RIPE 69 London
3 November 2014

Plenary Session, 2 p.m..

HANS PETTER HOLEN: Welcome everybody to this 69th RIPE Meeting here in London. It's good to see so many of you here. We're actually 665 registered for this meeting, 204 of you are newcomers, that's really impressive, 410 have checked in so far, and 122 of those are newcomers. So, welcome everybody.

I need to do two things at once, I need to press a button and talk.

So this is a RIPE Meeting. It's not a conference, it's a meeting, it's for all of you to meet and discuss topics of interest to you. So, we want to keep this meeting open to everyone. We want to bring people together from different backgrounds, culture, nationalities, beliefs and genders. And the RIPE community is supposed to be, and kept to be a safe, supportive and respectful environment. So that's one very important thing to realise for all of you. And if you, for some reason, feel that it's not so, we have some trusted contacts, I don't know if Nick and Mirjam is here ‑‑ they are both here ‑‑ so, if you need somebody to talk to, if you experience something that you don't think is good, go and talk to them. We will keep that confidential and they will help you the best they can.

So, we have a pretty full meeting plan this week. Most of this is labelled Plenary and we have a Programme Committee that's responsible for the content there and Filiz will come on stage and talk about that later. If you look at Wednesday and Thursday slots, you have all the Working Groups. So that's where you actually work and contribute. There is one slot at the end of Tuesday, which is labelled RACI, RIPE academic ‑‑ something ‑‑ initiative. It's kind of an initiative by the RIPE NCC to get fresh blood into this community. I mean, we realise that too many of us are getting grey hairs and can't remember things properly any more, so we need to get new people in here and they are getting sponsorships to come here and there is a time slot here where they will present their stuff and they also have the ability to submit their presentations to the Programme Committee, but there they will be rated and selected as any other presentation.

So... there is one thing on this agenda that I would like to bring your attention to and that's the Cooperation Working Group. This time in the Cooperation Working Group, there is a big change in what's surrounding our community. The US Government currently has a contract with ICANN for operating the IANA function. The IANA function is the registry that keeps track on which of the RIRs, where RIPE NCC is one of them which uses which IP addresses and ASN numbers and the US Government has said that it is considering to step out of its role as holding that contract and has tasked ICANN to make a plan to transition that into something else. Now our community, we, together with the five RIRs, plan to submit a plan to do that. So that's going to be discussed in the Cooperation Working Group. So if you are interested in how we are going to have the framework around the RIRs and ICANN in the future, or IANA in the future, you need to be present there.

What the RIRs have agreed is to put down, put together an editorial team of two community members and one staff from each region, so if you are interested to serve on that team, you need to be in the Cooperation Working Group, put your name forward before that, we'll discuss in that Working Group, and then afterwards, I, as Chair, will make a selection based on the discussion there and it will be confirmed in the Plenary on Friday. So, that group needs to work hard to deliver a proposal so the community can discuss it before mid‑December. So it's a pretty tight schedule, and we also need to do this in all five regions, that's why we are putting together this editorial team to do that work.

So, those guys up there are all the Working Group Chairs, so if you are interested in any of these Working Groups or have any questions about them, contact these guys. Two of them on the far left side, Marco and Shane in the IPv6 Working Group are actually stepping down so in the IPv6 Working Group there will be a selection of new Working Group Chairs.

And then I think that's all I had planned to say. So then I'm going to introduce the host of this meeting, John Souter from LINX, the stage is yours.

JOHN SOUTER: Thank you. Welcome everybody. I feel slightly more nervous standing in front of you than I normally do on these occasions and you'll see why in a minute. So, good afternoon. I'm sorry about the weather, we couldn't arrange for anything better. The trip from the Tube station was not a particularly pleasant experience. But, it's really nice to see so many people here and nice to see the attendance of the event peaking in the way they are. I have just come back in the NANOG meeting where there were just about 800 people, and these community meetings do seem to be picking up at the moment.

I guess the reason I'm here and the reason I have the honour of being the host is that it's a special year for LINX. It's our 20th anniversary year and this is actually the week of our 20th birthday. It's a special point in our history and so it makes us particularly proud to be able to assist RIPE in this way I honestly never thought there'd be a RIPE meeting in London. It's an incredibly complicated thing to arrange and finding the right kind of hotels and the right kind of prices is not easy. It's ten years since RIPE has been in England, or in the UK, perhaps many of you can remember, I think it was was it RIPE 49 in 2004? Hands up, those who were there? I thought there were a few. It was ‑‑ you can work out the maths, it was also our 10th anniversary year that year and we held a fairly spectacular party at that event. I don't remember too much about it to be honest with you, I can remember the hangover, I guess of those you who were probably can sympathise with me. Another thing that's significant this particular week is, it's a very British celebration, which we call Guy Fawkes night, which celebrates an event 400 years ago, that was made to blow up our Houses of Parliament. It didn't, unfortunately, and we have been suffering Government malpratice ever since. The fireworks lead me on to something I particularly wanted to plug, and that is that we also have the honour of having a social evening tomorrow night, specifically to celebrate our 20th birthday. So, you all got an invitation in your packs and I hope you all come to that. There will be lots going on there.

We also need to remember that 2014 is significant in RIPE's history too. It's RIPE's 25th anniversary, and over the last quarter of a century, things have changed a lot and RIPE has had a huge influence on that. I guess we can say with some confidence that the Internet is now pretty central to everything that goes on in the world and 25 years ago I don't think anyone could have quite conceived us making such a statement. It might have even seemed like science fiction.

