IGF 2017 - Day 4 - Room XXIII - NRI Collaborative Session: Working Together on National Regional Level to Encourage IPv6 Deployment


The following are the outputs of the real-time captioning taken during the Twelfth Annual Meeting of the Internet Governance Forum (IGF) in Geneva, Switzerland, from 17 to 21 December 2017. 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 event, but should not be treated as an authoritative record. 



>> MODERATOR: Okay.  Welcome to the IPv6 panel.  Does this work?  Yeah.  My name is Jan Gorshi.  I work for Internet Society and I will chair this panel.  So in IGF 2015 and 2016, there were two outputs and two IPv6 best current practices documents that were intended to also bring the IPv6 information to the IGF community, how no ‑‑ how to implement IGF, what are the incentives to implement IPv6 and this panel is all about sort of the feedback from different national, or regional IGFs from the world.  How these documents basically improve the IPv6 in different parts of the world.

And let me introduce our panelists.  First we have Mr. Oscar Robles.  He's representing the Mexico IGF.  No?  Well, that was written.  Sorry.

Then we have Mr. Tsuyoshi Kinoshita and Mr. Erik Huizer from Netherlands and Alan Barrett, representing the Kenya IGF.  And I would like to invite first our panelists to present their short presentations and then we can go into ‑‑ into the discussion and we will start with Mr. Robles.

Oscar Robles has been involved in the Internet institutions in the LAC region.  And since 2015, he's executive director of LACNIC, the IRA for Latin American countries and parts the Caribbean.

Please, the floor is yours.

>> OSCAR ROBLES: Can you listen?

>> You need to say next slide, please.

>> OSCAR ROBLES: So I'm here on behalf of LACNIC and, of course, I'm ‑‑ I'm not representing IGF in Mexico.  I'm not part of that initiative anymore.  I used to be but I'm not anymore.  I have information about the IPv6 deployment in LACNIC region, which also covers Mexico as well.  The next slide, please.  What we have found a year ago, almost a year ago is that we have a very low IPv6 deployment in pack nick region.  It was around 2% of the total traffic, it was IPv6 traffic.

The situation was that we were doing some training in the regions since 2005, so we trained more than 5,000 professionals in IPv6 in the region during those 12 years, and just in 2016, there was six in training workshops on IPv6, physical workshops with more than 1500 participants and eight online ‑‑ online courses on IPv6, basic and advanced with more than 1400 participants around 1400 participants.

So we were wondering what else we should do in order to improve the IPv6 deployment, because from our point of view, we were doing all the training that was required in the region.

Next slide, please.

So the next question was about IPv6's distribution.  Just wondering if that could be a reason for not having IPv6 deployed, and the LACNIC region is ‑‑ in the LACNIC region being currently, most of the operators have received IPv6's space, around 90% of them have IPv6 allocated, assigned to them.

Next one, please.

So another thing is that we are looking for how ready was the region to ‑‑ to articulate and to translate the IPv6 traffic, the networks of the region and we found that the region was doing very well, actually very well compared to the rest of the world, more than 30%, December of last year.  We have 33%, one‑third of the networks were already enabled to the IPv6 traffic.  That was compared with the 22% of the rest of the world.  So out of the 5500 networks in the region were ‑‑ 33% of them were already able to transit IPv6 traffic that was a very good number, and that was the part of the understanding of what was happening in the region.

This number, 30%, has grown, now to 37%.  So it's still growing and that can give you an idea of how ready is the world.  Next slide, please.

In 2015, we finished an analysis with the bank for Latin American development called CAF.  Together we this analysis on the IPv6 for social, for the lack of deployment in the region.  And if you can check it out afterwards, but the main reason for the operators mentioned for not deploying IPv6 was it doesn't make business case.  Okay?

So we already think that was the reason, but we wanted to make sure that we have some real information, rather hand our haunch.  So next slide, please.

We started in 2016, a different approach, an additional approach.  We are still doing online training, with the IPv6, but we started a different approach, taking with it, and talking with decision makers in the region.  Basically government officials and telecom operators.  Those, of course, we ‑‑ we current talk with them, with the technical ‑‑ technical speeches, but rather more strategic and policy‑oriented discussions.

So the first message was that the IPv6 topic was not the technological aspect anymore, but a strategic one.  We shouldn't approach them in ‑‑ we should ‑‑ we started the approach in a nontechnical language.

And we mentioned the opportunities loss and the risk they may face in the future if the IPv6 wasn't going to be deployed.

One piece of information, one piece of information that was relevant for them is that what's in it for every player in every.  What's in it for the academy?  What's in it for the players?  What's in it for the users and for the government?  And everybody has a chance to get something to ‑‑ to get some benefits out of the IPv6 deployment and they were very keen to know those benefits.  We didn't start with the least IPv6 readiness countries, but with those least developed countries.

Chile has less than 6%, but we don't think we should be talking about the Chilean government and the representatives.  They will have IPv6 deployed in the next two or three years without any problems.  Those are not the ones we are approaching, but the Least Developed Countries like in the Caribbean and Central America.  So we started and made these approaches, talking with several people, seven, eight, meetings from legislative ‑‑ legislative bodies to vice president.  And so that was a very different approach and raising the relevance, the strategic relevance of IPv6 deploy.

Next slide.  Last one, please.

So this is the summary of messages that we delivers to the decision makers that ‑‑ why was IPv6 needed for strategic reasons, for the country.  And the five messages in summary that they provide.

IPv4 and CGN use may prevent the capacity to map ‑‑ to a single user or a small group of possible users making the grand decision and the more complex task.  We will need to connect 100% of the population, and we need to take advantage ‑‑ I mean, we, as the country, we need to take advantage of technologies like smart cities and full promise of Internet of Things.

And both needs will require massive numbers of IP addresses, internet addresses and IPv6 is the only sustainable solution to the previous challenges.  So this is the basic messages in a nontechnical language that we provide, and we haven't ‑‑ we are ‑‑ this is very early to know that the results of this conversations, but this is what we are doing in 2017, and what we are going to do in 2018 with a different set of countries, but you I still think ‑‑ we think that this is the right approach at this moment, after we have provided all of these technical training in the region.

