IGF 2016 - Day 2 - Room 1 - BPF on IPv6


The following are the outputs of the real-time captioning taken during the Eleventh Annual Meeting of the Internet Governance Forum (IGF) in Jalisco, Mexico, from 5 to 9 December 2016. 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. 


>> IZUMI OKUTANI:  So, hi, everyone.  Let's give it a minute or two until we might have people coming in and then we'll start soon.

Just trying to test so good afternoon, everyone.  Welcome to the best practices forum on IPv6.  I'm trying to test the sound to make sure that the remote participants can hear.

Does the sound work?  So this is Izumi Okutani.

Okay.  So we are really going to start.  Welcome, everybody, and thank you for attending this best practices forum on IPv6.  And my name is Izumi Okutani.  I'm the MAG coordinator of this group, together with my colleague, Sumon who is sitting here with me.

So today, at this session, I would like to have discussions in these three parts here.  So first, we want to give you a quick recap on what this ‑‑ what the document is and the work that we have done, the key messages.  So this part is going to be really brief update.  And then the second part, we would like to have discussions with you on what are the sector observations and the regional observations with you, as well as some of the speakers who are sitting here with us, who will be issuing their observations.

And lastly, we would like to do a wrap‑up on what are the key messages that we can give to policymakers and business decision makers.  So that's how we would like to have discussions of the session, and so let's start.  So first the introduction.

Why are we doing this?  Why are we trying to promote IPv6 and why are we doing this in the IGF?  I reckon many of you are familiar with what IPv6 is, but then I want to do a little quick recap.  So IP address is a technical identifier that you use to connect the devices on the Internet, just like telephone numbers for phone.  So you actually need these IP address to be able to identify yourself and who you are communicating to.

And current version, most of the Internet is using today is IPv4, it's already run out.  It's no longer available for coming, growing, and expanding users to come.  We have this new version called IPv6 where we don't have to worry about the finance resources, for IP version 4.  However, the issue to be able for users, the new users that are actually receiving the IPv6 address, to be able to connect and do things on the Internet, the devices, the access, they need to be adopted, to be technically adopted to version 6.

And this is what the technical community has been doing and working on, but to maybe there are global, wide spread deployment of IPv6 to happen, we need a wider network, such as business decisions to commercially deploy IPv6 in their network, inbound corporate, or vendor and ISPs need to have their products ready in IPv6 and there may be some countries where we need encouragement from the government and creating an environment that accommodates IPv6.

And this is exactly the work that we did last year for the best practices forum on IPv6.  We actually compiled a document, and what it measures, the government has taken or the community has taken to encouraging v6 deployment in their country.

And this year, we are actually building on that work, and we focus on the economic incentive, because after all, it's important that the people feel motivated and economically and feel the need to deploy IPv6.  So this is the work that we focused this year and we compiled the practices and the observations around the economic incentives of these six deployments, based, we have developed it into a document.  It's available for public comment until the day after tomorrow.  Until the end of this meeting.  And we have also connected case studies from different parts of the world which were all compiled in that document as well.

So I'm sharing a couple of key points we are introducing a document.  The general status is that v6 deployment is happening close to general Internet users without them being aware.  So a lot of the global content where people frequently look in different parts of the world, such as Google, YouTube, Facebook, LinkedIn, these are all v6 ready.  And if you also look at the back hand that's supporting the distribution of content, in terms of, like, cloud infrastructure, content delivery networks, it's also v6 ready.

If you actually look at your computers, the OSs of your computers such as Windows, the recent versions they also support v6.  So without, you know, again, users taking specific action, if an Internet service provider switches on the v6 access, you actually get substantial traffic in v6.  So let me show you an example of Butang, I heard from Sumon who was this in the operational community, actually.  So if an ISP switches on the v6 in Bhutan, 86% of traffic is actually in v6.  That's how the volume of traffic is there, based on v6.  This is based on the measurement from Google.  I think Google measurement, where it's showing the percentage of browsers available in v6.  Currently it's 15%, but it's speculated if the current rate of growth continues, it's going to be something around 20% in 2017, and by 2019, it's speculated to be around 30%.  So it keeps on growing.

And so if you actually look at other trends, mobile, it's being deployed rapidly, for example, like several mobile operators in the US or also in India, 70% of the traffic is in v6.  The IP architect board, where they see the engineering task force that develops technical standards of the Internet protocols.  They have made a statement that the new protocols, or the changes of the protocols that will be coming, they will not support ‑‑ they don't require the cooperation.

