Welcome to the United Nations | Department of Economic and Social Affairs

EIGHTH INTERNET GOVERNANCE FORUM

   BALI

   BUILDING BRIDGES - ENHANCING MULTI-STAKEHOLDER COOPERATION FOR GROWTH AND SUSTAINABLE DEVELOPMENT

   OCTOBER 23, 1013

   11:00 A.M.

   Session 48

   REMOVING BARRIERS TO CONNECTIVITY: CONNECTING THE UNCONNECTED

  


The following is the output of the real-time captioning taken during the Eigth Meeting of the IGF, in Bali, Indonesia. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the session, but should not be treated as an authoritative record.


 

 

   >> Good morning, everyone.  We'd like to get started in a minute or two.  If you could take your seats, please.

   Okay, good morning, everyone.  This is workshop 48, removing barriers to connectivity.  My name is David Satola.  I'll be co-moderating this morning along with Michael Kende from the Internet Society.  We've got quite a panel this morning of six speakers.  We'll try and get through their initial presentations rapidly.  Part of my job as co-moderator today is to be the grim reaper of time keeping, so I'd like to have each of the panelists give a short introduction, three to five minutes, of their main points and then we'll move on and open things up for discussion.  And I hope that you've brought your questions with you.  In case you haven't, Michael and I are prepared to take on that role, but we'd like to hear from the audience.

   I will briefly introduce the panelists by name, but I would like for each of the panelists to give a short summary of their affiliation and what their main point is.  So -- and if you could raise your hand.  Simon Milner from Facebook.  Okay.

   >> Hello, yes, Simon Milner from Facebook -- Jennifer Haroon, Pranesh Prakash.  Christoph Steck.  Raj Singh and Michael Kende, my co-moderator.  Simon, you may now begin.  (laughter) Oh, Martin levy.  Martin, raise your hand, please, sorry.

   >> MARTIN LEVY: Thank you very much.  As I was saying --

   >> I'm Simon Milner, Facebook's policy director for U.K. and Ireland.  I'm delighted to be here.  I thought it would be useful to talk a little about Internet.org.  You may have heard about this initiative that Mark Zuckerberg our CEO announced some weeks ago.  I'm going to spend my time explaining what that's about and how we see it fitting into the wider spectrum of initiatives around trying to connect the unconnected.

   So there are three pillars to Internet.org.  Data affordability, data efficiency and business models.  So on data affordability, we're looking at mechanisms to make mobile connectivity cheaper and to decrease the costs of delivering data.  And we're exploring projects in a number of countries around the world working with partners to achieve that so, for example, we're exploring smart friends that provide higher quality access at lower cost.  So that's the data affordability part of Internet.org.

   Then on data efficiency we want to build more data efficient apps and services.  We know that Facebook is one of the most popular apps that people use around the globe, and indeed spend inordinate amounts of time on their mobile devices using Facebook.  So we want to invest time in terms of our own engineering teams and those from device manufacturers and others to reduce the amount of data that our apps consume, whether that's about the main Facebook app or the messenger app which are the two people are using from the Facebook family.

   Our analysis shows that we can reduce the cost -- the -- that we can increase the efficiency by a hundred times, so that's what we're trying to achieve and we recently put out a white paper with -- Ericson and korkum looking at exactly how we might achieve that.

   Final area we're working on is business model.  These how we can device manufacturers and Facebook itself in order to make -- make accessibility as chief as possible.  And the startling factor is whereas for the developed world costs of Internet access are roughly 1 to 2% of someone's monthly income, in the developing world it can be 25%, and clearly if we are going to connect the unconnected that's one of the things we need to tackle.  We need to massively reduce -- we need to massively reduce the cost of access for people in developing countries.

   I wanted to highlight a couple of examples of how we've already been doing this within Facebook.  So in July 2011 we launched Facebook for every phone, and this is a Facebook app for basic feature phones and which it is a text-based experience and enables people to consume -- to have a Facebook experience on a much cheaper basis than you might experience with a data package in a smartphone.  About 1 million people, so that's around a tenth of the total population of the world using Facebook, is using Facebook via this application.  On more than 7,000 different devices.  So that's an example of how we've worked with mobile networks and device manufacturers to provide a means of accessing the Internet using Facebook, connecting with people around the globe as well as people in the next village using that.

   Another example is Facebook Zero, also known as zero.Facebook.com, which is a mobile site which provides all the different features of Facebook but is optimized for speed.  And we've now done deals with over a hundred mobile operators to bring this cheaper program to more consumers around the world.  And really Internet.org from our perspective is about building on those foundations to try and connect the -- whether it's four and a half or five billion people who are currently not connected.  So that's what we are bringing to this initiative.  Thank you very much, and sorry for speaking a bit too quickly.

   >> DAVID SATOLA: Thank you, Simon.  Jennifer?

   >> JENNIFER HAROON: Thanks, David.  I'm Jennifer Haroon and I work on the access team at Google, and I don't think I need to tell anyone in this room, but the idea of removing barriers to connectivity is a really big challenge, and we really believe it's not a challenge that's going to be solved by any one organization or one technology.  And so we think about technological and business innovation, but also the strong policy foundation that's needed to enable those technologies to flourish.

   So there are a couple of different things that I wanted to touch on.  One is the fact that multistakeholderism, which we're discussing a lot here at IGF, is really important to us, and just two weeks ago we were in Nigeria to discuss the Worldwide Web foundation and their launch of the alliance for affordable Internet, which really is a group of 30-plus organizations from Civil Society, governments, nonprofits and private companies like ourselves and Facebook, working on the policy issues that can bring about a more affordable Internet.  So that's one piece of the puzzle.

   Another are these new technologies.  So there are already proven technologies that can help remove these barriers to connectivity, but there's also an opportunity to look at new technologies, such as TV wi space, project lune, which you may have heard of, and we also look at others external to Google who are developing new technologies.  Just yesterday we announced some funding for researchers at Stanford and Berkeley, who are developing some network designs based on software defined networking principles to try to bring -- to deploy and also manage rural wireless networks more efficiently.

   And the last thing I wanted to touch upon, of course, is the take-up of Internet.  It's sort of they're all connected, and there we really want to think about full Internet access, so where users are not just utilizing what's already on the Internet but participating and contributing to what's on the Internet.  And as some examples, we really try to launch local domains including YouTube so that people are being to upload their local content into YouTube, and we also have some programs like get businesses on-line.  As an example in Nigeria we got thousands of small businesses on-line, and one example of a success from that was a truck-hauling company, basically a guy had one truck, and he hauled stuff for hire.  And after getting his business on-line, within a few months he had to expand to over ten trucks because people were able to find his business.  Thank you.

   >> DAVID SATOLA: Thank you, Jennifer.  Pranesh?