So I think that's it for now.  And good.

The next one, please.

Yes, that's it.  Thank you.

>> MODERATOR: Okay.  Oscar, thank you very much for your nice presentation.

Our next speaker is Mr. Tsuyoshi Kinoshita, he's with the Japanese community.  He served in various Japanese policy technology development Study Groups regarding the Internet policy IPv6 IoT green cybersecurity.  He held a variety of leadership positions in sales, marketing and business development in the Asia Pacific region and Cisco.  Please, Tsuyoshi, the floor is yours.

>> TSUYOSHI KINOSHITA: Thank you very much for having me with this distinguished panelist.  I'm honored to be part of this last session.

Can I have my presentation up on the screen?

Nevertheless, as Jan said, I have been partially involved in the department of IPv6, almost from the day one, the 20 years.  So I have seen ups and downs and many ups and downs, I should say.  So I ‑‑ I'm actually very happy to have this opportunity to provide our background from Japan.

First of all, many of you were attending the IGF last year and that document was actually introduced formally.  And that the ‑‑ since then, I'm sure it was touched on briefly, that we have seen quite significant progress being made by some of the countries before I actually get into that, the Japan perspective.  And the reason why I'm showing this is that I think that after two decades of the IPv6 initiative, what the industry and the community has kicked up, we are actually entering ‑‑ definitely entering into the tipping point, actually.  Now the adoption of IPv6 is now entering into the main stream and few countries I highlighted, Uruguay, and Brazil, are really none straiting how the global south is now the front‑runner in terms of driving and adapting the IPv6 we have seen the IPv6 adoptions pretty much around developing count, the global north, if I may say so.  But in the last 12 months, we are now observing that it's not about the adoption driven by the global north but the south is now taking the front seat.

And this is very much relevant to the ‑‑ the next SDG agenda, that we are going to expect that there are new Internet users, it's going to really be on the Internet based on the IPv6 going forward.

So on the Japan side.  I have should actually talk about Japan.  In the last 12 months, here's what we have done from Japan perspective, but as I briefly mentioned, that the ‑‑ and some of you know that Japan has been on the ‑‑ this long journey of the IPv6 deployment, and adoption since day one.  Our community got involved heavily in the technical standard development in the late '90s and we have also had the government initiative to really drive the adoption of IPv6 as a new infrastructure for the digital age.

So we have really gone through the ‑‑ I mean, really many long journey over the last 20 years, and I would also like to actually emphasize, it's not only for us in Japan, but we are pretty committed to the global community and also the region, so that we are pretty much enjoining the best practice sharing, not only within the nation, but also outside the country to date.

So to be specific on the ‑‑ on what has happened in the last 12 months, notably, we have had a couple of significant updates in terms of how the ‑‑ the IPv6 are getting further adopted or adoption of IPv6 got accelerated.  The one most importantly to highlight here is the mobile operator.  The IPv6 initiative to date has been centered around the wire line infrastructure, the ISP and the backbone providers and the wire line operators, their infrastructure.

Now in the last six months I would say, in 2017, we were very happy to see that all the mobile operators, the three major mobile operators we have in Japan, stepped up and now they have put IPv6 as a default as IP services on their mobile infrastructure.

So we think that we are on the right course of action in terms of really driving adoption of IPv6 throughout the older infrastructures involved, but with the new initiative seen from a mobile operator, we think the adoption rate will really pick up in the next coming few years.

So that's one.

The other stuff, we also like to call it out is content providers.  We are finally also seeing not only the infrastructure players but the content players getting pressure, because of the movement from the big giants such as the Google and Facebook and so on, who really made IPv6 stand.  Local content providers are now also seeing that they now need to move in terms of getting IPv6 ready.

So this is another positive sign that we have seen in the last six months, since it was updated last December in Mexico.

And that's pretty much it.  So let me move on to the next slide.

So one of the things that we may also share, why Japan has been the front‑runner in driving and adopting IPv6, there are two secret success, if I may say so, which I would like to share.  One is that we have very strong tech community.  So we have ‑‑ we are very fortunate that before facing the need of capacity building to deal with new technology adoption, we have those very strong tech communities, really staying on top of the IPv6.  That's one.  But more ‑‑ more importantly, I think that as a country, some regions could also somewhat have a look at the best practice from Japan, is the fact that we do have the IPv6 Study Group framework installed by the government initiative.  They are not only the government possibly going to the meetings, but the academia and the private sector and sometimes the citizens, the stakeholders who are invited in the Study Group, and the meetings, and what we do is basically as a collective group of the IPv6 concerns the stakeholders, we basically review which part is actually really making progress and which part of the critical infrastructure or the services have to take the action to really drive the nationwide adoption.

We have had 37 meetings since this meeting got installed back in 2009, and for four official reports published with two Internet progress reports and what we do is basically calling out what are the current nature of IPv6 adoption, but also encouraging and sometimes reinforcing the necessary action to the concerned stakeholders what needs to be done.

The other ‑‑ the mobile operators, the IPv6, the default standard began in the middle of this year is indeed the result of this Study Group's recommendation.

So we have a mechanism to really calling out and actions that concern the bodies to ensure that they will execute on the expected actions accordingly and we do the check and balance type of mechanism under this IPv6 Study Group.  So this is ‑‑ I would say, the biggest reasons why Japan has not lost interest or motivation in driving IPv6 since the late '90s.

Go to the next slide.  So another interesting or best practices, I would also like to call it out, is that how the government is actually taking their role model in adopting IPv6 is a bit hard to see on the screen, but this it actually the adoption rate observed in the last, I mean, five plus years from the government agencies, how the government agencies website, getting IPv6 ready and also getting the systems ready.  And we are very happy to see that, for example, the website of the central government agencies are now about 50% IPv6 ready.  While the industry, the website are about 20 plus percentage.  Of course size of the install base are quite different, but the government is actually demonstrating that it's not only the infrastructure to be blocked out in terms of IPv6 adoption but content sites also are getting ready.  We are happy to see that that demonstration is helping out the outside government to be aware of their necessity of IPv6.