So if the vendors start compiling accordingly, it's not a must to support v4.  So it's like it could be, that's in the future, v4 will be a legacy protocol.  These are the things that are happening, and interesting users such as the connection, the v6 user, TV or like measuring electricity or it's used as a platform tore large‑scale streaming and companies have deployed v6.  This is a map of great of v6 showing in different parts of the world.

And this is provided by AP McLabs measurements and you can see it on the web.  And the red colors shows where there's the deployments.  And then the colors not so red, like further away from red, the rate of deployment ask high.  So you can see, like, western Europe, relatively doing well and also North America as well.  But because there are other parts of the world with almost zero deployment, if you look at the global average, it comes down to less than 8%.  So we really need to do better globally.  Are.

But what is encouraging is if you look at the top 20 countries with high v6 deployment, it's actually not linked to GDP.  It's not just the developed countries that has the high v6 deployment.  Look at Greece, Peru, Ecuador and Trinidad Tobago.  So these are actually the countries that come into top 20v6 deployment, and they are actually less than 1,200,000,000 dollars in terms of expected GDP.  There a's chance for developing countries to do well and be advanced in v6 and prepare their infrastructure.

Okay.  And I think I skipped a couple of slides too much.

Okay so the last slide is that out of the successful cases, we have tried to connect case studies from different parts of the world, over 20 cases collected, and the common reason for the v6 deployment, first, which is probably not too surprising is for the long‑term continuity of our business.  Another reason is they want to give it like a brand ownership and be ready for the coming practical technology.  And there's also a cost simulation, where I don't know how applicable or widely it is, but then it's a case where it was observed.  It's less expensive to do v6 than v4.

The point a couple of years ago, IPv6 was more of an issue, but now it's more of a reality.  It's happening.  And we have to get ready.

So that's general status of when we are in v6, and like now, I would like to move to Sumon where he will be sharing the challenges.

>> SUMON SABIR:  Thanks.  It's mostly the reason behind the deployment, and along with that, also we collected some changes while deploying v6, and ‑‑ and mostly we have found from the feedback of the respondent that the challenge was on conference site.  It's a large portion of content, and most of the local content is not available.  The observation is if we use v6, it controls 40%.  So this is a complimentary issue, and whereas the lack of support in some critical product sets and it's limiting.  And also there's some support for mitigation in terms of v6.  So also, the training and the human resources and preparing them to run and deploy on v6.

And then there's a complaint from the ISP that the infrastructure was ready, but still clients have been are not happy that IPv6 is enabled but they don't see much rise in that traffic of the network.

So ‑‑ so the policies that they ask our permission to apply v6 because they don't want to take a risk that the customer will complain about service or logistics and so it's the separator.  And there's some cases, at some additional cost incurring and we can do it at scale to v6.

We deployed v6, we try to get some information from the countries where v6 is not deployed.  What are the challenges they are facing in deploying v6.  So we tried to get in feedback from different stakeholders.  Most in south Asia, Nepal, and the ISPs have the challenges, and there are two group of customers, and the broad band users.  And they don't have any challenge in providing a physics service to their kind.  In some cases.  The client is reluctant to take v6, because people are not really ‑‑ they are comfortable in the deployment of v6.

Now getting broadband users somebody has complained that there were some issues in ‑‑ see specially the ‑‑ so there is the challenge of shipping the traffic to packets.  So it's easier for deploy.  And also there's a lack of IPv6 knowledge and also on security and we are preventing them from deploying v6.

And also the common point here, like the other developed countries that the CPEs were born.  So even those who deployed v6, there's some reason not really able to use it in the interim network.

And so they are still waiting and they say that a number was still needed in the developing countries.  It's crossing 20%.  It's growing rapidly and now they are considering it's deployed to v6 quickly but still Thailand, where they have v6.  So it's a challenge that we are seeing and really content provider, they said we will probably deploy v6, but the solution that they are not (Audio distorted).

But other than that, but only one for content provider and local content provider deploys v6.  But if you see in the ‑‑ in the work studies, those who are actually ‑‑ they already have logistics but those will be more content.

Some of the challenges, actually, we founded in developed and developing countries.

>> IZUMI OKUTANI: Thank you, Sumon, and so we go, and Susan and Marco will be moderating part two.  Before we go, in case anybody has any questions?