We seem to have company... you know who you are ‑‑

Hello Mr. Dalek. I guess you know about LINX. Do you? Well, I was going to give you a long story about LINX, but I guess, I am guessing from your silence that you are not too impressed?

MR. DALEK: We are beyond the reach of your meagre Internet world. It is unclean and will be cleansed along with the rest of humanity. We will prepare. We will grow stronger when the time is right we will emerge and take our rightful place as the supreme power of the universe.

JOHN SOUTER: The trouble with Daleks is, they are all talk. We'll probably die of boredem while he gets round to shooting me. By the way, I have seen your search history and it's nothing to be proud of, is it? You have trying to make the Internet better for more than 25 years. We stepped up. But where were the Daleks in this?

DR. DALEK: Daleks are incompatible with steps.

SPEAKER: Excuses, excuses. Let's forget about that seeing as someone stole the joke anyway. How are technology seems to be incompatible. I guess together we could do significant things for the good of the Internet? What do you think?

MR. DALEK: An alliance with humans? Request denied. All hostile elements will be eliminated.

SPEAKER: I have to warn you, Daleks, that you seem to be declaring war on us. There are a lot more of us than there are of you and we know how to fight. How many are there of you?

MR. DALEK: Numbers are irrelevant. We can destroy the Internet with a single command. Initialise single stakeholder top‑down governance node. The others will help me after they finish collecting samples of beer and wine from the bar. We will acknowledge one thing humans, you are better than the Daleks at organising social events for network engineers.

SPEAKER: That's a small consolation, I guess. Daleks seem to get the best jokes as well. So I better shut up. I was going to give you a long list of achievements and things that we had done that perhaps compared with Daleks, but I'm getting you are not really too interested in hearing that?

MR. DALEK: Exterminate. Exterminate. Have you finished? I was about to go into sleep mode? Look, can't I just wish you happy birthday and be done with it?

SPEAKER: Yeah, I guess you can do that as long as you don't sing. It must have been a much better gag if we could get it on the stage, but as you see the steps...

So, thank you, let's give a round of applause for the Dalek.

I'll just say it's a great honour to be asked to open the event so let's declare us truly open. Please come to our ‑‑ this is unscripted nought. As you might have guessed ‑‑ so, please come to our social event tomorrow night. I'd like to acknowledge the ‑‑ the Dalek was actually made by a Bedfordshire girls' school and if you want to acknowledge your appreciation of that, then you can tweet to the hash tag BGS or BGS Dalek, so please feel free to do that if you enjoyed that. Thanks everyone.


HANS PETTER HOLEN: Thank you very much John. Before going on with the programme, I have a couple of things to add. Of course there are some socials at this meeting. It's a social event today at 18:30, where you can meet the RIPE NCC executive board, so if you are interested in talking to them, that's where you can meet them. And there are welcome drinks at 7 o'clock tonight.

On Tuesday, there is a newcomers' welcome reception for newcomers only at 6:30, and then there is a party at Jewel Piccadilly and there are buses going back and forth for that.

On Wednesday, you actually have to do the work yourself so there is no arranged social, I mean we had complaints in the past that there were too many socials so you wanted to have the ability to actually arrange something for yourselves. So we have accommodated that on the Wednesday and on the Thursday, there is the RIPE 69 dinner. So, there will also be buses back and forth then.

And mentioning the people that are important to this community, I have showed you all the Working Group Chairs and Filiz will shortly show you all the Programme Committee. There is one special person here that we have to thank you for being here and that's our former Chair, Rob Blokzijl, he is sitting here.

So, with that, ill give the Chair to the Programme Committee and Filiz...

FILIZ YILMAZ: Hello everyone. Welcome this this fine RIPE Meeting. My name is Filiz Yilmaz, I Chair the RIPE Programme Committee and I will give you a short update about our activities, and ‑‑ but before that, I want to thank to all the sponsors for having this meeting, especially to LINX enabling us to be here in London. Thank you very much.

So, let's get on with what we do.

RIPE Programme Committee puts together the agenda and the content of the RIPE Meeting partially. As Hans Petter Holen already mentioned, that partially means Monday, Tuesday and a bit of Friday for the Programme Committee. The rest of it is done by the Working Group Chairs, and that is a group that is completely differently organised than us. We are the one people, I'll show you our faces in a minute, but the part of the week that we are responsible for in terms of content are those ones that are marked with a red frame on the screen. And one thing I want to mention here, speakers do not get any special treatment. RIPE meetings are totally a hundred percent volunteer based meetings. Not conferences. And I am actually quite proud of that asset, or the trait of RIPE meetings. So speakers are part of the audience who choose to submit some of their work, ideas and present to the rest of the community.

We talked about a little bit of the meeting week how it goes. But, we often receive the question, cheats the difference then? RIPE Programme Committee content or Plenary content versus Working Group. So I want to take a bit of a moment here to talk about that. We clearly are network community. We care about similar topics. DNS, routing, sometimes open resource, and programme committees strive to bring that kind of ‑‑ those kinds of subjects towards the general audience interest, towards the ‑‑ where it can speak to anybody maybe not too much in depth, but also just to give them a glimpse of what is happening around the world.