So maybe next slide.  My last slide.

So my last message, that message is that we ‑‑ as I said earlier, that we are really coming into the mainstream of IPv6 adoption.  Not only from Japan perspective and making this statement, but actually the timing of the ‑‑ this main stream adoption is somewhat viewed as a perfect storm, meaning imagine countries need the IPv6 space, which are definitely going to be based on IPv6.  But finally after 20 years of the dream, we finally have an application for IPv6, which is the Internet of Things.  That's in relation to the user of the Internet, we actually do go in to see new users, Internet of things will require new addresses.  We do see that the adoption now taking off now is the best timing from a variety of other perspective.

That said, the work is not done yet.  We need to continuously work together to really make the shifting, to more than v6 space network or their system.  So I would expect that we are going to continue to have this kind of forum at upcoming IGF in years and that's pretty much what I wanted to say for now.

Thank you.

>> MODERATOR: Thank you Kinoshita‑San.  It's nice to hear from the progression from your side.

So now we have Mr. Erik Huizer from the Netherlands.  Let's see what's happening in the Netherlands.  So Erik Huizer is a CEO and he's the chair of the IPv6 task force.  He's involved in the Internet Governance.  He's an induct in the Internet hall of fame.

Please, the floor is yours.

>> ERIK HUIZER: Thank you, Jan.

I would like to share a couple of lessons learned with you from the Netherlands.  I have left out most of the things that you have heard from the previous speakers.  We have ‑‑ I can confirm their lessons learned because they are similar and we have taken similar measures.  I tried to go a little bit beyond that.

Next slide, please.

First of all, IPv6 is not necessary in the current Internet, of course.  We can easily do with a two bit address space for everybody's only going to Google and Facebook anyway.

Next slide, please.  So before I start on the lessons learned, it's just incredible.  How many nontechnical people come up to me on the IPv6 and try to tell me what the right solution would be.  You know?  And they have no technical clue whatsoever and they come up to me and they tell me how ‑‑ you know, how you can fix IPv6 or IPv4 for all that matter.

And usually, a lot of criticism is about the migration or the lack of migration and, et cetera.  So I would like to show this slide by Jeff Houston because I think this depicts the original plan, and the original plan, I would like to point out, was made in 1995.  Right?

So that is quite a bit of time ago.  We would start to deploy IPv6 which is the red line and the Internet would grow along the green line and the IPv6 ‑‑ the IPv4 would deplete along the blue line and that we would have enough IPv4 addresses to complete the full migration to IPv6.

Well, of course a lot of things happened.  First of all, the green line went berserk and almost went vertically through the roof and the red line stayed almost horizontal.  And so the two are not yet meeting and we have depleted our IPv4 pool.

Next slide, please.

So this is what I want to talk to you about, that IPv6 is like an emergency exit.  It lacks a deadline.  It's costly, if you wait too long.  It suffers from mips.  It doesn't need just naming but it needs shaming to force people to implement it, and dual stack, dual stack, dual stack, or you will run into problems.

Why is IPv6 like an emergency exit?  It's like an emergency exit because if you talk to someone who operates a restaurant or a cafe or a discotheque and you tell them, you have to implement an emergency exit, they are really not very happy.  They have know this he have to do it.  They know it's unavoidable but it won't get them any extra customers.  There's no business case for the emergency exit.  And the same goes for IPv6.  You know you have to do it, bust you also know that it won't attract any extra people, well, at least not in the short run.

Next slide, please.

And the problem with IPv6 is it lacks a nice deadline.  I wish it had a deadline, like the year 2000 problem, because then everybody would really get nervous, and what you see now is that it gets on to the priority lists of larger companies, ISPs, et cetera, but then other issues come up, and the priority is never high enough for IPv6.  So it always drops to the bottom, because there is no deadline and you can always postpone implementing IPv6.

And one of the really frustrating issues we had in the Netherlands, Vodafone has been changing ownership notable times over the last ten years and every time it changes ownership, the new owners want to clean out the whole priority list and, of course, IPv6 drops back to the bottom.

Next slide, please.

The rumor has it that IPv6 is very costly, and I recently was involved with an organization and it was really costly, but they were forced to go to IPv6 because they really needed addresses and they run out of IPv4 and there was no way they could get more.  So they had to do almost an instant transition to IPv6 and, yes, that is very costly.

However, I will show you at the end, an example of a well prepared IPv6 migration, an organization that planned ahead, and replaced all the equipment when it needed to replacing and made sure that the new equipment accepted IPv6 and they gradually implemented all the security stuff around the IPv6.  They trained the people.  They did it well he prepared and then costs retained to a minimum.

So don't wait too long.  And if you are only starting now, you have waited too long.

Next slide, please.

IPv6 suffers a lot from mips.  So you have someone come in, someone from the IPv6 task force explains what they have to do.  People go off and they Google and they come one a lot of myths.  One myth is, I have IPv4 addresses and I don't need IPv6.  It's a stupid way of reasoning.  I'm on at old Internet and I don't care if anybody is on the new Internet and I can't communicate with the new.

Or even worse, NAT is the main security feature.  We lose our security feature.  If your security feature is dependent on, that you have a bigger problem than just IPv6.

If you really think NAT is a great tool, you also have a bigger problem than IPv6.

IPv6 is like the Y2K problem.  It doesn't exist.  Well, wake up!  It does.  IPv6 is complicated.  Well, just as complicated as IPv4 and you have to calculate in 128 bits, but rest assured, it takes one hour to learn how to deal with 128 bits.

IPv6 is slower than IPv4.  That's a really nice one.  Yes, it takes probably a fraction of a fraction of a fraction of a millisecond longer to transport 128 bits compared to 32 bits.  So maybe it's slower but I don't think you will notice.