>> AUDIENCE MEMBER: So I have remote intervention.  I also work in the field of education for an international association, and as an end user, we recently started to deploy IPv6 within our network.  We put pressure on our ISP by arguing we will go away if they don't route v6 from and to international traffic, I think he means.  They achieved that because we are a good customer.

In Cameroon, if the universities were to raise the needs on IPv6, it can move ‑‑ it can help IPv6 be, you know, increase the usage there, and the most important ISP in Cameroon got IPv6 from AFRINIC and they didn't activate it because they didn't see the need.

>> IZUMI OKUTANI: Excellent.  Thank you so much for sharing this case.  I have think pressure comes from customers to deploy.  Thank you so much for sharing this.  We appreciate it.


So if no other comments or questions, let's go to part two, where Susan and Marco will be moderating.

>> MARCO HOGEWONING: Thank you, good afternoon, my name is Marco, I work for the regional Internet registries and we are deploying IPv6 and we are part of the cause of the problems because we are the ones running out of IPv4 addresses.

Also me with Susan.

>> SUSAN CHALMERS: My name is Susan Chalmers.  I work in the office of international affairs at the national telecommunications and information administration for the US Department of Commerce.  I'm very pleased to be here.  I think this is an excellent Best Practice Forum, and it's building on some work from last year.  So just, again, thanks for allowing me to moderate.

>> MARCO HOGEWONING: Thank you.  So we've got quite a number of speakers lined up, who each contribute to the best practices forum.  What I want to do, to make a quick round of the speakers to allow them three or four minutes and explain their stake in IPv6 and what have you done to sort of deploy IPv6 or carry IPv6.  And I suggest we start with Aaron, if that's okay with you.

>> AARON HUGHES: Hi, good afternoon.  My name is Aaron Hughes on the I'm on the board of American registry Internet numbers, and 6connect.  I have been in this industry in the technical industry from the perspective of primarily operations for nearly 25 years.  I have done all kinds of v6 implementations, whether it's working as an integrator in automating provisioning solutions, to simply helping people with decision making around what implementation methodology, as a global provider.

I have been working with a number of different flavors of vendors on many years on getting standards to have future parity between both v4 and v6, as well as done many direct consultations with organizations on approach.

I have done it myself for all kinds of different ISPs.  I have been developing in lab with the Internet of Things space, as well as was on a project called TMOS TSAT and had implementation of a router on a satellite network.

>> MARCO HOGEWONING: Yes, I can probably just go down the line.  Carlos? Carlos, I work for LACNIC.  As Marco said, we are somehow part of this problem.  Let me tell you a little bit about my problem.  I'm currently LACNIC's CEO, although I have been working the communications and Internet for the last part of 20 years, I would say.  I have a funny story to share that is ‑‑ first day after I got my degree in engineering, I went back to work, and they decided to give me something to read.  Something different to read and they gave me something about this new thing on the Internet.  This new sort of IP protocol that was the new thing and it was back in 1997.

So since then, I have been involved in different aspects.  In fact, with some colleagues that are present here in this room, we configured the first IPv6 network for a country in Uruguay.

We did some very early trials of v6 networks and then when I came to work with LACNIC, I have been doing extensive training and helping to run IPv6 networks.  We make a point of doing IPv6 everywhere we go.  Every conference that was run and even that we host has to have a v6.  And you have perhaps seen me complaining about the lack of IPv6 and the network, which I'm glad to say it was solved this morning.  So you guys have IPv6 on your laptops even if you don't realize.

And that's part of the thing.  You are not supposed to have ‑‑ to do anything, to have IPv6 on your computer, right?  So if you didn't notice any change from yesterday to today, that's fine.  That's the way it is supposed to be.  So that's my introduction.

>> LISE FUHR: Thank you, my name is lice fur and I'm from ETNO, that's the European network operators association.  It's a long and boring name but we represent the European telco operators and, of course, many of my members are deploying IPv6.

But as I actually represent all of my members, I'm talking for the members and I must say that we as an industry actually support the IPv6 deployment.  We have think that's the early and long lasting solution for this problem and we also, of course, committed to do it in a timely manner.

But, of course, as we heard, it's actually ‑‑ it's a complex issue.  It's costly, and we actually are trying to respond to that, to the growing needs, but it's not done in just a day.