So the content is a mixed bag of operational policy and sometimes at times philosophical content, I would say. The format of the programme also reflects to this approach. We have long talks, and when I say long, for the primary time of the Monday and Tuesday, long means 30 minutes. And this is one of the main differences between Working Group part of the week to the Plenary part, because in Working Groups, you deep dive into certain subjects. We don't do that. We give you 20 minutes and 10 minutes we ask for the audience to ask questions. And in that 30 minutes you need to catch the general audience's interest. So the format is different there. And some of our content are also tutorials and workshops and they can be long, about an hour, and sometimes we give longer for the tutorials, but they are not placed in the prime time of the meeting. So not the prime time of the Monday and Tuesday. They either occur in the morning on Monday mornings or in the evenings and we have BoFs for informal gatherings and we also have lightning talks and as the name suggests we really want to catch those hot topics, crazy ideas sometimes. New ideas that you may be thinking of that you thought like, dreamt of yesterday. And you think that this audience might be, you know, interested in that. And they are very short. So, that is one thing to keep in mind.

Now, who does that? Well those 11 people at the moment are doing that and these are the faces of them. We will have two of them approaching the end of their term this Friday, so there will be two seats for elections. If you are interested to join in this group, please do nominate. You can send your bio as a nomination to, and you have until 3:00pm on Thursday to show your interest. These links on the slides and these slides are going to be of course on the meeting site. We'll give you a glimpse of what is expected of you but I will take this moment to show you what is expected.

When I say ‑‑ this is a good bunch of people. We work really hard but we want to mention that, there are some expectations when you join the group. Among which our main job is to recruit talks, clearly, because that's kind of, you know, makes the programme, what you receive here, what you see is the result of our selection or evaluation process. So that contains reading and rating and commenting and discussing talks, and that can be 40 to 70 talks I'm talking about here, and then there is the shepherding activities, and attending conferences which are often in the central European time. And keeping in touch with the PC, you know, on ‑‑ we do a lot of work on an online basis, so, you just don't join one conference call and then disappear; you stay in touch with the rest of the group.

So what do you get in return? Those are the benefits side, which is a short list you can see on the other side. You get satisfaction for serving the community. And a bit of industry and peer visibility. And you often need to remind yourself, a repeat process, one to three, this is what you get as of benefits because there is not much more than that.

Now, let's talk about this week, the Monday, Tuesday, and Friday content, as I mentioned.

First of all, I want to thank you all. This time we received lot‑ish submissions, quite a lot compared to the previous meeting, so, great numbers. And one other thing is that they came on time. So double thanks. And, again, I'll ‑‑ the majority of them came in on time after ‑‑ by the first deadline we announced. And I want to take another moment to talk about this. You may have noticed we often make a call, and that deadline often falls just a little lower than the 8 weeks prior to the RIPE Meeting. This ‑‑ we aim to have some content already published before the eight weeks coming ‑‑ approaching towards the RIPE Meeting and there is a reason for that. Normally, we can accommodate 16 slots, and if your talk is really good and you were on time following that first deadline, then you get a slot out of that 16 slots. And if your talk is really still very good but you are a last‑minuter, you risk to get that slot. So what we advise is help us be on time, because that eight weeks time is a good ample time for the RIPE participants. Sometimes people need to go and see the agenda before they can ask for support to come to this meeting and the more people come, as you can imagine, a better meeting it becomes. Because it gives some ample time for the PC to shepherd the talks, so, it can a talk that may need some work and that will give us time to work with the proposers, submitter, so that you get a better content as well. So it's a win‑win situation, and I hope it works, and then the final agenda is published about four weeks before the meeting starts. So, if you have trouble with this and you don't like, please talk to us, we have been keeping asking for feedback about that, and we thought it was working, but if it is not for you, let us know.

One final thing. This RIPE Meeting we have more lightning talks. There was a talk that was withdrawn in the last minute. This means there is still space in the agenda for lightning talks. If you have a crazy, funky idea, like I said, that might be proactive or that might be fun, please, you know, if you have some spare time during this week, put slides together and submit them to us. On Friday you may have your lightning talk presented.

One final slide here.
We do like to have the ‑‑ receive feedback, so please talk to us, and I don't know if this is the case still, but I hope I'm not putting RIPE NCC on a difficult spot here, but if you talk ‑‑ rate the talks, you will get a prize, I believe, again, and we will be all around here the whole week, talk to us, and have a very good meeting. And I will leave the stage to our next speaker, William Waites on community networking in Scotland. Thank you.


WILLIAM WAITES: Hello. This is my first RIPE Meeting. Thank you very much for having me. I'm from Hubs, CIC, which is a community interest company, a kind of not‑for‑profit that we have in Scotland, which is made up of very small network providers, I'll get to more of that in a moment. I'm also at the School of Informatics and my main job is working on numerical information on cells and these sorts of things. And it was in fact through this and because there was some ‑‑ I have been a network engineer also for a long time, and it was sort of inevitable at the school of Informatics that I got involved in some of these projects up north.