Dual stack means double the amount of work.  No, it doesn't.  There's no other argument there.

It's better to wait until everybody else deploys IPv6.  This comes from people who never install the latest version of IOS.  They wait until that version has already been in existence for two years before they deploy that and in the meantime, they have a lot of bugs and make them vulnerable to security.

Next slide, please.  We tried to promote IPv6.  We put everybody in the spotlight who did wonderful IPv6.  We worked with our government and this statement by the way is not even of my government.  The statement that you see there is from the European Union, and it's coming from Commissioner Cruise at the time and all the affirmative action from the government, from the IPv6 task force, it's not big enough.  It didn't work.

We bought Internet.nl.  I don't know if you have online access.  If you do, click on the Internet.nl.  Internet.nl is a website that allows you to test your domain.

Or you can test your website or your email address and it will text it for all the modern up to date Internet standards.  So it will tell you if you have IPv6 connectivity, which we don't in this case.  It will tell you whether we ‑‑ whether you have DNS sec.  It will tell you if your DNS is appropriately configured.  And it will also shame you if you don't have those things implemented.

>> All green.

>> ERIK HUIZER: I wouldn't expect anything from you, Jan.

This is not only useable in the Netherlands, and you can use this all through the world and I encourage you to use that tool and to show organizations where they are failing in implementing their stuff.

I recently started at Jeoon.  This has been used to promote stuff a lot better.  Back to the slides, please.  Next slide.  Many of the ISPs are now ready and deploying IPv6.  That is not to say that all the citizens in the Netherlands have already got an IPv6 and we are not very high yet on Google's rankings on IPv6, but our major ISPs are now IPv6 prepared that.

Means that every new customer that gets connected gets IPv6.  The old customers get IPv6 as soon as they replace their modems because replacing the modems is an expensive thing.

However, several ISPs are using DS light and the problem with DS light is that it gives a lot of problems.  You get NAT at your home address, plus another NAT at a carrier level, which makes two NATs and one NAT is problem enough.  Two NATs on top of each other is way more problem than most people can handle.  And, unfortunately, you get the DS light as soon as you do IPv6.

So people associate the problems with IPv6 and people start complaining that IPv6 is not working.  Well, IPv6 is working perfectly.  It's IPv4 that sucks because you have DS light.

So try to avoid DS light.  Do the whole dual stack.

Next slide, please.

So I finish with a very positive example.  This is an article in Dutch, unfortunately.  They looked at the cost and the costs weren't too high and all in all, they said the whole thing to implement IPv6 was really smooth and easy and I hope this article gets translated because it's really positive in that it's not that much effort if you just prepare yourself well.

Thank you very much.

>> MODERATOR: Thank you very much, Erik.  I think Netherland.nl is very useful to do some testing.  How are we doing with time?  I think we are doing quite well.

The next speaker is Alain Duran of ICANN and IPv6 pioneer since 1993 and worked with Comcast in the US, in 2005 to 200, and chair of several IPv6‑related Working Groups at the ITF and many groups.  It's looking at the evolution of the IPv4 market.  Tell us a little bit more about the global picture.

>> ALAIN DURAN:  So as you mentioned, I have been working since early 1993 and since everybody asks me, when are we going to see IPv6 and the answer is, well, in two years.  So it has been now 24, 25 years.  Hopefully it's going to be soon.  Are we there yet?  That's really the question that I went on to try to answer.

So I did some work with Jeff Houston from APNIC and we are correcting a lot of measurements.  Jeff is collecting about 6 to 10 million measurements a day, all around the world.  Through the Google system.  It's targeting users and presenting users with an IPv6 address and an IPv4 address and try to see if users can reach both or just one or the other.

And what's interesting with the system, is it gives the user perspective, actually people use.

If things have been done, like service provider I have done some work, but they have not done it on or not make it available, then it doesn't make any difference to the users.  So that's really a user centric perspective.

The flip side of the story is that it does not capture machine‑to‑machine communication.  Anything from society is not capturing the statistics.  So next slide, please.

So when we look at IPv6, we can do two things.  We can look at the ratio of IPv6 of over population and that gives the penetration of IPv6 in the country.  Oops.

Or we can look at the ratio of IPv6 over IPv4, and see how many of the users have actually migrated to IPv6.  It's slightly different picture.  So I have done both in this study.  So let's look at the first one.  So it's looking for over ‑‑ or most of the countries, so the first question is how do you define a country?  And it's not actually a very easy question to answer.  So I came up with my own definition.  I took all the countries that have a country code a two‑letter country code but has a population that's larger than 1 million.

Why do I do that?  It's because the statistics that we use about populations are ‑‑ well, the data is not very clean and the country, the less reliable the information about population is.

So if we take countries that have at least 1 million people in there, we have higher likelihood to have better data.

So I look at the total IPv6 adoption in 2017 and 2017, the data is from yesterday.  So what we see is that there are countries that have been accelerating in IPv6 over the last four years.  So the study goes down from 2014, up to now, and we have some countries that have not been so active.  So essentially, I clustered them into three categories.  The countries that have more than 5%.  We can say, yes, this is happening.

The countries which are between 1 and 5, and they are starting to bubble up on the radar and countries below 1%.

So in this view, we have 158 countries.  Yes, 158 countries.  103 are still below the radar.  They are still below 1%.

And that is worrisome.  The other side is we have 31 countries that are above 5% and if you look at them, year after year after year, those are the same countries that have been increasing the IPv6 penetration.

So this starts to paint a picture where the left‑hand side of this graph was 31 countries, are actively doing this and in a few years, they will be on course to get to full IPv6 deployment.

But open the right‑hand side, not so much.  Are we going to go in a world where we have two Internets, 31 country Internet and the rest?  And that's a question, but I'm asking.