So we acknowledge our responsibilities as being the traffic interconnection, and we have the leading rolls as the large telcos and ETNO members, we work with the Working Groups and we share a lot of knowledge and experience on this issue.  And we are working very hard to pass on the message to our members on ‑‑ on IPv6 is important.

Before doing this ‑‑ before saying yes to this panel, we actually looked into what is the experience of the different operators and why is IPv6 not there just like that?  And what we found out, of course ‑‑ and that's a very easy way to excuse ourself, but it's not an excuse but no size fits all.  Every company is different.  And if I look at some of the feedback we got from our members, we got that there is a technical need from some of our members.  They actually see that IPv6 is a better deployment for services, for some of the services and this is actually very important because as businesses you need to actually convince the business part that you need to do this.

Others actually found ‑‑ and this was more ‑‑ what do you call it?  Well, they found that they actually wanted to try it out before it became a necessity.  So there was someone who was willing to say, okay.  We know this is coming, but we will try it out before it's absolutely acute.

Also others are actually looking into embedding IPv6 in whatever projects they do, in order to make them future oriented.  I'm not saying futureproof, I don't think there's anything that is futureproof.  So we have companies that are now actually looking into embedding IPv6 in whatever they do.  But we still see it's actually driven by technical reasons and not business reasons and for telcos who are very ‑‑ under a lot of pressure, especially in Europe, where the competition is fierce, you actually need to also have a business argument, but it's actually starting to be much more apparent for our companies, but there is a business to this too.

We also see that the number were actually deploying IPv6 are doing it in different ways.  Some are actually making specific, setups do it and others have it for the technical department.

Actually what we see is the still slow uptake of IPv6.  We are willing to look at how we can accelerate this and we think there is a need for this.

>> PAUL WILSON: Good afternoon, everyone, I'm Paul Wilson from APNIC, we are another one of the regional Internet address registries.  We have serve the Asia Pacific region.  We have around 13,000 network operators, including ISP and enterprises across our membership.  So responsible for allocating, providing them with IPv6 addresses and of course we track those activities.  We publish information about ‑‑ about IPv6 uptake in terms of IP, in terms after dress allocation.

And we also as a membership organization have a whole range of services to our members including these days trading in technical assistance and conferences and fellowships and so forth.  And we periodically survey our members formally.  We have various withs of understanding our members' needs but one of them is a periodic survey of every two years in which we are asking about members expectations and challenges and needs and what they would like from APNIC in terms of helping with their requirements.

So we have found over many years that the demand for IPv6 support is very strong.  That, of course, although, I guess everyone knows the actual process of deployment is being a long process, and a slow process.  The interest has been maintained over many, many years and I think one of the points of this session will be to show why the level of interest isn't matched yet in the level of deployment, but suffice it to say, like the other ARIs, we are providing strong levels of support for IPv6 and our latest survey showed that that's still an ongoing demand.  Plans are coming to fruition.  We are starting to see as you have also heard deployment of IPv6 around the region.  So we are tracking that.  We are trying to understand where it's happening and what are the best practices that are allowing it to happen and happen in the best way possible.

So that's the introduction to APNIC as an ARRI.  We have a division called APNIC Labs as well and we provide one of the global sources of reference points for the actual utilization of IPv6 by end user customers.  So you can find your way to IPv6 measurements and see the measurements as Izumi showed earlier of the deployment rates around the world, and I'm quite able to talk about the various scenarios and cases in our region in the Asia Pacific, in particular, where IPv6, as I say is starting to be deployed quite actively.  These days.

So I think that's all for me for now.

>> MARCO HOGEWONING: Thank you, Paul.  We have one more panelist who is not able to join us in the physical presence but is open remote.  Afifa Hariz is online.  We can patch her in.  Yep, I see thumbs up.

Okay.  Afifa can you say something?

>> AFIFA HARIZ: Hello, everyone.

>> MARCO HOGEWONING: What's your stake in IPv6?

>> AFIFA HARIZ: What?  Okay.  Can you hear me?


>> AFIFA HARIZ: I'm in India.  I'm with the higher education university.  I'm a member of the IPv6 at our university.  So (Audio distorted).

And the technology ‑‑ the technology.  So we are ‑‑ we have status to deploy IPv6 in our university.  It was deployed with collaboration.  The importance of this technology with collaboration, we ‑‑ our university ISP had to pick up, and with the Internet and the history ‑‑ the IPv6 preference.