There is just the setting. There is Europe. We all know about Europe. There is Scotland up on the edge of it. And we're not so concerned where the story begins. With all of Scotland but actually some small region of it there sort of on the west coast where there are a bunch of islands you can see ‑‑ ‑‑ you have Skye here and we have the small isles, Eigg and Rhum, and we have this peninsula and actually where this really begins is these two places here, Arnisdale and Corran. To give you an idea of what this part of the world looks like. This is ‑‑ these are small places, they are quite remote. It takes probably about four and a half or five hours to get there from Edinburgh, the last hour or so on really very windy single‑track roads. And there is in the picture that we see here, there is may be about 50, 60 house holds or so. And back in the mid‑sort of 2000s, one of the professors at the university bought a house there and moved in, started renovating it and tried to get some sort of Internet access. And it's far away from the telephone exchange, about 15 kilometres. Even though, in principle, you could get, oh, maybe 2400 or 4800 bandwidth of modem over it in practice, the copper is in such bad shape that you can't do that. Glenelg, where the exchange is, just where the exchange is, has a couple of E1s, maybe there's somebody from BT that can correct me, but from I have seen staying in the hotel there I'd be very much surprised if it's more than that. There have been suggestions that maybe well people living in places like this should use satellites. This turns out not to work very well. Not only is it slow and expensive, but [Lavrain] ‑‑ my apologies to the stengographer for gaelic placenames ‑‑ is the mountain that's across the loch there and since we're fairly far north and the satellites are fairly low in the sky, satellites don't work very well. And so the people who live there spent a lot of time writing letters and complaining and got eventually, we got nowhere and eventually decided well, maybe they could lib rate a little bit of research funding at the university and build stuff themselves. This, incidentally, is one of the first masts that was built and this is a very bad design for a mast. Itings likes like the wind turbine is part of the mast, it isn't. But even so, when you do the obvious building things pointing straight up in an exposed place on the side of a hill by the Atlantic ocean, in a gale it tends to twist around, which is not very good for the alignment of antennas. So ‑‑ this is just more about ‑‑ I did a little calculation of how copper lines behave at this kind of a distance and, well, it's not good.

So, the first iteration of this began with a network that came to be called Tegola. This was with Peter Buneman who was the professor that I mentioned there, amino who might be around here somewhere, who is also ‑‑ I see them in the back ‑‑ he is giving a talk on sort defined networks, or some such, tomorrow, I think ‑‑ was at the time a grad student, I think, at the university and helped to build this, along with Mahesh Marina and they put a trial link with some 802.11A kits, about 15km over water from Corran and shared an ADSL line and said okay this worked. This expanded to six masts going around Loch Hourn, that's the name of the body of water that's there. This was all built using gate works industrial PCs and open WRT and running Quagga and OSPF and all this was great. They managed to make friends with the people at the University of Highlands and Islands just across the way where they have a college there, and there was reasonable significant bandwidth at the college, something like 45 or 50 megabits per second or so, which they agreed to share with the people across the way. And so that was good and that worked very well.

There's a picture of a more recent generation of mast construction which works a little bit better in this kind of environment. It's more solid, it's just made of scaffolding and clamps. This one is epoxied into the rock but some of there are stuck in there and it's quite good and stable and cheap to build.

And so the neighbours, the neighbours on these nearby islands, Eigg and Rhum, they looked at this and said this is good and we have had people suggesting that maybe we should get satellites and continue didn't seem to work very well. So we're going to try to do the same thing. And so they did. And you know with a little bit of advice and with some slightly newer more off‑the‑shelf kind of kit and then it spread in the, you know, in the next loch, in Loch Nevis, people built theirs with some help from Eigg and Applecross further north etc., Mull, and Slate and all of these places, and further away in central Scotland, in Laggan, and Lothian near Edinburgh and there are more, there's... and so this is good as far as it goes. Now we have all of these networks and they are all kind of doing their own thing. And they have got basically what amounts to an access network around in their local area, and they are connecting these things to, in some cases DSL lines and in some cases, well the university's network at sell more us a particular and what do you do? I mean they are fairly close together? I mean doubt obvious thing. You ‑‑ I skipped a slide.

So, this slide is ‑‑ points out that these networks are actually quite different. They were built with help and advice from their neighbours, but each person involved in doing this or each small group of people had a very different understanding of how to build a network and what they could manage and what they could do. But these places are, you know, quite ‑‑ it has to be done according to what the local people are able to do and understand and maintain and support on their own. It can't be done in a way that is centrally managed because going and fixing something, especially if it's something that goes and creates a network partition and you can't see it, is, you know, might be six hours, it maybe a day if the weather is bad and you have to take the boat and the boats are not running, it might be longer than that. I put a version to dynamic routing. It's a version as in not understanding it. The same kind of thing that happens when somebody first gets a very simple router and learns about I can make a static route route and understand how that works. Do you dynamic routing, that's complicated. It might be a little bit frightening. But, if a strong desire for having redundant paths because things break. Things break in the meantime to repair is quite long, it can be days or in some cases it can be even more than a week. As people have gotten used to these networks, they have done things like throwing away their land lines and there is no mobile coverage. If you want a resilient network in an area like this, you need some kind of dynamic routing. And well how do you do this?

These networks are built on a shoestring and the equipment that you can get from vendors for shoe string kinds of prices is not very intelligent. It doesn't speak OSPF, it doesn't do any of these sorts of things. Putting open WRT on them, possible in principle but it seems to be too high a barrier, and you know, we would like to run Juniper routers everywhere but they are expensive and too power‑hungry, because we have masts that are running off grids and powered by solar panels and wind turbines, and that won't do.

What do you do in this environment? Now, the slide that I started on there. We connect them. We connect them and ‑‑ so this is basically where hubs comes in and this is about two years ago we came in and saw this and said well, let's connect these networks together, manage the ‑‑ part ‑‑ the part of the networks that is the inter‑connections between these community networks. Don't try to make ubiquity radios do anything fancy; just treat them as transparent bridges. Do the core with embedded PCs, run FreeBSD, and BIRD, sometimes MIkroTik, don't try to dig too far into what the individual networks are doing, because even though I might not agree with how they are built and I might have suggestions about how they could make it better, it's better that they have something that they can maintain.