   >> PRANESH PRAKASH: Good morning.  My name is Pranesh Prakash.  I'm policy director for the Internet Society and access knowledge fellow with the Yale information society project.  Today I want to just highlight a few items right now that I hope to be discussing in more detail later.  One, I want to talk about the relationship between cheap mobile devices, which are the most important things that are connecting people to the Internet right now, and patents, and how the market is right now failing to provide the solutions to this problem.

   Two, I want to briefly touch upon spectrum reform that is necessary, also to provide access to the greatest numbers.  And what needs to be done to encourage local language up take of the Internet and also I'd sound a cautionary note saying that what we're trying to move towards is not to provide access to walled gardens, because making that cheaper is generally easier but to have easier access to an open Web and an open Internet, and that I think is an important point that we need to keep in mind.

   And should I just touch briefly upon each of these points or should I save that for later?

   >> DAVID SATOLA: Yeah, why don't you take a couple minutes to do that.

   >> PRANESH PRAKASH: Sure.  On the issue of patents, right now we have, for things like 3 G technologies and 4G technologies which are being complemented in various phones, there are competing -- there are competing patent pools and the same patents aren't of course -- haven't been recognized across jurisdictions.  So there's a great amount of legal uncertainty in terms of which -- in which country do you have to actually pay for which patents, and this legal uncertainty coupled with the costs of the thickets that have been created, make both innovation impossible and make creation of cheap devices impossible.

   So what we are seeing right now with the flurry of cheap -- of the -- with for instance the Indian government, which is trying to create something called the Akosh tablet, as a cheap computing device, well, its creation is going to run into patent-related problems, because there are so many patents that are involved in it, that selling it for a sub-$35 price as the government is trying to do is just impossible.  So that issue has to be addressed, and right now there is no compelling interest amongst different market players to address that issue, because each of them -- each one of them is trying to maximize revenues and as things stand there is no government intervention or regulation of this.  So as we saw, for instance, in World War I, when patents relating to -- to airplanes were brought governmentally into a pool, perhaps it's time to think about doing the same to our mobile phones as well which are providing cheap access.  So where the inventors would also be benefited by having -- by being provided royalties and the regular users will also be benefited by -- by actually having access to cheap devices through which they have access to the Internet.

   So this is something that -- one of the ideas that the Centre for Internet & Society is doing research on right now.

   One other -- so I've -- I actually save a little bit of time later to talk about spectrum sharing and the other things.

   >> DAVID SATOLA: Very good.  Thanks, Pranesh.  Christoph, five minutes.  You have the floor, please.

   >> Okay.  Good morning.  Christoph from Telefonica.  I'm the public policy director.  It's my first IGF and the first time we speak about connectivity so I think that's fantastic, because I think it's the first step when you speak about the Internet, that first people want to get access to it and I think -- so thanks to the organizers for setting up this workshop.  I think it's very appropriate.

   Obviously in contrast to some of the things we heard before, I am working at a company which call business is -- the business is to give connectivity to people.  This is what we do.  This is why we're there.  And therefore I would like to focus a little bit on some of the aspects we had already, which is that basically if you want to increase connectivity on a world scale massively the first thing you have to do is think about private investments and how you can improve conditions for private investments.  It's a fact that investments in digital infrastructure comes from private companies.  This is behind the huge success of the -- of the mobile sector in the last years, has been basically private companies do these investments on a commercial basis, obviously.

   So I think the first step, then, is to think -- you know, to -- about the right conditions to achieve these investments, and obviously this is summarized under the nice headline of investment-friendly policies.  And what we mean with that is that if you invest money in a network, a lot of money in a network, you have to have the predictability at least that you might over time, you know, get back some of that money.  It's as simple as that.  And obviously a lot of times, you know, when you speak to governments, you know, it depends to whom you speak and you get different answers, the finance part of the government might ask you to pay more for spectrum licenses, for example, which obviously, you know, makes the in the end the final product, more expensive, because it needs to be -- over time.  So it is really a complex issue, but what we advise usually is that you have to think about how you can make, you know, basically your territory, your country more attractive for these investments, which as I said is around -- it's around having the right conditions to have private companies investing.

   Just to give you a number, my company last year invested around 9 billion euros in the world in new infrastructure.  Out of that 60% in Latin America, which is obviously a part of the world where you need to improve connectivity quite a lot.  So I think this is an example for one company.  If you put all the telecom operators together you might get too much high numbers.

   Second thing is, and this was touched already on by Pranesh as well, you have to think about the whole value chain of connectivity.  It's not just about the connectivity part.  It's about, you know, how cheap is a smartphone, because access in the future will be through smartphones.  It would not be through PCs.  It will be tablets at least, but it will be some form of mobile device.  So how expensive is that?  Is that affordable as well?  You have to think about the conditions under which network can be built.  There are network infrastructure -- network infrastructure needs to be built out and there are conditions to that as well.  You have to think about the whole value chain.

   Let me finish with speaking about the problem we have obviously because there are always regions which are not connected, and sometimes the -- what we call the business case for connectivity, there is not working out.  And the reasons are very straightforward and simple.  Usually there are not enough people there so basically building up expensive infrastructure for a few people in the end is not a commercial business.  So this is where we get into areas where we have all kinds of corporations.  We touched on a couple as well.  We do similar things.  There's an interesting project running in Peru currently which we tried to roll out in the future to other countries, which is Connect To Me, and this is basically about getting a private public partnership to make the case better, which means basically that it's usually involvement of the public administration, the work that was included as well, to improve the situation for the business case in the end and we connected more than 200 villages in Peru in rural areas to the Internet, which is of course from an economic and social perspective fantastic advantage.

   Just to finish with a little positive side, connectivity is improving actually around the world.  The latest number I've seen is that you have today average global speed for Internet connectivity of around 3 megabits per second.  That's quite considerable.  When I started in the industry ten years ago everything over 200 kill bits was broadband.  We are now two times over that average.  There is massive improvement in connectivity but we have to go on.

   >> DAVID SATOLA: Thank you, Christoph.  Martin, please.

   >> MARTIN LEVY: Martin Levy from Hurricane Electric.  Like Christoph, I come from the infrastructure side of the business.  We move bits around the world and what I'm going to talk about is going to take you deeper into the guts of the Internet, away from the original points about mobile phones and things like zero Facebook, with lower number of bits but into the place where the plumbing moves bits around continents, around countries, and around cities.  So a couple of key points that I want to talk about.  The subject that has kept me busy for a ridiculous number of years, the deployment of IP version 6 as a follow-on to the IPv4 protocol that has lasted us 30 years on the Internet.  This, when now provided at the core of the Internet, means that mobile operators, hosting providers, content providers now have a future path for massive connectivity as we increase the number of people, the number of devices, the -- just the total size of the Internet.  And it can't be understated that this vector, this addressing vector of providing more -- more addresses, more ability for more devices is fundamental to anything else that we do.