So despite the actions (Audio distorted).

And it's the first university in India that permits approval of the IPv6.  So our expansion is that we have started from the top down.  We have ‑‑ (Audio distorted) so every efforts are deployed.  To push up the other departments.  To take serious steps to deploy IPv6 and the strategy for the for IPv6.

>> SUSAN CHALMERS: Thank you.  Hi, Afifa, this is Susan, one of the moderates.  I think with that said, now we have some mic feedback.  I don't know if someone can help with that?

Okay.  All right!  Before we dig into the second section of this session, I was wondering if we could just get a show of hands for those of you who are in the room, if you have worked with IPv6, either in deploying, implementing or providing services to somebody else or working around IPv6 policy, would you please raise your hand?

Okay.  Great.  Thanks.  Much appreciated.  I think during this section, along with the panelists, please consider the questions and the comments that you would hike to share as we move throughout this section so we can all just pull up our knowledge and have it there for the record.

And so I kind of want ‑‑ I want to begin with question, Aaron for you and it's kind of an ice breaker question.  Recently NTIA, we published requests for comment on the challenges in implementing IPv6, but what are the incentives and the motivations and in some of our submissions received, one in particular was from state CI Os association.  And these are all public.  They are on the website, which I can share later.

But they said, look, we just do not have any sort of incentive because we are completely fine with our pool of IPv4, there's absolutely no energy within the ‑‑ within our department to implement IPv6 at the moment because we don't have that demand and so I was just wondering what you would say or how would you explain the importance of IPv6 and how it might impact other parties within the Internet ecosystem to that submitter.

>> AARON HUGHES: That's for that question.  It's quite a mouthful and quite a challenge to answer.

I guess we could start with the simple fact that corrected companies, organizations, private and public sector around the world are implementing v6 with different approaches.  Sometimes they use the method of dual stacking where they implement both v4 and v6 at the same time.  Other times, like in cases of, say, large mobile providers, they take the leap to jump to v6 only and use translation devices to connect to v4 infrastructure.

And that simple fact alone means that what needs to be your native connectivity, that is end‑to‑end reachability of v4 to v4 goes away without you being involved at all.

So your user experience in accessing websites and services from a company that would move to a v6 only infrastructure and use v4 translators impacts your user experience.  So doing nothing actually hurts you.

Certainly as a leader in the public sector, you know, setting the example of implementing v6 is a very good thing in many aspects.  Certainly in economic aspect and I'm sure we are going to get into this a bit later, but, you know, when it comes down to it, at the end of the day, doing nothing, we're waiting, leaving your infrastructure on v4 only continues to cost the world more and more money.

That cost goes up, whether ‑‑ if you do nothing or if you take more time before you do IPv6.  We certainly have got some great examples of cost analysis from various perspectives of different types of organizations which I'm sure we are also going to go over.

But I would encourage someone who thinks that they are fine by doing nothing because they have plenty of resources to come to understand that v4 technology is legacy.  The current protocol is v6.  It's not about the number of resources you have.  It's about the future of connectivity, end‑to‑end reachability and future growth in ‑‑ in the protocol itself.

>> MARCO HOGEWONING: Thank you, Aaron.  I will quickly go back to ‑‑ like the one size fits all.  That's always a custom job but there again, here we are and trying to come up with sort of best practices.  How are we to sort of generalize the deployment experiences or how easy is it to sort of link between one successful adoption and how that can help other profilers to deploy that.  How do we evolve in trying to bridge that knowledge gap?

>> LISE FUHR: Actually, what we are trying to bridge is have one of our Working Groups involved with IPv6 and it's important to exchange the knowledge or the experience that our members have with IPv6.

I know IPv6 is a standard that's fixed, how to deploy it is actually different from company to company.  And that's because there are ‑‑ they have different services, they have different sizes.  It's more a matter of what are you using of technology too.  So for us, it's not ‑‑ I cannot say all of our members can do it in the same bay and at the same pace, but what we see is actually that we see a growing interest.  We see a growing understanding, and we also see, as I said before, a use for it in the business perspective, not only technical perspective.

And I think that is actually key.  And if you get the demand side, you will get it faster deployed because if there is a demand from the customers, which we know and we hear here, it's difficult to explain to people.  Why do we need IPv6?  Because that's going to need a lot of change of equipment, like the end users which is also tedious and expensive thing to do.