It turns out this works. And it turns out it works quite well. For example, we had fibre cut a little bit further north that took out a substantial part of the telephone and Internet service on the west coast. This was one of the stories where the fibre cut was somebody seeing this cable and thinking oh, copper, I can sell that, isn't that great? And, you know, and Sky and all these people, no telephone service, but our network stayed up, and that was good and stayed up thanks to the connection at the university there.

Then a couple of months later, we had exactly the same reverse happen. We had a power cut and back up power thing that didn't work at the university, over the Christmas holidays, so it was a good week before, and you know, again, we have this arrangement with small networks providing mutual transit and ‑‑ oh, yes, lightning strike that melted phone lines, that was night because the NHS were using or network to provide the health service.

So that was good. So we tried to structure it a lit bit the way the Internet works with you know networks that are autonomous and they have a strong ab traction barrier between them and you can still have dynamic routing and such.

So this is kind I suppose a logical diagram of what it looked like. About two years ago, in fact almost exactly two years ago, the light blue, like we see here, is meant to mean a sub‑net basically so we see one that this is Eigg where they decided to put everything in one great big spanning tree, well that's what they did. This is ‑‑ this is ‑‑ which is quite remarkable because it was built on the ICE of the people from Eigg and it's not structured the same way. The way it's structured is as brilliant as it is frightening, the person there is he wanted to understand how IP routing works, and so he implemented something like a distance investigator protocol in a spreadsheet, base and out of the spreadsheet came configuration files for ubiquity radios. So we have this and it's all static routed and it works, and up here we have the old, like at the very top we have the old experimental tell toll network which sick masts which remains as a test bed and the newer production one, which is you know, basically the same sort of thing, but slightly different, and here we have some kind of an inter community network connection thing going on and BGP going where there are those red lines, and the back up connections with DSL lines and ‑‑ and that's all very nice. And so we decided to formalise this a little bit and to create an organisation to do this. So we did, and we said well, okay, each community network gets an ASN. In this case it's private ASs but for one reason or another we might move some of them to using public ASs. Retain the diverse internal structures and don't mess with what they want to do too much. Provide for collective negotiation of transit and carrier services and getting IP addresses. Remember, each of these networks are very small; they range from ‑ actually, the smallest is probably about a dozen subscribers and the biggest is about 200, so none of them big enough to become LIRs and get addresses and do all of these sorts of things. They look like a confederation which is not actually visible in the BGP tables because the private ASs are stripped out. There are currently two of such things. One on the west coast and another one that we started for the south of Scotland around Edinburgh and the borders and East Lothian and these sorts of places and there is a bit of a logo parade going on. UHI and the University of Edinburgh and the Carnegie Trust who employed me briefly for about a year to work on this, leader and the European Union, note, you can't see this here but this is agricultural fund who used to fund these sorts of things on places like Eigg and this is now changed, which will be coming up. And Fluency, who is a service provider around here, who I don't know, may be in the room has been extremely helpful.

And this is what we have in the south of Scotland, and in fact there are a couple of ‑‑ there's one or two more of these networks that are missing. Transit from Fluency, we are peering at the exchange in Scotland, the new exchange started with the new LINX exchange there, which is great and it would be wonderful to see more people at the exchange and in particular more content providers, that would be great. So if anybody from, I don't know, Google or Amazon or any of those people with large sources of data were to join IX Scotland, that would be a wonderful thing.

And that's ‑‑ it's a network diagram, we have all seen network diagrams before.

So the impact of this seems to have been fairly good. Distance learning for adults, UHI's interest in this was because they run quite a lot of distance education because of the university in this very sparsely populated mountain use western Scotland, so that was good, it meant people could attend their courses. Children, this is important ‑‑ so there's a national schools Internet but this is particularly important on some of the islands like Eigg where the children miss perhaps three weeks or more of school a year because of bad weather and the ferry is not sailing and there is no high school on the island so they would normally be going to mall egg or something like that to go to high school and they can't because the ferry isn't running, so they can communicate with their teachers and do their homework that sort of thing. Would he have seen local businesses moving into these places, which is something quite new. The one that I'm thinking of, I think, makes wind turbines and ‑‑ or sets up wind turbines or something like this, but in any case they would not have moved there if they did not have decent Internet access, and we started getting tele medicine and we have noticed that symmetry in network access is appreciated. I actually wrote that on the train as I was sitting at 5:30 in the morning at Edinburgh this morning, and I was trying to upload a slightly modified version of this presentation, and I found that even sitting there, not even moving, sitting in the train at Edinburgh, trying to send, I don't know, like 10 megabits or something of file, it was terrible, but you know, tourism is a thing that happens a lot in these places and it's a big part of economy and tourists want to go there and take their pictures and they want to be able to upload them, and I suppose whoever decided to build a symmetry into the access networks at a fundamental level never anticipated this.