   This isn't an IPv6 session so I'll be terse about it and leave it at that until asking questions.  I'm going to talk about optimizing bandwidth because when you look at how the Internet works, the use of the word Internet, the interconnection of networks, we go back and at our level at the bulk wholesale moving of bits level, we look at the need to have more efficient Internet exchanges, to make traffic stay as close and local, to enable caches where content can be brought from one place to another to find an efficient place for that data, all of which makes the end user experience much better.  They may not see that infrastructure, but it actually is an important part of the cost model, and that brings me on to this other aspect of however you look at it, making bits cheaper, making it more cost-effective to use the Internet in any which way that an end user wants to use it relies on the large infrastructure, the under sea cables, the interconnects, the fiber in the ground, all of those items to be accessible to ISPs, big or small, and to enable them to expand their networks.

   The last part that I want to talk about within this section is local knowledge.  You can go out to North America, you can go to Europe, you can go to many places in Asia and you can find fantastic groups of people with -- with years and years of experience.  Occasionally some of them have gray hair, myself included.  Some of them are just young kids that really got it and go out and do it.  But it's just as important to find that level of expertise when building infrastructure, whether you're here in Indonesia, whether you're in Bangladesh, which I was in the beginning of the year, whether you are anywhere in the world, and you make sure that there are people being trained that are finding these avenues to -- other to ear geeks other geeks, people who understand how to run bigger networks.  I'll give you an example that's relevant to here.  There's been a few rumors, rumblings of people complaining about WiFi being a bit iffy, especially yesterday afternoon.  That's true, but that to me is a local issue inside a building.  I spent yesterday morning and a little bit on Monday going backwards and forwards with two of the major Indonesian telcos in Jakarta used for the infrastructure in this facility.  I went and said, I'm seeing bits flowing from here to Amsterdam only to come back to Hong Kong.  That didn't seem right and isn't right.  Of course it's expensive.  And in going backwards and forwards I was able to find the right piece of information and give it to the right engineer and all of a sudden about midday yesterday, the routing changed, things got more efficient.  It wasn't about the network, the single network here.  It was actually about every user in the whole of Indonesia using this particular service provider and getting their network to improve and just be -- just get that little bit of network knowledge imparted and move on.

   So it doesn't scale to have somebody like me sitting in a hotel or in a conference going, ooh, that doesn't quite look right, but it does scale if you can get the appropriate entities to do that localized training, to understand how to build Internet exchanges, to understand how to run BGP and the core routing.  So that ultimately the bits flow to the end users that want to use those cheaper tablets, cheaper mobile devices and simply increase the number of on-line users.

   So that's my view from way down in the guts.  Hopefully something that most people never see.  Thank you.

   >> DAVID SATOLA: Thank you very much, Martin.  Raj is batting cleanup today.  He's our last scheduled speaker.  If he doesn't break the microphone.  (laughter) So I'll turn it over to Raj.

   >> RAJ SINGH: Thanks, David, and I promise I won't touch the microphone after this.  So I am Raj Singh.  I look after the Internet societies in the Asia-Pacific region.  It was interesting listening to all the panelists talking because I don't think they have left me much to talk about, which is great, so I can maybe leave.

   One thing I do want to say, though, is that -- I know someone mentioned spectrum, I think it was you, Pranesh.  I was on an earlier panel this morning where the issue of spectrum also came up.  And wireless is the in thing in a lot of ways because people say, wireless is the solution to everything, you know.  It will help us connect quicker, faster and maybe cheaper.  But one of the things I do see and count this as a barrier to connectivity is the fact the spectrum has been auctioned at ridiculous rates by government.  Once you start spending billions of dollars buying spectrum, which has been the case in some countries in this region, that cost has to come from somewhere, and unfortunately that will be passed on in some way or form to the subscriber.  Therefore your subscription rates go up.  In the two countries that I call home, service providers have recently released L G-based subscriptions in both instances I find it ridiculous to subscribe to them.  It's too expensive.  It may get faster but I'm only consuming X gigabytes a month anyway and I'm happy to do it at a slightly slower rate.  So I do think that's a major barrier going forward, particularly if we talk about connecting rural areas where the cost ratio would be much higher.  I would imagine someone mentioned that there's not a large enough population to support commercial infrastructure roll-out.  So that I think is a major issue going forward.

   What I do want to talk about, though, is something that the Internet Society and a couple of other partners that we got together in 2010 decided to -- you know, it was more a trial.  Let's see what happens.  And what we did was conceptualize something called wireless for communities, and basically what it meant was go to a rural area, set up -- which was pretty much unconnected, set up a rural WiFi network, wire up the community and just see what happens.  And the experiment has been interesting.  In fact, it's been so successful that our partners have been really gung ho about it and gone on and done great things.  Now they have installations across India.  But the key thing there was that we decided that, you know, there's barriers at all levels, there's the policy barriers, financial barriers and so on.  But there's also barriers at the local level.  When you bring connectivity to a population which hasn't been connected there are cultural and social barriers as well that haven't been addressed.  We involved the community.  We talked to the village elders.  Once we had their go ahead then we decided to -- the network, showing them the impact the Internet could have.  Since then it's grown tremendously.  11 out of 13 schools in one particular locality we did are now connected.  They're now invested in the -- schools.  The health centre has been wired up.  They had equipment lying under a blanket for something like two years because they had the equipment that was donated by some very kind government, but there was no connectivity to use them.  But now they do.  So on and so forth.

   I want to talk much more about that but at the granular level that's how we made a difference and someone mentioned that no one organization can solve the problem.  That's very true.  That's why we involve everyone in the community, and the community itself is the ones we actually train to deploy and operate the laptop.  We don't -- we actually train them.  That was our USP if you like.  So they maintain it and keep it running.  The other criteria we had is that we use very low cost off the shelf equipment, something, if it breaks they can walk to the local town and buy it and replace it.  Those are other criteria we used.

   In ending my spiel here I want to say that the Internet Society has a paper called removing barriers, which we focused in Africa in the last round.  We are looking at doing further papers like that perhaps next year, including some for the Asia-Pacific, and, you know, what the paper does is identifies what some of the issues are infrastructure-based, capacity-building issues technical issues, so on.  Watch, we'll have it out soon.  Thank you.

   >> DAVID SATOLA: Great, thanks very much for those introductions.  I thought that was a really good table setting for a further discussion and now I'd like to open up for questions from the audience.  I see two over there.  Let's just go in order.  Dan and then Mike and then pepper, Bob.