But actually, we could encourage governments to ask for IPv6.  We could ask them to, in their procurement processes say, well, we want services that are IPv6 ready.  So there are ‑‑ because many of the smart cities, many of the Internet of Things will actually have this demand of IPv6.  So I think by pushing it and having the governments helping us, I think that would be a way to accelerate the deployment.

>> MARCO HOGEWONING: How will the governments help you?  That's a clear call of action.  Left's stick a little bit to the government role.  You have quite a bit of experience and also talking to governments.  What's ‑‑ what's the role of governments and regulatory environments and especially in the APNIC region to sort of motivate people to deploy IPv6?

>> PAUL WILSON: I think the role is important.  I think the role in the Asia Pacific is particularly important.  Unlike in much of the world, people actually do listen to their governments in much of the Asia Pacific.  And so statements from government and urging from governments and incentives from governments can be very meaningful, but I have to say there's very little experience with success in that area.

I think a government of any ‑‑ any type is kind of reluctant to legislate, for instance, to force an industry to do something, when they really do not understand whether that think is possible or feasible in the town, whether it will cost a huge amount.  And they will be interested in what their industry has to say and what oversee experts might have to say.  So I think ‑‑ I think the role of government needs to be sort of well understood.  I think procurement as Lise mentioned is very important, but I would actually say that that same principle goes for everyone.

People often ask me, what will it cost to deploy IPv6 and it's sort of a silly question to be asked, because the deployment of IPv6 is completely situational.  One size fits nobody and the cost of IPv6 depends entirely on what is your existing usage infrastructure and dependency on the Internet.

But for any person who has a dependency on the Internet, a person, a government, or the business, it's now ‑‑ it really now is incumbent on you to understand the future of your reliance on the Internet and in particular, IPv6.  So the Internet ecosystem is actually one of the services and service providers and service users of all kinds, connectivity, equipment, software, even consulting services and staff services.  If you are receiving any Internet service from anyone, you should now be asking and placing importance in proportion with your role on the Internet.  You need to now be asking those service providers, what are they doing about IPv6?  What is their recommendation for IPv6?  When are they going to support IPv6?  And you kind of do need to have some understanding of the answer.  So it's quite true that some ‑‑ some folks actually don't need to move now.  They don't need to move just yet.  Some would, on the other hand, gain a great benefit from moving now in terms of particularly purchasing and deploying new infrastructure, building new networks.  This can all be done today, using IPv6 and for anyone to ‑‑ to actually undertake an exercise like that today, without actually putting IPv6 into the picture, and deploying it now, it's a huge wasted opportunity and potentially a waste of money.

So I think the interesting thing about the Asia Pacific region, which I guess most people understand is the rate of growth is pretty phenomenal.  It accounts for the majority of Internet growth around the world, and that's going to go on for many years.  The interesting thing about that, as I said, where networks and infrastructures are growing, that's where you are deploying new services, new equipment, you are purchasing, you are deploying new infrastructure and that, as I say can be done with IPv6 now and we are actually seeing not on a country by country basis.  As I say, there's not much that happens at a country level but we are seeing on an ISP versus ISP basis, that they can do that with IPv6.  And so we are seeing quite large jumps in deployment being driven by individual ISPs who are building new services and deploying new services and it's happening in the Asia Pacific, probably more than elsewhere.

But it does explain something we heard earlier which is that there's some interesting and unexpected cases all around the world, where developing countries have actually leapt into the leaderboards of the IPv6 deployment in a way that many people don't expect.  But it's for exactly that reason.  Thanks.

>> SUSAN CHALMERS: Thanks, Paul.  My next question is for Carlos.  So Carlos, there is a bit of discussion earlier in this section how if IPv6 is implemented and then the customer wouldn't really notice or that it's ‑‑ that we are looking at the providers and certainly other parties to take a lead but not the customer.  Given your anecdote, as we opened, thank you by, the way for lodging that ‑‑ you know, that concern.  What do you see the role of the customer being in this in IPv6 transition?

>> CARLOS MARTINEZ:  That's an interesting question.  There's sort of a contradiction here.  As I said, the customer is not supposed to notice anything, yet companies and it's reasonable for them to do so, expect somehow to have a demand from the customer in order to deploy IPv6, and have this case, as Lise said.  And that makes sense.

That's what makes the case for IPv6 so difficult in the past.  IPv6 actually is a proposition that does not necessarily bring new things to the Internet, but what assures you that you are not going to lose what you have now.