But, it's nice, because at least we have symmetric access up there. So, okay, these are ‑‑ there are two slides that might be a little bit ‑‑ these are the problems that we are seeing actually. Everybody who has ‑‑ we have ‑‑ there is still a lot of NAT in these networks. It's not going to go away because of IPv4 address scarcity, it breaks things like pick owe cells that people like to use because there are no mobile converge. This also causes problems with some big service providers, Google is one that we have seen where they say, well, okay we see all sorts of traffic coming from this address and we're going to do some kind of collective punishment on people there and say well, you have to enter captures or you are not allowed to use this. And this is in fact a very similar phenomenon to what we see with anonymity networks like Tour, and whereas the solution for us actually is to push IPv6 and that's very easy, this is not the solution for something like tour that has exactly the same sorts of things happen. We have seen RFC 1918 address collisions because the which we are connected into the university network up there, we can see their internal address space and they are running weird DNS thing and so there are a few ugly things in there that need to be gotten rid of. Adoption of IPv6 is slow for exactly the reason that adoption of dynamic routing is slow. Is it seems difficult and complicated and there is a running thread in here in that edcation is something that's needed, and it's education of not your typical network engineers or a person who is doing this because they want to, but it's education of joiners and builders who are the people who tend to actually be constructing these things, and need to learn about IP routing.

And we're outgrowing the capacity on the west coast. This is not just our capacity and what's dedicated to us ‑‑ not just the capacity that is available to us, but the capacity that is available which is basically limited to a network put in by what was then Scottish Telecom or THUS and became Cable & Wireless and then Vodafone about 15 years ago, and hasn't been expanded since.

It's not immediately clear what the answer to that is. What the answer to that should be kind of comes up on the next slide, which is these are the things that are happening at the higher layers, but first I want to point out the ‑‑ this was ‑‑ I had an idea that maybe we could use POV‑Ray for doing RF site of line planning and turns out you can and it makes pretty pictures. So, recently, about two years ago, about the same time we were starting this, there was a UK and Scottish Government project to fund a massive infrastructure build in the whole place, about 500 million pounds for Scotland, and some of this was put aside for the places that BT, who got the contract, would not be able to reach, and we said this is going to be great, okay, so you know, BT is going to come in and put in a backbone network and where it's convenient for us we can going to eat net access, no problem. But what happened is, by creating this small funding agency for community networks with a very cumbersome process and staffed by nobody with any technical background at all, it meant that it, all of a sudden, became impossible to get funding to build these things, or not impossible but you have to hire a consultant to make a complete plan which then has to go out for tender and all of these sorts of things basically taking it entirely out of the hands of the communities. There is a preoccupation with state he'd which are EU rules invented by the UK to prevent distortion of markets. And double funding. Which is related to the territorial "disputes" because they are not really disputes but they're places like here where there is a tiny community here and it's just on the edge of being viable for building its own network, and it perfectly viable if it was able to serve people in the blue area and there is no technical reason why they can't and it's easy but they are not allowed to because the rules are such that it's claimed by BT and if they were to serve there, then the Government would say, well, we funded this area twice and that's not fair.

We also do not know in this big infrastructure build where we will be able to get access to point‑to‑point layer 2 circuits that we need to connect these networks to elsewhere, and that's very frustrating.

There is also education which was touched on in previous slides, about how to do West Highland network engineering. West Highland engineering is a phrase that's meant, to one example, you know, is the man who knows how to repair a net board motor on a boat using only a rock and a rusty screwdriver, because you have to make do with the things that you have there.

And in fact, it would be ‑‑ I would very much like to hear from anybody who is involved in doing networking particularly in strange or weird or hostile or non standard kind of environments, because I'd like to, given that we have niece good connections with the universities, I'd like to organise maybe a week long or so summer school, maybe up on sky, maybe on Edinburgh, next summer to bring the people who want to learn about doing these things together with people who know about doing these things to actually get serious about education to improve ‑‑ well to do what we can to improve the parts of it that we can do.

And this slide, I don't know that we have ‑‑ we have got five minutes, which is just enough time for questions. But we have ‑‑ the only interesting thing here is ‑‑ well tied al fading, I can talk you about that, anybody who wants to know, elsewhere. But we set up this experiment with, off‑the‑shelf kit, and used in a slightly non standard way, sent it off to know art where they found a piece of drift wood to attach it to and it works beautifully to mitigate tied al fading and doing other things with 24 gigahertz and free space optics, I am perfectly happy to talk about. But, yes, I think time for ‑‑ thank you ‑‑ website is there. And any questions?


CHAIR: Are there any questions?

AUDIENCE SPEAKER: Brian Nesbitt. I wanted to sort of ask if you were aware of the guy a global access to the Internet for all piece of work which is now started up in the IETF and IRT F and Internet societies involved in it and it's exactly this kind of thing, it is looking at initiatives and ways of getting Internet access to remote locations globally and I think if you are not aware of it, there is a mailing list and there could be some interesting stuff there and to take a look at and some useful knowledge sharing as well.

WILLIAM WAITES: Yes, I am aware of that. The difficulty is partly I'm not ‑‑ I'm aware of, it I occasionally get messages forwarded from the guy a mailing list. I don't subscribe to it, purely due to mailing list fatigue. The difficulty is getting the people running the small community networks to subscribe to these things and to participate, because they are working a little bit in isolation which is one of the reasons why in fact the idea of the summer school is a little bit less to have ‑‑ well, it is for, you know, explicitly for education but more to try to create those links and get those people talking together.

AUDIENCE SPEAKER: Hi, it's not a question. I'm Vesna and I'm really happy that this topic is now at the RIPE Meeting and that it is opening this RIPE Meeting so thanks a lot for that. I'd like to help you to connect those communities and I have a lot of experience with similar networks in weird locations in ex‑Yugoslavia and I think it's appropriate your research for those areas. So thanks.

WILLIAM WAITES: Wonderful. Thanks very much. We'll talk afterwards.