   >> Good morning, everybody can hear me okay?  My name is Dan McGarry.  I work with the Pacific Institute of Public Policy.  We do -- we advocate for better policy processes throughout the Pacific Islands region.  I thought I might provide a little context, you know, sort of the view from the coal face, if you will.  First observation I'd like to make is I don't want for a moment to denigrate the great work we're seeing from Facebook, from Google especially, in making that -- you know, the social interaction, all these possibilities that come with the Internet, making them available to people in the developing world.  I can tell you for a fact that it's hugely, hugely popular and quite influential in many ways.  And it's -- I think the value of social communication, you know, on-line as a development tool is -- in spite of all the rhetoric flying around, I still think it's underrated and I don't think it's as well understood as it could be.

   But that said, and you knew there was a but coming -- that said, it's a necessary but not nearly sufficient condition for, you know, really leveraging digital communications in the developing world.  The observations from a friend from Telefonica on the mark in many ways.  We've seen good public/private partnerships.  We've seen some failed public/private partnerships as well, partly because in the Pacific we have very, very unique conditions.  In Tonga recently, which they just turned on their -- their fiberoptic cable about a month ago -- no, a few months ago, with the generous assistance of David and his colleagues at the World Bank as well as the folks at the Asian development bank.  I spoke with the Chairman of the Tonga cable company, and he talked about all the lengths that they had gone through to get commercial, you know, wholesale bandwidth down to about $500 U.S. megabit, and that translates, you know, at very high contention rates to a wholesale cost of about $25 per customer, which is about five times what the average Tongan can afford.  That's for a single megabit of connectivity.  I spoke further with some of the telecoms people including finance people, and they just got this deer in the headlights look when we talked about making Internet affordable.  They really do genuinely want to do as much as they can because of course it's in their best interests.  If there is a way to make money they want to do it.  And certainly some of the countries operating in the development world right now are very aggressive in terms of their margins.  But the plain fact is that the cost of infrastructure, the cost of maintenance, you know, just the sort of opecs, I guess, and the capacity, Martin, as you rightly pointed out, are -- the conditions are prohibitive, and so when we talk about multistakeholderism, from my perspective it's a much broader umbrella, a much bigger umbrella than we might consider.  You know, you hear a lot about government business, Civil Society.  With a lot of sort of hand wavy stuff around the edges, but the plain fact is that we need to have a different -- a fundamentally different understanding of investment in terms of how this infrastructure gets built out, because the last mile, it's not just a matter of putting out connectivity.  You need power.  You need a huge number of secondary services, financial services, retail financial services in the Pacific are a joke.  We've actually got fairly decent bandwidth in a lot of the municipal centres, but you can't do a thing with it.  So the bottom line is we do need a fundamental rethink about how we're going about it.  There's been great work so far but there's much more that needs to be done, and I for one am just not thrilled that we don't have many development institutions here at the IGF.  So congratulations on everything that you've done and I'd just like to underline the fact that there is a huge amount more that remains to be done.  Thanks.

   >> Okay.  Thanks, Dan.  Any comments or thoughts on that?

   >> I'll be brief.  Thank you for what you -- and I suppose one of the things we recognize as a company, that there's much of our growth is demand driven and is being demand driven in terms of often consumers telling other consumers, you should get on Facebook, then we can connect that way.  So we've seen get to 1.1 billion (this is Levy), and the one after that and the one after that, cannot just be demand driven.  So I absolutely support what you say.  I suppose the question is, what can we do around this table in terms of that fundamental rethink.  It's all very well saying it but what does it involve, I suppose.  I'm here to listen and learn from the colleagues who are here about that.

   >> DAVID SATOLA: I love the deer in the headlights statement, and so just so that people have a reference point, if the wholesale price once delivered inside Tonga is delivered at $500 a mega beat after an undersea cable is installed, the other end of that cable, if we consider that to be Los Angeles, California, the wholesale price of bits in that city is below 50 cents a megabit, 1/1,000th of that price.  That multiplier alone just stops you in your tracks trying to work out how to do deliverable bandwidth.  And I don't know the economics of that cable, but it's better than it was.  But unfortunately not a brilliant way forward.  I'm going to bring up a quick point, if we talk about geography, this is actually a key point.  The places that have been successful at the infrastructure game, at coming up with affordable prices, have been places that become natural gateways for other places to interconnect with.  Tonga is not in that camp, unfortunately, but slightly to the west in Fiji, they were in that camp.  They bid and successfully got to be part of the landing site for undersea cables towards Australia, a massive marketplace, and they've benefited as a side result of that, because the cable stopped.  99% of the bits continued but the fact it stopped gave them an ability to take advantage of infrastructure costs (this is Martin Levy), far nicer, I think, than you have in the example in Tonga.

   And that's hard for some places to do, but for other places it may be worth thinking about.  Understand how -- and this is all about looking at a map and looking at geography, how could you become a gateway to other places, share the load, share the cost and therefore reap the benefits of a more efficient pricing model.

   >> DAVID SATOLA: I think we'll move on to Mike and then Pepper.  I see a hand there.

   >> Thanks, Michael.  I have a question for the representative from Facebook regarding the exclusive focus on mobile technologies.  I was wondering about the rationale for that when we we've heard already from the panel about the need for more of an ecosystem approach with the importance of WiFi technology, for example.  There are other delivery mechanisms such as TV wi space and the need for electrical power delivery as well.  So I was just wondering how Internet.org justifies this rather narrow focus just on mobile technologies.  Thank you.

   >> Yeah, the focus is very much on mobile devices.  So it's not just about mobile networks.  We also think WiFi is incredibly important in terms of that.  So we're certainly exploring opportunities for working with people who are trying to bring WiFi to unconnected geographies.  But very much expecting that most people will want to access by mobile because that's why the big group is -- of 1.5 billion years many are using Facebook on a mobile device, and they spend much more engaged, spend much more time on it.  So that's one of the reasons why mark flipped the company to become a mobile first company, because that's where we see users most want to enjoy Facebook and use it.  That's why we focus there.  I was in Nigeria recently with Jennifer and that was very much the focus there.  In terms of the markets we're focused on it's mobile devices typically (that was Simon).

   >> Great.  Pepper?

   >> Robert Pepper, Cisco.  So actually I want to come back to -- I was going to pick up on Raj's comments, but then the last one on -- is related, which is mobile wireless.  So one of the things I think that's important is to note that, number one, virtually all of the devices that people are using today are wireless, are connected as wireless devices.  All right.  The number of devices that will being connected with Ethernet is declining dramatically, and if you take a look at the data -- the amount of data going to wireless connections, whether it's mobile data, WiFi, which is off-loading mobile data, or fixed WiFi, that is going to be -- it's already over half of the data.  It's going to be two-thirds of the data.  So the devices will be wireless devices.  That's number one.

   Second and related to that, you know, there's this false choice debate that people sometimes raise, which is, is it going to be fixed (audio difficulties) operators talk about mobile networks.  There's no such thing.  The networks are not mobile.  The networks don't move.  I do, you do.  Right?  The networks, in fact, are fixed, even if they are wireless networks.  Mobile satellite is sort of a footnote exception.