So next slide, please.  And then exactly the same calculation now, using the ratio of v6 over v4, and it's slightly more ‑‑ now we have 37 countries but really under the radar almost 5%.  But the overall picture is the same.  We still have more than 100 countries that don't show up on the radar at all.  So it's interesting to look at what is top 37 countries.  So the next slide.  I will go straight into the following one.  The next one, please.


So this ‑‑ I have population of IPv6.  So if you look at this one, what we see is ‑‑ well, the first thing that strikes me is there's a negative.  So there are places where we see a decrease of IPv6.  So I'm still studying this to try figure out, is it a glitch because some of the data that we have is still dirty and we need to clean it up a little bit more?  Or is it really a decrease.

Sometimes a decrease could an increase in disguise.  For example, if we compute the ratio of IPv6 over IPv4 and your IPv6 stays the same but your IPv6 increase, ratio will decrease, right?

So I don't necessarily want to look at this as people stop doing v6 and remove IPv6.  That's not necessarily the case.  We need to be more careful with that.

Putting aside a decrease may or may not be a real decrease.  What is interesting no me is are we doing better or not?  So if we look at some countries like India, for example, the he can one in there, there we start in 2016 and 2017, it just exploded.

If you look at countries like Belgium, next to it, well, Belgium started earlier.  So 2014 includes things that were done earlier.

But they are now reaching about 60, 70% and we are already in the phase of the curve where things are not accelerate as much.  They are flattening an plateau.  But there are some countries like Peru, where we see a lot in '14 and a lot in '15 and then not so much.

So in the next part of research, my next phase of this research, I want to go in this also of 37 countries and really understand what is the landscape there.  I was wire line providers that are driving this, are motivated by what government mandate may have been there or not.  Why is it that we reached this plateau in some cases?  Is it the only one of maybe four providers in the country did that and the other three or four did not do it?

And what was the motivation behind this?  Next slide.

So this is the same thing looking at the 1 to 5%, and same observation, we see some countries that did a little bit last year, or some countries that did a little bit year or two years or three years ago but haven't agreed on much since.

So I would like to go back to the ‑‑ one back.  One back.  Thank you.

So I want to conclude on this, and the optimist part of me is looking at the left side of the graph and says that our countries are really, really looking good and great course.  And not so optimistic side of me is looking at a new divide and I don't like that.

And when I do this graph again, the red will go further to the right and less and less countries that are not showing up.

Thank you.

>> MODERATOR: Thank you, Alain.  This was a nice global overview of what is going on and we all want that red line to be as further right as possible.

Already.  Mr. Alan Barrett from Africa.  Alan Barrett is the CEO of Africa.

We have two CEOs of two IRAs.  He's involved with technical and policy of the interknit since the early '90s.  Please, Alan, the floor is yours.

>> ALAN BARRETT: Okay.  So thank you.  I'm here in my role as representing AFRINIC.  On the agenda, you will see that it's the Kenya IGF and that that's not really the case.  The speaker from Kenya wasn't able to come.  So basically I'm filling in for him.

So I see my slides are up there.  Okay.  Next slide, please.

AFRINIC ‑‑ sorry, Africa has a fairly he low Internet penetration rate, and a lot of use of IPv4 with network address translation, so that multiple users share the same IP address and I'm also very low IPv6 penetration so I hope we can improve that.

So the population is growing faster than the rest of the world.  You probably can't read the legend, but the green line at the top is the world population of Asia, shows that it would ‑‑ it's projected to increase from now until about 2055 and then maybe it will start reducing or staying more or less the same.

The gnat lines near the bottom are Europe, North America, Latin America, the Caribbean and Oceania, showing that their population is projected to stay fairly stable over the next 50 years, from now until maybe 9 end of the century.

Whereas the purple line going up there is the population of Africa, projected to increase from about 1.2 billion now to possibly more than 4 billion by the end of the century.

Now, of course, these are just projections.  They could change.  Nobody knows exactly what the population is going to be.  But it is interesting that Africa's population is growing so fast, and that will bring an increased demand and so there's a lot of business opportunities there and businesses providing Internet access can take advantage of the increased demand from population growth to increase their business.

Now, currently, Internet penetration in Africa is ‑‑ excuse me, about 31%, and that's definitely growing, and possibly in the next three years, we may have to increase to more than 50%.

Next slide, please.  We are, of course, like the rest of the world, running out of IPv4 addresses in Africa.  There's one point per year from 2012 to 2016, showing from how 2012, the amount of IPv4 in the AFRINIC pool.  So what AFRINIC had available to allocate to our members throughout Africa has decreased from about 4/8 in 2012, to about 1.5/8 at the end of 2016 and then on the next slide, we zoom in a bit ‑‑ oh, no, we don't.  We just add one point to 2017.

Now, in early 2017, we crossed that line of having less than one/eight available and when that happened, the ‑‑ what we call the soft landing policy took effect, and under the terms of the soft landing policy, AFRINIC members, who are mostly ISPs have to try harder to qualify for additional allocations of IPv4 space and we expect that we'll cross another boundary of having less than one‑quarter, less than 1.25/8 available sometime next year and when that happens, the rules will change again, and AFRINIC means, they will only be able to get us 1/22.  That's about 1,000 IPv4 addresses at a time.

Now open the next slide ‑‑ so we are rapidly running out of IPv4 space but we have got plenty of IPv6 space.  More than 40% of the AFRINIC members have some IPv6 space allocated, and they can easily come to AFRINIC.  Those who don't, can easily come to AFRINIC and request IPv6 space.  It's very easy.  The amount of justification that they need to receive, say, /32 is minimal, and there's also no extra charge for that.  The annual membership fee for a member who has some IPv4 space and some IPv6 will basically be the same as for the same member if they didn't have any IPv6.  So there shouldn't be a financial barrier to receiving IPv6 space from AFRINIC.

Only 41% of our members have IPv6 space and only half of those, only about 18% of the networks in Africa, which is let's say about half of the 41%, are announcing their IPv6 space in the BGP routing table.

Of those, very few are actually using it.  So that the situation in Africa looks like much lower deployment.