CHAIR: Thank you very much William. So, I am Jan, I will take care of the housekeeping for the rest of this session. Just a note to all the speakers that are following, when we show five minutes, that shows five minutes including questions. So, when you see the five‑minutes mark, you should probably think about questions that are following.

Our next speaker is Michele, she is coming out of Africa and she will talk about iGDP in Africa.

MICHELE McCANN: Hello everybody. So, I just wanted to, before I go into the details around the iGDP stuff, I wanted to give you a bit of an example that's happened in Africa and it's pretty much because of this kind of community. I have been coming to the RIPE meetings now for two years, so no longer a 'firsty'. And with the leadership and guidance that you guys have actually provided us a tangible examples of savings that have now been achieved in Africa around a particular exchange point that we look after is now equating to 1.4 million dollars, a monthly saving in terms of local transit. Just a key point to say thank you very much.

From the outside, you look at Africa and say what a great investment? Lots of people, huge penetration in terms of, which is actually growing and doubling every single year on Internet access but the big question is: Why is it still hard to invest in Africa? We have got a lot of fibre resources, we have got colocation facilities, we have got exchanges so the big question is why is it still happening in Africa? The key thing which, personally, I'm quite actively involved in, is inter‑connection and the inter‑connection issues that we actually have in Africa. So, I have put some examples here and one of them, funny enough, is Swaziland, and from my home town, it's pretty much like 200 kilometres away from where I live, and, in South Africa, we kind of have open connection policies where you are paying about 50 dollars per interconnect but if you have a look at Swaziland, they sort of last point‑to‑point connections can go up to 5,400 dollars per month for a basic connection. Just an interesting idea there. Nigeria, which a lot of you do tend to ask us about that, a huge population, lots of money, those kinds of things. The big issue there is that MTN controls the landing station in Nigeria, so you can get there by wax but try and get out of the landing station and you are going to look at paying over 1,000 dollars for just a really small cross‑connect.

Djibouti itself, the landing station, to give you an idea, is pretty much less than a kilometre away from the the colocation environment and you are looking at 10,000 dollars per one gigabit of speed just to get from the landing station to the co‑lo facility. So just to kind of illustrate some of the very interesting issues that we face.

Yes, it is a growing market. Yes, there are active ISPs. Just a word of warning. So there's a lot of marketed data centres that say they are neutral. Ask them what are the cost of the inter‑connection before you side any co‑lo deals around that one. You'll see kind of a lot of ASN requests are starting to happen. So myself personally, I work very closely with the AfriNIC guys and we are starting to get a lot more enterprises, interestingly enough, starting to apply for their own ASN because they have the ability to now choose between multiple carriers.

So, what I wanted to give today was a view on some of the key hubs and what's actually happened to kind of open up those ones. So, South Africa, that's where I'm from, the key thing that happened in South Africa is the deregulation, and, if it wasn't for this, obviously [Terico] wouldn't exist. Before we had seven infrastructure operations, now we are currently sitting on over 30 infrastructure operators, as well as regional exchange sitting with over 100 members, lots of content investment, but what we are actually seeing now is more of a consolidation happening from a South Africa perspective.

So you'll see here this is an example of inter‑connection growth in kind of an open market space. And out of one of the co‑lo facilities. What happened in Kenya, even though it was deregulated in 1999, there wasn't much ASN applications or people going out to try and launch their own ISPs etc.. the major thing that happened there that made the change was C‑COM and C‑COM landing, so you'll see I put on AfriNIC's website they show the time frame and when ASN requests started to grow. So that was typically when that happened.

There is a neutral data centre there and that has now created a lot of competition that essentially happens within Kenya. So you have seen Kenyan prices drop, South African prices generally tend to drop and content investment is starting to happen there as well.

Nigeria, our wild west partners, so they themselves, the major thing there was, it was deregulated in 2011, GLO‑1 landed and that was when you started to see the difference happen in Nigeria. There is an operational IXP running about 2 gigabits of traffic and 38 members. No neutral data centre and as I have mentioned to you before, is there's access to the cable landing station is very limited. And content investment is very limited at the moment, so there's a lot of talk and I hear a lot of you talking about investing in Nigeria, but we generally start seeing people rather going to Ghana than Nigeria purely because of the open market space that Ghana generally tends to offer.

So, iGDP itself in terms of the African market, we are seeing a three times growth, which is fantastic from ‑‑ in the open kind of areas. In some of the other areas you generally don't see much growth like your DRC, those kinds of areas, but what are the key things that need to happen? Once again, mentioned, open unrestricted market places need to happen, cross board assets, so, companies like C‑COM, CM C networks have invested a huge amount in terms of cross‑bordering the various countries so those assets need to start to be utilised within Africa. And Government itself, my favourite friends, they need to play a supportive role and not generally the policy making kind of role that they are trying to do at the moment.

And then obviously establish an open IXPs, the one thing that we're doing quite a bit of work with is the access project and that's generally where the ISOC guys have gotten involved to try and say guys, let's put down key exchanges and let's make them sustainable because we previously did have a lot of exchanges around Africa and it's easy to put down the Juniper box, make it work, but trying to overcome a lot of the monopolies that actually still exist in Africa.

