcomes. Do you think this is a bit of a problem?
MARCO DIAS: I think unfortunately it's going to come down to a political decision where the opposition wants to use FTTN, and take a gamble on the last mile. The technical studies talk about trying to transform FTTN into FTTP at a later stage, and more than likely it would cost even more. So, I think obviously the NBN is a bit of a gamble, and there are issues in implementation. It's not an easy project to achieve. Some people say it's utopian, but I think it's a gamble in the future of the Internet, and if you took at other countries, such as south Korea, Japan, etc., that are at the forefront, I think FTTP allows you to implement a solution that's long term, and easily upgradeable without having to go the extra length, and one of the issues as I mentioned, is the fact that the coalition seems to think that Internet is just downloading movies, and doing basic stuff, but in my opinion, they are not looking at the long?term picture. So I think it's a gamble that's worth it and I think it's important that the Government gets involved, although a lot of people fear that Government involvement is intrusive, but I think in this case it's worthwhile.
AUDIENCE SPEAKER: I'd absolutely agree. I think the important word is ubiquity of the solution, versus the FTTN approaches is not.
CHAIR: Thank you very much. All right...
(Applause)
So since we just had a talk that came to us from the RACI initiative, Paul Rendek has asked to say a few words about it in general.
PAUL RENDEK: I am Paul Rendek from the RIPE NCC. Thank you, Marcos, for that. You are the first of our lightning talk on behalf of the RACI initiative. I just wanted to say I am soliciting feedback from the community on what you think about this. I think it's great that we have three five students here that have come. We have reached out to the academic community, the technical, the youth actually. We do this not only just in the technical community, bringing them here to RIPE, but we have made some of these connections through some of the work that we do in Internet governance because it was easier for to us reach out and find these youth because they are programme that are out there that are accessible for us. We wanted to make sure we brought the youth of the technical community here. I am happy with this. I hope that you are as well and you think it's a good initiative from the RIPE NCC's part. Thank you very much, Marcos, for coming here and taking up that opportunity. Thank you.
CHAIR: Before we finish here. I'd like to remind everyone that we really need your feedback. Please go to the website. Please rate the lightning talks.
And it's also possible to rate the tutorials. If you went to a tutorial this morning. It wasn't possible but you can do that now. So please also rate the tutorials. Again it's important to us in the programming committee to know what you think. Otherwise we are just getting.
And finally, Rob Blokzijl would like to say a word or two and then we are done.
ROB BLOKZIJL: Don't be afraid ?? this will be extremely short, even shorter than a lightning talk.
1989, we had the first RIPE Meeting. Three years later, we still had RIPE meetings, they had grown in frequency and groan in number of attendance and we came to the conclusion that we needed the sort of a secretariat because volunteers are a great resource, but I don't always have the time they would like to pledge. So a couple of us wrote a proposal to set up a secretariat, we called it the RIPE NCC, the network coordination centre and we the we had covered aspect to we proposed to hire three people to run this NCC. If you hire people, you need to be part of a legal entity, at least in the Netherlands, so we were looking around and we found a friendly legal entity, the European association of academic and research networks, TERENA, so the RIPE NCC became and activity of TERENA as part of TERENA, the RIPE NCC staff was employed by TERENA. That went on for a couple of years, quite happily, where both organisations were growing in size and growing in distance from each other in their activities, so it became more and more an unnatural situation, so somewhere in the mid?nineties, we decided that the RIPE NCC should become a legal entity on its own and that turned into what you see today. You can imagine that to disen attending he will these two organisations from the under the same umbrella was a complicated administrative, financial and legal operation, and this was mainly driven by the first managing director of the RIPE NCC, Daniel Karrenberg, who should be giving this speech right now, were it not for the fact that tomorrow he has some medical stuff to deal with in the Netherlands, he will be coming here on Wednesday.
So that was the one party. The other party was Carl [Fich]e. Secretary general of TERENA and it is with great pleasure now that I can show my one slide, because today Carl, Fiche for all thinks work for many, many years for the networking community in Europe, more specifically the academic and research community, was decorated by the king of the Netherlands ?? we have a king now since a week or two ?? and he got the official title of officer in the order of orange and NAS awe, if you want to understand better what that means, look at me, I got one two years ago. So, we are taking over this Royal Order. And I think I can ?? yes...
(Applause)
Congratulations, Carl, and I think I can speak on behalf of all of you when we request the current managing director of the RIPE NCC A he will Pawlik to send a note of congratulation once behalf of all of us to Carl. If you want to do it personally, go to the TERENA website, TERENA dot org and there you will find a link to a dedicated page, it tells a little bit more about Carl, the history of TERENA and there is a little box where you can drop a /TPOET for Carl. Thank you. And now I hope there is nothing more between me and the drinks.
CHAIR: Thank you.
(Applause)
Okay. So, we have got two BoFs scheduled for tonight. It's certainly not required to attend them. One will be in this room, that will be Jan /SORZ's BoF an I think it's next door the RIPE Stat's live demo. That starts in 20 minutes. And then there is a social tonight at $7, and we start here tomorrow morning bright and early at $9. Thank you everyone.
(Applause)