We see in Alian Duran, and whereas in Africa, the highest is less than 10%.  So this graphic, you can hardly see it there, but most of Africa is shaded white, which means very little IPv6 usage.  Zimbabwe is the only one that looks a little bit screen there.  And it has IPv6 usage of between 7% and 9% depending on whose measures you believe.

And that's the leader in Africa.

In second place we have Egypt, between 1.4, to 1.4 IPv6 penetration.

Next slide, please.  This diagram shows the darkness of the green shading.  It shows the ratio of IPv6 prefixes that are rooted in PGP, and it's allocated.  So the leader there is Mauritius.  And the 68% of the IPv6 are actually announced in BGP.

Tanzania, the green blob in the middle of the East Coast, it's a little bit behind at 56%.  And, Zimbabwe, which showed as the best in the previous diagram, is only 42% here.  And the reason for that is there are many Internet service providers in Zimbabwe, and one of the big ones turned on IPv6 and that pushed the country's IPv6 usage up, but many others have not turned on IPv6, and so that puts the ratio of rooted space versus allocated space.  It reduces that ratio down to 42%.

You see a few more green countries here in the diagram, but there's a lot of white, or very green, and those are countries where either there's no IPv6 at all, or the IPv6 as been allocated to the ISPs by the registry, but is not being used, not being announced in BGP.

Next slide, and there are some good news cases, however.  Here is South Africa.  The black line shows the number of allocated IPv6 prefixes and you can see it goes ‑‑ it goes over about a three‑year ‑‑ a three or four‑year period from the middle of 2012 on the left‑hand side of the graph to about the third quarter of 2017 on the right‑hand side of the graph.  And so the units on the Y axis are in number of prefixed which closely corresponds to number of Internet service providers.  Typically they only need one prefix each.  The IPv6 is enough to serve their needs unlike v4, where you need many IPv4 prefixes.

So we see over the last four years or so, the number of IPv6 prefixes in South Africa has increased from about 85 to more than 200.  That's definitely good news.  That's the black line.

The green line near the bottom, it doesn't look so good, though.  That's the prefixes that are actually alive, where we can see some actual usage coming from those prefixes.  They are not just allocated sitting on the shelf somewhere, bust they are actually being used.  And there we see ‑‑ sure, it's increased over the last four years from, oh, from 10 to 20 or 25, but that's still not looking very good.  There's room for much more improvement.  On the next slide, this is Zimbabwe, and we are looking at the top 500 websites in Zimbabwe.

And, oh, dear, the bad if yous is that only 100.  Top 500 websites in Zimbabwe are IPv6 enabled.  That's bad news, I guess, if it's only 100 out of 500.  But if we look at that green line ‑‑ a year ago, it's in the middle of 2016, it was only 50.  So over the past year, well, let's say the past 18 months, the number of v6 enabled websites in Zimbabwe, shown by the green line has increased from 50 to more than 100.

So there's some progress there but I wish it was faster.  On the next slide, we see the same kind of data for Egypt, a more rapid increase.  So, again, we have about an 18 month period ‑‑ oh, sorry, you know, this one the X axis is longer.  It's about a four‑year period.

We show the top 500 websites in Egypt, and now about 150 of them are IPv6 enabled, and there was a big jump earlier this year, when it increased from 100 to 150.  I'm not sure the reason for that.  Why did they all suddenly enable v6 all at once.  It could be something as simple as one of the ISPs turned on v6 and all the web servers behind that ISP were able to use v6.

The next slide, this is the result of a survey that we did.  AFRINIC survey, the network operators in Africa, asking them do you have IPv6 or maybe you are working towards it.  It's in progress, but it's not done yet.  Maybe you are only planning it.  And we did this for several different types of networks, data centers, university campuses, enterprises, Internet service providers and the last category is just other unspecified.

So we see here if you can interpret all the different colors, that 34.5% of the Internet service provider networks are at least doing ‑‑ sorry, no, I got that wrong.

About 34.5% of the networks that we surveyed were from Internet service providers.  About 11% ‑‑ 11.9% of the networks that responded to the survey were enterprises, and about 14% of the networks that responded to the survey were campus networks, and about 5.7% were data centers.

Now ‑‑ and the last 33% were other.  Now, within each of those categories, we have broken out the results using shading.  So if we look at ‑‑ let's say, the orange ones at the bottom.  Okay?  34.5% of the networks we surveyed were service providers.  And looking at the shading, of those, about 40% of them say that they have done their IPv6.  It's done.

Another 10% or so says it's in progress, and it looks like another 40% says we're thinking about it.  We are planning.  And finally, another 10% say they are testing.

So there is some progress there, at least people are thinking about it or planning, but the deployment is a lot lower than I would hope.

Okay.  Now, the same survey, we asked people, what are the challenges.  Why don't you have IPv6?  And the answers we get when we ask managers are different than the answers you get when you ask engineers.

When we ask managers, nearly 20% of them says we lack confidence that we can handle IPv6; whereas for the engineers only 13.5% said that.  A lot of engineers looking near the bottom, there's the second red line from the bottom there, 16% ‑‑ 16.4% of the engineers said our management doesn't really support us in deploying IPv6.  Whereas only 6% of the managers said our management doesn't support IPv6.

A lot of them said it's not a priority on both sides, about 11.6% ‑‑ about 4% of managers said it's not a priority.  A lot of them are saying that knowledge and skill are the biggest reason why they are not deploying.  Nearly 18% of managers and nearly 16% of engineers say our knowledge and skill, that's the biggest reason why we are not deploying IPv6.

Next slide, what is AFRINIC doing?  We give training on IPv6 and this year, 2017, we have taught 20 workshops in 18 countries and we have trained more than 600 engineers but that's not enough.  We need those engineers to go off and use their training and deploy IPv6.

We have also set up a certification platform called 36, where we were offering exams.  You can ‑‑ you can pass an online course and receive a certificate which is accredited by the world IPv6 forum and we have signed up some partners to help us administer the exams but it hasn't really taken off yet.