   What that means is that when you connect these devices with wireless technologies, whether it is macro cell, which is licensed for outdoors when you're moving around, high velocity, right, you're in cars, or it's small cell when you're indoors or outdoors but not moving around, and that's predominantly going to be WiFi.  The answer is, yes, both, but none of that will be scalable unless it's connected back to fiber.  Right?

   So the future architectures clearly are going to be heterogeneous network of licensed, small cell, WiFi, PICO, but mostly WiFi, all aggregated and back -- into fiber.  Without that it won't work.  So when we're looking at affordability, where are the barriers on the affordability, it's not in the low-cost WiFi -- it's not in the low-cost devices or the low-cost WiFi connections, and, you know, Mark can talk more about those.  It's the bay -- the bay years tend to be middle mile or back hall to very low density areas, and that's what we need to be thinking about and also other ways to affect that are IXPs, which lower the cost within continents and within regions also to keep traffic within regions, and I love -- one of the things I love about Martin is that he actually looks at where the traffic flows and traffic flowing from Bali to Amsterdam, huh?  Right?  So these are things if people know they can fix.

   One last point on Raj's where I actually disagree with Raj, his statement about auctions.  The reality is that we actually know that what drives pricing is not spectrum auctions.  It's the level of competition.  And if you have competition it drives prices down.  The price paid for spectrum in a properly designed auction is like any other input, right?  80% of the cost of building the wireless networks is civil engineering of massed towers, back hall, electricity, security, backup generators.  The spectrum is a rounding out error of the radios, and we can have a second conversation but empirically spectrum auctions don't raise retail prices if the market are competitive.  If the markets are not competitive you can give away the spectrum and you're still going to have prices that are too high.  Thank you.

   >> Okay.  Thank you.  Looks like there's no comment.  The next question -- yes, please.

   >> Just a -- for Martin, keeping data local, is that going to stop other people from having a look at it?  So if -- if the president of Mexico had all his email in Mexico, would that have stopped somebody in the U.S. from reading it?

   >> MARTIN LEVY: Good question, a little out of my -- out of my league, but I'll talk about it from a where bits flow point of view as opposed to anything else.  One of the wonderful things about the Internet and the reason why, you know, I could -- I could help do this quick fix yesterday and why I do look at bits, actually, to hit Pepper's sort of like, you know, who I am, is that the Internet, the routing of Internet bits is unbelievably transparent.  The global routing tables are truly global.  This network sitting here, which was built for IGF, is part and parcel of the global Internet.  It has what's called an autonomous system number, a unique number that identifies it in the world.  It has a set of addresses, IPv4 and IPv6, allocated from the local country because it's Indonesian but out of the registry, and the addressing and routing of that exist on every Internet router around the globe.

   So, for example -- or in comparison, Anne actually Mexico is a brilliant example of this.  There was a comment at the microphone during the wicket in Dubai in December where the gentleman from Mexico said, we don't like our voice traffic flowing via another country and then another country -- I'll ignore all the countries that he mentioned.  I should not be country-specific here at IGF anyway.  And then end up at its destination.  We want better control of that.  And in the voice world it turns out that's actually quite hard, the way that voice at the wholesale level is resold, the way signaling system 7 works.  But in the Internet world it's totally transparent.  So if the traffic did flow from Mexico, and I'll pick it arbitrarily, over the border where 99% of all the fiber is anyway, into the United States, and then flowed around somewhere and then flowed back to somewhere else in Mexico, maybe a different service provider, the thing is that all that is transparent and findable and seeable.  Oh, hopefully, and fixable.  Because it's obvious.

   There is -- there are a couple of the major players in Indonesia that I can see buy bandwidth and some of the traffic flows to all, but some flows up to Singapore and then flows back to a different service provider here in Indonesia.  We see that around a lot of the world.  It's been pretty flushed out in Europe, and it's been flushed out in most places in Asia.  You will see Japanese traffic stay within -- inside Japan.

   So I don't know the specifics of the Mexico example, but the point is in the Internet that routing is 100% visible and 100% therefore fixable.  So dealing -- the answer from the infrastructure point of view, I want to hit one subtle point.  There was a project in Canada done out of Toronto called, I think it was IX maps.  I think it's IX maps.CA.  I'll find the URL, and they basically said, we want to map all the traffic in Canada that flows down towards the United States, maybe from, let's say the West Coast, Vancouver and end up going back over the border to the East Coast, maybe Montreal.  There's a reason that happens.  Trans-Canadian bandwidth is still quite expensive.  You pop south of the border in the United States and you're dealing with commodity pricing of bandwidth, much cheaper, but all of a sudden you're in another country.  So your question could well be very valid for the Canadian, and there is this example and they're explicit why they don't want to see this happen.  I want to find the right URL.  However, it's a great study and the economics of fixing Canada would be wonderful, so that traffic truly does stay in their case north of the border.  You -- you need to look at examples like that to see how this is done.  But the point about that study is, it's totally open.  It uses Internet tools to measure it, and it's absolute.  If a bit does flow down to the U.S., you see it.  It's all there.

   >> Great.  Thank you, we have the two gentlemen sitting next to Pepper, first on the left and then on the right, and then Lorenzo, I think.

   >> Mark Sommer was in -- venue is a nonprofit organization.  We focus on getting connectivity and access to ICTs in general, really to underserved rural areas, mostly sub-Saharan Africa, but actually underserved in that context means often secondary towns as well because real bandwidth -- real broadband opportunity for use ends up stopping in places like mom bass a and nigh robe nigh robe yeah.  Technology is a key thing, things like compression, efficiency gains.  A key component to it but ultimately I think it's more of a business and a market failure, an opportunity to grow those kind of inputs.

   It's typically a very non-diverse ecosystem of bandwidth delivery in these countries.  You have maybe two or three mobile operators who are competing for the market, and that really stifles, I think, the competition to provide services ultimately.  Those two or three operators provide most of the infrastructure they need by themselves and what I see happening is when regulation opens up and gives more entrance to the market access, suddenly you see middle mile providers providing fiber independent of those actual carriers who sell it at wholesale rates to smaller new providers.

   I think another one is spectrum.  Spectrum is often purchased on a nationwide basis but often not used on a nationwide basis.  So if there's ways to reallocate spectrum more dynamically, reuse spectrum in certain areas service providers don't want to use the spectrum because they don't want to invest, I think that would be a huge change and huge opportunity to get more competition into the market.  And ultimately really encouraging once these systems are in place, local entrepreneurs to provide bandwidth services and establish basically service providers in that country.  I think Martin and Raj alluded to getting that ecosystem together I think is a key component, because then you can have competition in the actual technology sector, mobile versus WiFi versus lune or whatever it might be.  But as long as the markets are correspond by one or two -- are cornered by one or two operators, all the -- they're losing out, the content providers or the people who actually provide services because there's not enough choice in the market and that's one thing I think is really important, to see how we can change that.