So in a situation that parallels somehow the investment, you are investing money, not to have something new but not to lose something, it makes it very difficult to somehow ‑‑ somehow to build this business case.

So taking this to the case of the customer, what are the possible things that a customer may lose by not having IPv6?  Well, as was said, actually, losing the end‑to‑end connectivity the application that we are used to, normally used, like, for a particular case, there become more and more degraded and difficult to use.  Gaming is another.  Video streaming and downloading large files are different applications that do work now, but may have some issues in a post‑IP world.  For example, in a world where your IPv4 connectivity goes through an ICGN box.  There's an interesting study by people from NTT in Japan that, they purposefully degraded an IPv4 connection and restricting the number of floors and see how the different applications degrade.

I highly recommend taking a look at it because it clearly shows how by not moving to IPv6 the user experience will be degraded, and actually, some of the cases we have in our region.  One of the regulars for IPv6 has been the growing body of complaints from users.  I mean, the user does not require IPv6 by itself, but it complains about the things that they are losing.

So this is sort of ‑‑ I would say inverted by proposition, if you will.  So that's what has made it very difficult, essentially to build an economic case for many years.

Now, I believe the economic case has been, you know, emerging and the picture of this economic case is more clear now.

>> MARCO HOGEWONING: Thank you.  I want to take this to the audience.  Get your questions ready.  So what does the team quickly say.

>> PAUL WILSON: If you visit the APNIC website, you will see probably an IPv6 address given to you, which is a long string of letters and numbers shown near the top of your screen.  If you don't see that, then you are probably using an older computer or you could do it with some help with your configuration.  I'm sure one of us up here could help with that if you would like to experience the IPv6 sometime here in this week.  Thanks.

>> MARCO HOGEWONING: There is a row boat in the village that have the experts and Paul has decided is the IPv6 help desk.

Questions in the room?  People who have deployed IPv6.  And anybody who wanted to add something to what was said.

Please introduce yourself.

>> AUDIENCE MEMBER: Hi.  Many universities also not only companies to deploy IPv6.  And not only to work on the connectivity issue.  You have to work on the applications.  Many are waiting for IPv6‑only application and they don't realize what they have to do now is to have maybe the same application or conversion of the application, one which will supports IPv6 and one with IPv4.  They have been doing it for many years.

So I have been working and trying to teach people that they have to learn how to deploy applications or continue using the applications but now we have the protocol.  Yes, they are going to spend more money, continue working with old versions and in some cases, like in the mobile scam.  That we have seen, and you have mentioned, he's deployed only with IPv6.  I'm not considering IPv4.  So you can work on many aspects but also the applications.

>> MARCO HOGEWONING: Thank you.  And somewhere in the process of writing up the BPS, you have mentioned something about applications and then specifically ‑‑ and now ‑‑

>> I assume you are referring to the Apple Store requirements.  I'm not sure that there's a killer application, necessarily.  There may be killer advantages to using v6 for some things down the road.  One of those is vehicle to vehicle or Internet of Things area, or Internet of stuff, connected things.  Where we can actually get away from the client server model and actually move to more models that go client to client, communication or client to open controller communication.  All kinds of benefits that can come from having unique numbers, uniqueness on the Internet and not using network address translation and v4.

But also to your point, in terms of application, sometime ago, at an Apple developer forum, Apple announced they will require all application developers to support v6 and test the v6 environment before they would be permitted to update the application on the app store.  And this was ‑‑ this was done with the knowledge that in particular T‑Mobile in the US had intended to eventually cut off all v4 underlying connectivity on Apple iPhones.  That's actually happening very soon.  So the IOS 10.2, which comes out in a couple of weeks.  If you are one of those 65 million users and you will be a native IPv6 only and that was the reasoning behind the Apple's slow drive to encourage developers to test it on the v6 only environment.

As of that release, you will see a large jump in direct native traffic.  It's not across the global Internet, in handsets and the app store and those applications running on those end devices connecting back to their infrastructure.  So while I'm not sure it's a killer app, I think things are happening in that space that are going to be pushing for that better connectivity, for the native end‑to‑end connectivity again.

>> SUSAN CHALMERS: Thanks, Aaron.  We want to turn back to the audience, oh, it looks like we have a few hands over here.