Okay.  That's the end of my presentation.

>> MODERATOR: Thank you.  That was a nice overview.

I would like to open the floor if there are any questions.

>> PARTICIPANT: Yes, this Alan Alina from Africa.  So I would first direct my question to Alain.  The survey said that the ingenuous are complaining about getting support from the managers, but you are saying that you will be training engineers.  So maybe it's time to train the managers.  So is it something that AFRINIC is planning to do in the short term?

Maybe the other panelists can also pep me.  If I got him correctly, he said don't deploy your IPv6 with DS light.  Okay?  So I don't know if I got it very well.  If we don't do DS light, so what is the alternative?  Yeah, he's proposing?

>> MODERATOR: Yeah, unfortunately Mr. Huizer has to leave for the airport to catch a flight.  We have Mr. Alain Duran.  Maybe he can answer.

>> ALAIN DURAN: Okay.  Let me quickly answer the question about training.  Yes, we are using the survey to help us figure out what we should do next, and so as a result of the survey, we will be designing training courses for managers.

It's not yet operational, but we will do it because the survey tells us it's needed.

>> PARTICIPANT: Something question did about ten years ago and the idea is if you have to do IPv4 and IPv6, you are essentially doing more work and you are not really getting a lot of benefits because you still have to do IPv4, right?  That seemed to be an interesting interim solution when we still have IPv4 addresses, but when you don't have IPv4 addresses left, or you only have a few left or not enough, you can't do that any more, right?

And in that case, it's kind of an exit strategy and you need to do something.  As you have seen all the data about the deployment, we are very far from the situation where we can say, oh, just do IPv6 only and nothing else.  That's not a viable approach for most providers or most services.

So you need to do something.  And when you are essentially pushing the caller like this when you have to do something, and the rest of the world doesn't help you, then you can't offer a service that's 100% as good.  And that's really what the situation is.  So it's not so much that there are two NATs there's one NAT.  There's a place where you your share your addresses with other people and sometimes that creates problems.  That's kind of expected because you are painted in a corner where you don't have enough IPv4 addresses and the world has not yet moved to IPv6.

>> If I may add, for mobile networks, the majority of mobile networks is deciding on IPv6 only, plus 464, which is basically the translation.  And the land line provides several.  So there's also another mechanism called A plus B and you have map E and map T as a stateless version of how to share the address and the port.  So you have a variety of tools in the tool box that you can use today, to do this.

The biggest problem is that you need to change the CPE.  With the slides, with A plus B, you need to carefully ‑‑ to carefully man and choose which way you want to go.  But like with anyone from the panel.

>> PANELIST:  I would like to ‑‑ yes, you have many mechanism, especially my fold because I was sharing some of the IGF Working Group.  When we started in the late '90s, it was let the flowers room, and now we have many mechanism and now we created more confusion than anything else.

And that's not necessarily the best place to be.  They are small deltas.  In the end, when it's not enough IPv4 addresses we have to share and when you share, you are not the only ‑‑

>> MODERATOR: Would anyone else like to add?

>> PANELIST:  Looking at the IGF from last year, it's not clear what technologies ‑‑ for the brownfield transition, which ones are good for the greenfield deployment.  So I would suggest that if there is an opportunity to update the BPF in the future we should actually incorporate some of the best practices based upon the situation, what the need from the IPv6 issues people are facing.

>> MODERATOR: Thank you.


We have an online question.

>> Hello, I'm Luca, from Webster University, your online host.  We have a question goes:  He's an individual community member of the community in Japan.  Interesting fact of preservation by Alain Duran on the IPv6 deployment.  I had a chance to talk on local content service providers and they listed a few challenges and possible ways forward.  I wondered if they could help in addressing some of the divides on IPv6.

Number one, access lines or IPv6 ready technically but still bugs in middleware.  If more information on debugging can be shared, it helps.

Question two.

In addition to the announcement by the I. AB, if open source makes IPv6 a default, it could have substantial impact.  Part number three.  Cost and time on training for help desk is often more ‑‑ more challenged in terms of commercial service than technical issues on deploying IPv6.

>> MODERATOR: Okay.  Thank you.  Who would like to take the first two questions.

>> PANELIST:  I will take the third one.  But speaking the different one.  It's a chicken and regular problem.  We have no uses.  We have no reason to put any services and there are no services and that has been essentially solved with Google, Facebook and others went ton deploy IPv6.  What is missing is the local content and in some places we are seeing this happen, and some places less, and some other data that you showed about Zimbabwe, it was inning to see local content there.  That's really one of the key.  But the other thing I want to do is turn this around and say, we are not yet at a point where it makes sense to put your new content on IPv6.  You still have to also provide it on IPv4.

Because if you only provide it on IPv6, then you have only essentially 30 some countries that can reach you and that's not the entire Internet.  If you want to make sure that you have one Internet and not two Internet and you provide content, you cannot do it just over IPv6.

>> MODERATOR: Who would like to take the second question?

No?  Apparently I have to.  Well, the IAB said that IPv6 needs to be the default in the new developments.  Yes, this have ‑‑ this can have the impact but we need to understand that the standards that are coming from the IATF need some time when they are implemented on the devices so it may take some time before we can see this effect.

And on the third question, about the help desks, it's a very good question.  In the right community, we basically brought ‑‑ for ‑‑ how to implement IPv6.  The first one was rite 554 that says ‑‑ it's explaining how to ask for IPv6 capabilities in ICT equipment for the procurement.  The next one, where we actually ‑‑ while talking to the wide world community, where they have problems with IPv6, we heard the common problem is help desk, because they don't want to implement IPv6 because their help desk will just burn down in flames if they do so.

And so we have a document called ripe 631 and it's called the IPv6 troubleshooting for the help desks and that's basically the description to put your to your help desk and they can follow the procedure how to troubleshoot the IPv6.  We also build a tool as a part of that.

It's called IPv6.IPv6.com.  If you are deploying IPv6, go and have a look at the ripe 631 and you will find a very good solution for your help desk.