   >> Pranesh?

   >> I wanted to pick up on the discussion on the spectrum.  To address an earlier point about spectrum costs, I don't think that -- well, cost is very important, but it's equally important to see spectrum permissions, okay?  So discussion -- I would like to see a little bit more discussion around licensed spectrum in terms of who -- in terms of openness as to who all can use it rather than just in terms of cost, of whether it should be auctioned or that method of allocation of spectrum.  Right?  And secondly, the -- the counter-problem can also exist to the nationwide provisional spectrum, where you have different circles being broken up, okay, and nationwide players who haven't actually got the right to that spectrum -- to the same spectrum on a nationwide basis, have difficulty in actually providing a service on a nationwide basis, especially when the rules also prohibit relicensing or sublicensing.  And so there can be the opposite problem as well existing, and that has happened in India, as a matter of fact.

   >> I would just like to build on that.  I mean, I think -- I think you're right in the sense that as we said earlier, the full ecosystem approach is the right one.  I have to disagree a little bit on that, if you are three mobile operators competing, you do not have competition in the market.  I mean, there are -- there might be discussion about if it's enough competition or whatever, but the experience is that competition in the world have gone down dramatically.  There's no other industry sector I know where prices have been so deflationary as in our industry.  They're still -- this is not perfect.  There might be exceptions and countries where this isn't happening but then the reason is because maybe the competition is not good enough or the conditions are not right.  Having said that (this is Simon), of course you might imagine that a company like ours investing billions every year, we're looking into ways to make it cheaper.  We don't like that.  Buying spectrum and investing, are the most difficult decisions for management every year.  And so of course we are looking into all ways we can improve that and make it cheaper.  In the end we want to sell a product, we want to sell connectivity, it's not that we wish no one gets on-line.  This is our core business, as I said.  I think we're looking into that and we're trying to improve it, and as I said earlier and I think Pranesh touched on the point as well, what we've seen is very often that, for example, cheap devices, yes, it is an issue.  It's not that smartphones are very cheap compared to -- at least to the income in some countries, and we've started therefore with a foundation, a project that I think is fantastic, which is working on Firefox OS, which is an open operating system.  There are no license fees, no royalties paid, which is one of the reasons why smartphones have become much cheaper.  And we have devices now on the market.  For instance one device that the Chinese put -- producer.  You can buy them for 50 euros, which is fantastic, a fantastic price compared to others.  So I think these are the solutions we have to improve and obviously we have to look into new ways of giving the last mile Lexus -- and every idea is welcome.

   >> JENNIFER HAROON: Mark, I just wanted to comment on what you said about spectrum, particularly in rural areas potentially not being used, because as Christoph said, the investment, the business case for building out networks can be very difficult.  And one thing we've been working with governments on is mapping spectrum, because as you said, you can see that while licenses tend to be given out nationally, that's not how they're being used, and we've mapped spectrum for South Africa and Senegal recently and it's amazing to see, even though everyone says spectrum is a scarce resource, how much spectrum is actually available to be used if you consider location and also time, particularly when you're talking about TV bands or even radio bands.  And so one of the benefits of looking at TV wi space technology is how can we use that available spectrum to deliver broadband in the right places and at the right time, and that can be done, for instance, through spectrum database that can then dynamically manage when different frequencies can be used to deliver broadband.  And that technology can be used for -- by mobile operators, by ISPs to add competition to the mark.  So that's one of the things we're really interested in supporting and we're wrapping up a trial in South Africa that is using TV wi space back hall to deliver broadband to a number of schools, and it has shown that there is no interference with other licensed users of the nearby bands.  And so now it's about working with regulators so that they enable this technology to be used, whether it's unlicensed or through other managed means.

   >> Please.

   >> Thank you.  I'm Jayantha Fernando from Sri Lanka.  First I thank ISOC for putting on the panels for connectivity.  Just a quick question and thoughts coming from my own jurisdiction.  In terms of removing various connectivity, we are mindful about the spectrum issues and the spectrum point issues.  We try to provide a homegrown solution as much as possible, but one area which helped improve connectivity was to set up a mirror of one of the root servers in our country, which was facilitated by my agency in conjunction with the API NEC and of course support from ISOC in that whole exercise.  So just a thought and a question perhaps to the panel.  How important are mirrors of root servers, and to what extent do they help remove barriers to connectivity and empower people with faster and easier access to the Internet.  Thank you.

   >> I think Raj should talk about the work at ISOC.  (audio difficulties).

   >> The example I want to give was presented a week or so ago in Europe at the -- at the right meeting, which is the RI (off mic) for Europe and the Middle East and Russia.  And the example was actually somewhere that was fairly close to the middle of Europe.  This was a -- a monitoring of the root name server response time that was done in Belgrade, Serbia.  (audio difficulties) took them from about a 44 millisecond (audio difficulties) couple of hours on a flight at best, not even, actually, now that I think about it, down to -- down to literally a handful, 3 or 4 milliseconds inside the city.  That latency translates to each and every user as they are doing a Web search, as they are typing in www.whatever they feel like, hopefully with the local ccTLD at the end of it.  But that root original question is fundamental to part of that query.  The DNS system is fundamental to the way the interpret works.  It's not just the installation of one root server.  There are a total of 13 servers.  Putting three or four inside a local facility, moving a copy of one into Tonga, for example, makes all the difference in the world.  I keep picking on Tonga but it's so far away from everything it's a great example.  So it wins.  It wins hands down, and there should be hundreds, if not many hundreds, of these instances.  Not all the same letter.  The letter by the way, just for those people -- a quick aside, root name servers are named A through M.  There are 13 of them.  They are distributed around the world to different organizations to manage.  But then again they then redistribute the instances, the multiple copies of the same thing, mirror images, even further around the world.  And all of this again, the wonderful thing about the Internet, totally available, a quick -- a quick search on something like root name server locations will give you a wonderful map, which changes and improves.

   So it does make a difference.  It makes a difference to the end user.  It isn't a lot of bits, but it is the latency, it is the responsiveness that changes, and Raj, you can talk about the wonderful work that -- pushing them more and further away.

   >> RAJ SINGH: I think I'll leave it there.  You've covered the technical stuff pretty well, so yeah.  There are a lot of organizations that run the root servers and most of them are quite happy to talk with you and set up mirror instance in your locality, and if anyone here is interested issues ask me and I can point you in the right direction.  Thanks.

   >> Lorenzo, did you -- no, okay.  Jane?