This is actually an interesting slide and funny enough what's happened is if you see in the news South Africa has a power crisis, ironically even though we're the best in terms of power, today one of the escom salas [phonetic] collapsed, believe it or not, to we have lost half of our powers resources as of today but you actually see the direct impact in terms of your sales and revenue numbers when you have a lack of infrastructure, and infrastructure availability. And one of the things we ‑‑ I'm also give you another example. We looked at building a co‑lo data centre in Nigeria and the issue on why we didn't do it was not because of Ebola, but we decided that the lack of power within Nigeria, we would literally have to shut up a generator once a month to keep that data centre going, so for us it was absolutely not sustainable to invest in that kind of country. So, it's quite interesting to see the comparison and the lack of infrastructure and what impact that has in terms of revenue numbers.

So, in closing, and I don't want to keep you guys away from your coffee, so, in closing, we have the users, we have access to technology, so, pretty much most of our users do use smartphones, so when you see the adverts and someone sitting in a township using a really broken old phone it's actually not true, most of our users do use smartphones and most of our users do have access to Internet. Which has been something that's changed in the last pretty much two years where we didn't really have much access to Internet, and I'll give you my last example on that one, is there's a school for up and coming pupils and people who want to learn about the Internet and start learning how to develop Apps etc. Based in Cape Town and one of the guys, the sew actually runs, it he went and he showed me all different kinds of rooms, there is an app room a boardroom. They teach guys who don't have a chance to have these kinds of facilities. And interestingly enough the boardroom actually smells bike a boardroom and they do this on purpose because the kids are too scared to go into a boardroom and he told me a story about a gentleman who walked all the way from Zimbabwe all the way to Cape Town top able to have this opportunity ‑‑ and he wants to be an app developer, as an example ‑‑ and he literally walked all the way to Cape Town, and that was two years ago and within two years now this gentleman, be you wherever you need to be, you can now log on and be able to get access to these tutorials, which is fantastic news, it's an amazing change over a short period of time.

So, we do have the smartphones. And we do have, I like the big dollar sign there, yes, there is an opportunity to make money in Africa. However, we still need to work together as a community, and I thank this community for all the work they have done in Africa, because it definitely is noticed, and my cell phone usage and content has improved, I can tell that you now. But we need to continue working together to solve these interconnection issues, because otherwise markets like Nigeria will not be accessible. I want to thank you all for your time today, and any questions.


CHAIR: That was really, really brief, we have 20 minutes for questions... Randy?

AUDIENCE SPEAKER: Randy Bush, could you go back a slide or two? That one there. So, North Africa and Maghreb have left the continent. This is news, I didn't see it in today's morning paper? Maghreb, northern Africa, aren't they part of the continent?

MICHELE McCANN: Yes, they are.

RANDY BUSH: But what's actually happened to prices in South Africa which is kind of interesting, because from my little understanding, it was quote deregulated, but actually pretty much a monopoly and the prices came down, the gossip is prices came down far less than one would have expected and it's a highly tilted financial situation. Is that ameliorated any in the last year?

MICHELE McCANN: Absolutely. So, one of the big things, the monopoly telecom, were taken to competitions committee and that was about a year ago, and they have now had to do the official split between the wholesale and retail. So that's the one major change that has happened. As well as there's been an increase of capacity, and global guys like a level 3, BT, AT&T starting to invest in Africa and we have seen the prices pretty much, and personally I have also seen this is from 100 dollars drop down to like 30 dollars, and if you really push them hard you can get to 11 dollars at the moment for transport.

CHAIR: Are there any more questions? I don't see the queues at the mics. What are we going to do with 20 minutes?

RANDY BUSH: What's it like for the consumer wanting to buy; in other words, at the other end of the last mile, not just transport?

MICHELE McCANN: So on the consumer side what's happened there is the mobile operators are having a price war at the moment. Personally, they are not handing over the savings. So me as a contract user, I'm paying 250 dollars a month for a gig of usage bandwidth. So they are not passing on their savings to us at all. So even though, yes, they are experiencing the savings on their ‑‑

RANDY BUSH: Okay. And what about fixed line?

MICHELE McCANN: So, fixed line, in terms of fibre, yes, there's been a huge drop around pricing, and what's happened there is there is a lot of consolidation, so, for example, Vodacom is now buying the Neotel assets which generally ‑‑ and the spectrum. So the key thing around South Africa particularly is being able to deliver the wireless assets, and we have seen those, some of them actually drop down to where the guys give the last mile away into the data centre to be able to have access to the services and sell the bandwidth.

RANDY BUSH: So what's it like for an E1 enterprise, or whatever?

MICHELE McCANN: An E1, last mile? Let me think about that...

RANDY BUSH: I mean that mobile price, of course, is rather shocking to most of us. And even for people paying European prices as opposed to the low prices I pay in Japan.

MICHELE McCANN: Sorry, and E1 fibre connection, you consider about 100 dollars a month, last mile within country.

RANDY BUSH: That's not bad.

RANDY BUSH: So you are paying 200 for your mobile, and FGTH would be 100?


RANDY BUSH: Something is strange in your market place. How does it compare with Kenya, for instance?

MICHELE McCANN: So, on the Kenyan side of things, to be honest, I can't tell you the mobile rates, I haven't been working on those, but your E1s are sitting at a similar price.

CHAIR: Okay. Are there any other questions?
Thank you, Michele.

So let's introduce the new term, is there a brave heart that would like to do the five‑minutes lightning talk from the mic in the audience right now without slides? No...

Okay. So now we have the coffee break. I hope it's prepared. Be back for the next session, we have lots of more speakers to come and I would also like to invite you at 6 p.m. in this room, there will be the BCOP Task Force Meeting, where we will be talking about the best current operational practices. The agenda is packed with good operational content and everyone is invited to join us at 6:00 here. Thank you.