And I hope this ‑‑ this answers the question.

Are there any other questions from the floor?  Please, don't be shy.  Yes, please gentlemen.

>> PARTICIPANT: My name is Justin.  I'm the director of research and ‑‑ and the authority.  So during this IGF, I found that the government is ‑‑ maybe you will be a user of IPv6, but also a facilitator and the deployment.

In the presentation, which was done I found that the country don't ‑‑ the countries don't have the same level of implementing IPv6 and I would like to know maybe which kind of steps or incentives countries have can follow and put in place, I mean the government to speed up the deployment of IPv6.  Thank you.


>> MODERATOR: Thank you White Sox would like to take the question?

>> There are a few things that governments can do.  I think some of the easy things are putting IPv6 as a requirement in your procurement, and if the government is buying some kind of an IT service, make it a requirement that the service provider must implement IPv6, otherwise, they will not qualify to win the tender process or the bid.

Regulators could make a requirement for licensing.  Often the equipment needs to be approved.  So, you know, if every cell phone has a type of approval and it could be a requirement for type approval that the regulators says we will not approve any cell phones that don't support IPv6.  So just a few examples of things governments could do, without passing any laws, just using the regulatory capabilities they already have or using the financial capabilities they already have to ‑‑ to decide which types of services to provide.

>> MODERATOR: So a couple of things from my point of view, the government can actually set their direction with some time schedule target, with how their country likes to see the adoption of IPv6 to be dealt with.  By having such a nationwide direction, it would help 89 cross industry to be on the same page, otherwise, you will see the different stakeholders taking a different action in terms of how to move towards IPv6 adoption.

So that ‑‑ having a common goal ‑‑ it was set or published by government would definitely help to bring all the stakeholders to be on the same page.

And the other thing I briefly touched on, it is ‑‑ it's also, I think the government's role to facilitate how the different stakeholders to meet regular basis, and then exchange their state over the IPv6 adoption from the different parties.

Otherwise, of course, again, the industry has different agenda.  How each of the articles of industry have a different agenda to actually look at the IPv6 adoption.  So the government could provide a platform and facilitate their discussions long the different stakeholders regarding IPv6 adoption.

>> MODERATOR: Okay.  There are more questions.  We are out of time, but I think we can ‑‑ if you are very, very brief, please.

>> PARTICIPANT: My name is Serge.  From the republic of Congo.  I'm from the IGF in Congo.  So my question goes to Mr. Alan of AFRINIC.  I love the way the survey put things between the view of engineers and managers, and what I want to know, when the engineer says that they don't have the support of the manager, what type of step do you really need to deploy the IPv6?

>> I'm not sure that the survey asked the question in that level detail.  When they said my manager doesn't support it, the common problems is my manager gives me so much other work to do that I don't have time to work on IPv6.  Or my manager won't hire any more people to be trained on IPv6 or my manager won't allow me to go on IPv6 training.  So there can be many reasons that offered foot category that the manager is not supporting the engineers to deploy IPv6.

>> PARTICIPANT: That's why I ask.  It can to be support ‑‑ to let the IPv6 be more efficient or more quickly, because some university can set it as a course log for the courses.  It can be very good so that the new engineer coming or the new technical guys coming, they can be ‑‑ if I can put it like that by the end.

Thank you very much.

>> MODERATOR: Okay.  We started the panel a little bit late and maybe we can finish a little bit late.  Please.

>> PARTICIPANT: My name is Bevill Wooding, I'm the director of the Caribbean group and my question follows very nicely on that comment.  We have been doing IPv6 training for the better part of five years in collaborate with the IRAs, LACNIC and ARA.  We haven't seen the type of uptick in IPv6 deployment or engineering skills we were hoping nor when we are talking ‑‑ when we are talking the technical community.

And so we have begun discussing getting into the community colleges and into the universities across the region.  My question is, is there already available courses that are prepared, that we can use as a starting point for outreach to the universities and the colleges?  Is that something that has worked or that is being explored in another jurisdiction that you can share?

>> Yes, you can use the IPv6 basic online course that we have available.  It's free.  There's no charge.  And for our members, we have an advanced course that is also free of charge.

We are going to start in 2018 to keep it alive because we used to have additions, six ‑‑ four to six additions a year.  So in 2018, we really could go in and start any time they want.  So it's a self‑teach course.  So you don't knee toad know when to follow up.  But you can use that tool.  And what else.

So we have another training course based on training applications.  Those are free of charge to our members.  So just let us know if some of them could be useful for the community.

>> MODERATOR: We have a follow‑up comment from Mr. Izumi.  He says I agree with the conversation by Kinoshita, I think the key to ‑‑ was they don't stop in developing milestones.  Follow up in progress gave soft pressure to players to make sure that they get their part in IPv6 deployment, done and raise issues with all stakeholders in case they face challenges which needs to be addressed by multiple stakeholders.

>> MODERATOR: Thank you.  I think it's time we wrap up this panel.  Maybe some quick closing words 10 seconds each, please.  We can start from this side.

>> PANELIST:  Well, I think IPv6 is not a technological issue anymore but a strategic one for the countries.  So we will keep this new approach with the least developed countries in our region and see if that could trigger some deployment in our region.

>> MODERATOR: Thank you.

>> PANELIST:  I think key for me is to look further into the data and try to understand our patterns in countries that have successfully done something.  So that things could be used, maybe input to others.  So I'm going to keep working on this study.

>> MODERATOR: Thank you.  Alain.

>> ALAIN DURAN: I'm sorry, I don't have any answers here.  I just recognize that we still got a lot of work to do.

>> MODERATOR: Thank you.

>> TSUYOSHI KINOSHITA: I would say that I hope that there is April opportunity to update the BPF by incorporating some of the best practices, getting developed by the global south countries, because those are pretty much relevant for other countries.

>> MODERATOR: Thank you to our panelists and thank you to you.  Good audience for all the questions and let's ‑‑ let's wrap it up.  Thank you very much.