   >> JANE COFFIN: To add to the point that Raj just made about the instances, this is something that the Internet Society is very keen to promote when we're installing and helping local partners, and I stress partner because we don't go where we're not asked to go and we don't do anything usually by ourselves.  It's either in partnership with, ons on the ground, companies, other nonprofits, but we're very keen to make sure that when we're building out IXPs with the partners, that the -- these instances are part of that process.  So I just wanted to put that out there.

   And I did want to sort of elaborate or note an issue that has come up about Internet exchanges and monitoring, and IX isn't designed to do deep -- inspection.  That's not what it's all about.  There may be issues related to ISPs that have been asked to do that but I want to clarify when we're going into develop and emerging markets and putting in basic capability for an Internet exchange point, we don't do that, and they don't do that.  The objective is to bring competitors and others together to exchange local traffic, keep it local, create a cheaper environment because the quality of service goes up, latency goes down, volume of traffic spikes, we've seen it from the Democratic Republic of Congo, to Ecuador to other places we're working and the fellow sitting next to me is our regional director for Latin America and he's shaking his head.  I work closely with him.  We want to be clear that the Internet -- organizations are trying to build local connectivity, get a stronger more resilient infrastructure are doing -- there may be places where they're monitoring but when we use the word monitor we're talking about how much volume there is from a daily or weekly or monthly basis so that others can see there's the investment potential in that country.  When we say measure and monitor we're talking about the health of the infrastructure, the efficiency of the infrastructure.  So I want to be clear that when we speak in that language from our Internet Society and other partner perspective we're talking about how you make the network better, more efficient and manage better.  I wanted to add that perspective.

   >> I'm from New Zealand.  If I can go back a step to the spectrum discussion before New Zealand has implemented a somewhat experimental policy designed to be a halfway between the unlicensed regime and the typical auction/management regime.  It's now in a spectrum -- features of the approach, heavy user or loser provisions.  No national operator across all the licensing -- across all the bands.  The spectrum is harmonized with band 48 -- band 41 and band 38 and LT 8 TDD.  And it's actually regionally divided.  So I think from memory there's about 20 subranges within New Zealand that operators can license up and use it.  Also another feature is multiple operators can license in a region, but conflict resolution between them and how they manage interference is organized between them.  So if anyone is wanting to look at different approaches, I suggest have them look at managed --

   >> Thank you, any comments on that?  Other questions?  Yes, please?

   >> Hi, Mike Kellye from the American Bar Association.  We've heard a lot about the front end problems that you encounter with connectivity, spectrum power, all these kinds of things, but I'm kind of curious about the back end issues, and that is sustaining connectivity, once it's connected.  I was really interested to hear what the Internet Society does.  You know, if you go to a village in the Congo and you get connection, I mean, do you select someone to train in that local society to be tech support or do you leave an Internet Society person behind for a period of time to do that?  How does that all work on the back end of sustaining connectivity.

   >> So our USP on the project is by the community for the community.  What that means is we don't parachute in, leave the stuff and leave.  Our whole idea is we train the commute to maintain the network.  It's a bit of a challenge because not all communities can support those skill sets.  We started off the project by first doing the training.  It was people training, getting the locals involved and it's typically people who had done high school, they finished high school or secondary school, whatever you want to call it.  Some of them had some basic technical education.  Some IT skills, they knew what a computer was, how to operate it, and then we spent two weeks training them.  Then we went back three months later and did it all over again so that they were refreshed on what -- they knew what they were doing.

   The other part to the sustainability equation, if you like, is the fact that, you know, the back haul costs money as well, the equipment costs money.  Maintaining the equipment costs money as well.  So what we want to do is come up with a model for them, where the network actually generates revenue, and the way we do that is, for example, one of the government offices wanted to get connected so they pay for the equipment cost, and they agree we can use them as a relay station to get further out into the area.  And then there are things like sidewalk cafes swung up, didn't exist before.  Schools got wired up.  Everyone pays a minimal cost to access the network because there is a cost associated with it, but it's not a commercial operation.  It's basically whatever it costs is divided by the number of people on it.  But we let those decisions be the community decision.  We don't decide for them.  You know, what we show them is a, withing model.  This is how it operates.  Here are suggestions how you can sustain it and the ones who are truly committed are the ones we see they become successful.  In terms of choosing the community I think was your own question.  We have local partners on the ground who do a lot of that selection process, and the essential -- essential criteria is local buy-in.  The village elders or the elders council must have buy-in.  They must support it.  If they don't the project fails.  That's what we've seen.

   >> Actually if I can is ask a follow-up question on that, and kind of I guess take the moderator's prerogative.  So we've heard all these great initiatives to lower the cost, make the Internet more efficient.  Clearly that's going to increase usage for people who already have access and get more people on-line.  But I was wondering if we could just delve a little into the demand side and build on what you're saying about the literacy, local languages.  I know Google is working a lot in local languages in Africa and other places to make sure once access is available that the most people see it relevant and take it up.  And Pranesh had his hand up.

   >> PRANESH PRAKASH: That's precisely what I wanted to touch on in response to that question.  I'll point out a few things.  One, in India what the government is doing is it's rolled out more than 100,000 kiosks, community service centres, where the main focus is on government services, but it also serves as a place where Internet availability is spread through.  Not all of them are equally functional but it's working much better in a sense than one would expect.  So -- and there are some studies on it as well.  If you go on-line and search for common -- centres in India on the government Web site you'll find studies about it.

   Secondly, one very important thing in actually keeping demand up is entertainment.  People don't just go on-line, you know, with a thirst for Wikipedia all the time.  They are looking for entertainment very often.  And one very important thing that emerges out of this is piracy.  So piracy is actually something that does drive -- drive the demand for Internet access.  One way to actually -- to help the situation is to -- and what's happened -- what's happening right now in India, one of our studies has shown that piracy is actually great for the industry and for the entertainment industry, and what many music companies are now doing is that they are co-opting pirates, especially those who are at the mobile phone store, who are doing site looting on to mobile phones, et cetera, they're actually making -- they're not demanding their regular royalty fees anymore.  They're actually -- they've actually realized that what the market finds sustainable in terms of royalty fees is much, much, much lower in this digital age and they're profiting out of it.  So unlike in the west where music companies are facing difficulties, while movie companies are doing great but even despite or even because of piracy, in India the music companies are actually starting to change what they're doing now and earning money out of it by co-opting erst while pirates by lowering cost point.

   A few other things I want to quickly mention.  One, text-to-speech and speech to text, is incredibly important in getting illiterate and semi-literal people on the Internet.  Unfortunately this costs more amount of bandwidth, okay?  But we still have to focus on -- and it's in my opinion an area that's way too neglected, and so that's something that if we want to get -- get to people who can't type, that is something we have to focus on, and secondly, government regulations in some ways are good, so parts of what I'm saying, shared spectrum, unlicensed spectrum -- shared spectrum needs good government regulations.  Unlicensed needs less government regulation.  So there's no one level of government regulation that's good for everything, right?  And one area where government regulation is good is to say that when you're selling a mobile phone, you have to have it enabled -- you have to provide multiple local languages on it if you're selling it in this particular country.  And that actually isn't that very costly, but just because -- there isn't enough of market incentive generally to get that done because the folks who can't talk in English or who can't use the keyboard don't constitute the market for many companies, right?  So -- and they use the Internet in very different ways, with video being a more important constituent of the data for the illiterate people.  That is something that could also do with a little bit of light government intervention.

   >> Great.  I think we've got time for two more questions and -- from the lady there and then from Edward.  Please.

   >> Thank you.  The earlier question, I got lost looking for this room.  Apologies.  My name is Nina and I'm from the Worldwide Web foundation, and one of the things we do is the alliance for affordable Internet that was spoken about during the opening ceremony.  As I followed through what --i followed that you've covered most of the issues, but I just wanted to come back to the things that the alliance is doing and will continue to do.  One is encouraging discussion and advocacy among partners.  The second is research, but the third and most important I want to speak to is public support, especially in areas where bandwidth is expensive.  Why is it privilege?  What are the policy and regulatory side to it?  And one of the things that we are doing, we'll be sharing best practices within the countries.  Recently we signed an -- something with Nigeria to kick off because Nigeria has the biggest market in Africa and we believe we can make a success story of Nigeria that will ripple across the continent.  So one thing will be liberalization of the market with an open competitive environment, which we feel is very important.  And I want to speak to practices that encourage lower cost structure, because that is actually one of the huge issues, what's the cost structure in place and what are regulations around it and what can be done to reduce this.

   So that will be streamlining processes for substructure deployment and sharing.  The other will be effective spectrum management.  It's been spoken to.  I followed it on-line here.  Then enabling ability of the usage through a licensed spectrum.  Jennifer has spoken a bit about TV wi spaces.  And -- taxation or excessive cost on tariffs.  Countries don't know.  I was speaking with the minister two weeks ago and she said the country needs taxes.  Yes, we agreed that we need to strike that balance, but we need to keep policy support to these countries to know how -- while at the same time maintaining this balance.  I'm available to speak more and to connect anyone who would like to know more about the initiatives of the alliance for affordable Internet, and once again, thank you to our advisory council members who are here.  Thanks.

   >> Thank you.  Edward and then if we have time -- oh, sorry, of course.

   >> I want to thank you as well.  You are a prolific tweeter about this subject.  I've been following you for quite a while, and (this is Martin), and it is one of whether you're here or are you back in code -- ivory, I get to see an aspect of your interest in the Internet from all aspects.  I wanted to thank you for that.

   >> Okay, Edward?

   >> Yes, Edward from idea source.  We are a venture capital incubators based in Jakarta and we're investing in Indonesia as well in the region.  I just wanted to comment and add about the accessibility on the demand side.  I was on a panel about the IXP ecosystems.  I think for -- for example, in Indonesia, TV is very popular, and the way I see it, you know, is that the movie, the smart TV right now, and development from Google of chrome cast, we have $35 devised I think is going to be -- add more accessibility and also to create more demand.  I think that's my comment.

   >> Can I say -- one of the things that struck me about the session is we've been talking a lot about, you know, how much Internet is demand driven, how much of it is availability determined.  Most of our countries, I live in the U.K.  We've got 9 million people in the U.K. that have never used the Internet.  It's not an availability problem.  It's a relevance problem.  They don't see the relevance of it.  So I think we have to recognize there will always be some people, even when you make the Internet available and cheap, they will still not want to use it (this is Simon).  And one of the things I think has come out is a sense of actually peer-to-peer engagement in learning is fundamental.  It's not about representatives from Facebook, Google, or anyone else coming in.  It's about people like them showing them how to use the Internet that can be very helpful.  Also we shouldn't lose sight of Internet safety and some of the things that we've gotten used to in -- in our own countries in terms of these debates, actually can be quite fearful in environments where the Internet is very new.  Therefore contrary to what somebody said, there may be a notion of an anchor tenant service that is used to try and entice people into the Internet, where actually they're not necessarily encouraged to explore too widely initially, might be a good way of -- and I think, of course, being from Facebook, I think Facebook is potentially that environment, but there are certainly other anchor tenant services that we might use to encourage people to use the Internet.  We shouldn't discount it.  And we should explore different pilot exercises like that.  I know to try and tackle that relevance of fearfulness factor as well about access.

   >> Great point.  I think the final question, then, Pepper?

   >> I wanted to respond to the point about taxes.  Right?  So one of the things that the UN broadband commission has recommended is that this sector should not be subject to luxury taxes.  There will be taxes.  It's not about no taxes because minister, as you point out, says he, yes, governments need taxes, but taxes on the sector above normal are counterproductive to the goals of affordability and connecting people.  This is a sector that's characterized in too many countries by taxes that are equivalent to sin taxes, the taxes imposed on alcohol and tobacco.  This is not a sin, all right?  We should be promoting affordability by lowering taxes but taxes are a reality of life.  They just shouldn't be so high as to -- to raise prices, create barriers and be counter-productive to the goal of affordability.

   >> Mr. Chair, I'm removing the Web foundation heart and I'm wearing my normal scarf, African scarf.  I was in New York during the UN week.  I bought two laptops.  I come from New York.  And I hate the country I live in.  It's a West African country and I -- two laptops and this old one.  So I go to pay the -- is it taxes you pay?  Yeah, customs.  It is 26%.  I have the receipt.  And the man looks at me and says -- actually I bought those cheap laptops that were on sale, and it still -- he still says he does not believe the receipt I have given to him is authentic from Best Buy.  So instead of evaluating the computers at $700, he's taken the liberty upon himself to say the computers are worth $1,000.  And 26% is $260 for laptops that were worth $570 at the most.  I have the receipt.  It's in my name.  And at that same place where they are charging me and I felt like strangling someone, some lady had come in with a printer and had declared the printer as office material.  And the guys were saying, you did not declare the ink in the printer, and so we are going to charge you for what -- something crazy.  And then I left my -- I blasted this guy.  Have you seen a printer that does not come with an ink?  Thank you.

   >> Does anyone have a final -- it's going to be hard to top, but does anyone have a final (laughter) -- a final word?  Martin?

   >> MARTIN LEVY: I have to do this if you've been wondering why I've been hitting the keys, but the under sea cable that just turned on from Tonga to Fiji had no Wikipedia page.  I have to throw things out in the public and it's there now, in the Pacific fiber categories and has its own Wikipedia page so hopefully more people will know about it.  Maybe it the get more use.  Hopefully the price will go down.  Use and price are related.

   >> On that positive outcome, please join me in thanking the panel for an excellent discussion.

   (Applause)