>> AUDIENCE MEMBER: This is Mike Levy from CloudFlare.  There are a bunch of different things that I could add to this and those people who know me know that, but one of the things that was brought up was this issue of using v6 without even knowing it, and I'm going to, if I may, expand upon that, because the aspect of knowing it, if one was to project to v4 is did the end users, that's us in the room and 3 billion other people know that they were using v4?

No, they knew they were using the Internet within a very high ‑‑ within a group of technical people, of course there, was an understanding of what they were using.  So it is actually incumbent on us to understand that what we are doing when we are building infrastructure, or delivering web pages or providing mobile access, that we are providing access to the Internet, and it so happens that is the v4 and v6 at the present moment in time and it is a combination of maybe only v4 for a set of players that have yet to implement v6 or in the mobile world ‑‑ and T‑Mobile is not the only provider in this case ‑‑ where the tools are out there to deliver v6 only over the air connectivity and v4 becomes a ‑‑ a poor stepchild.  The way of implementing connectivity for v4 is a slightly degraded tumbling technique.  Funny how that got swapped in the world, where v4 connectivity not only costs them more in deployment, but also cost the end user in the quality or the connectivity.

So that's the ‑‑ that's the preachy bit.  Now let me deal with some numbers.  Our company is a commercial company.  We provide content delivery network services and we recently went through a process of making sure that as many as possible domains on our system are ran with both v4 and v6, and we have been doing this very voluntarily for many, many years.  We had decided earlier ‑‑ well, summer of this year, that we would make a major change and we went back and did a migration of maybe 4 million different URLs, different zones, so that the v4 and v6 ‑‑ because we do this.  We provide the connectivity on behalf of the customer.

So instead of asking the customer, do you want v6?  We really went back to the basics.  What are we providing?  We are providing connectivity and acceleration and DDOS mitigation for websites on the Internet.  And for a very, very large percentage of our customers and there are some subtleties that we discussed in a blog, this turns out to be 98 point something percent of our customer base.  I want to pull out some interesting stats that helps with Izumi's argument earlier.

If you look at very specific operators out there, such as certain mobile operators, such as certain connectivity broadband providers and you look only at those specific operators, you start seeing very high percentage of v6 penetration.

If you look at your overall stats including that long tail of many, many tens of thousands of networks spread around the globe, then you start seeing a much lower number, in sort of ‑‑ well, depending on how you measure it, 8%, 12%, sometimes even lower than that.  So I'm just going to just share some stats which is a published ‑‑ we published about two weeks ago.

In the US, we see AT&T broadband or Comcast cable sitting at about mid35% penetration with v6, but that ‑‑ sorry that really is not penetration.  That's actual traffic.  That means collecting connectivity from a web browser or for a noble device is actually in 35% of times when it hits us, actually becomes v6.  Let's pick some good numbers out of this.  T‑Mobile, which is also matched by Verizon sits at plus or minus of 1%, at 50%.  Half of their requests are coming out at v6.  There must be a good reason for that and to go back from the gentlemen from Mexico, from Yuma, about the ‑‑ but those operators and the deployment of code and the way that they operate have deemed that if they can use v6, they should use v6.  They know something about the cost basis, because ultimately nor that many millions of end users they want v6.

Going higher up the list, we basically look at ‑‑ think people like telnet.  Something really good is happening in Belgium, outside of beer and chocolate.  We see 80 something odd percent or 90% odd percent of traffic on that particular provider which is where the stats for Belgium go up.  They are pulled up fairly high because of that.  So as I said, looking it per provider basis, we see very specific choices which are biasing towards v6 for good reasons.  Orange Poland is up there as well.  A few other players.  Go look at the blog.  You will see that there's both hope from people like myself who spent a long time trying to get people ‑‑ to convince people to implement v6, but also we know from talking to these very specific operators that they have come up with good economic models because ultimately, they are scale.  They have to do v6 for a good reason.  I can give you wonderful academic reason to do v6, but in the commercial world, that's where things happen and it has to be a good reason for it, and this seems to be that.

I will stop now if anybody has a problem finding the blog, find me afterwards.

>> MARCO HOGEWONING: Thank you, Martin.  I have read the blog and it's certainly worth looking up.  There's a lost interesting things there.  And Antonio, you also wanted to make some comments here.  Thank you.

>> AUDIENCE MEMBER: Thank you.  And I'm Antonio from Meek Brazil.  We would like to first, talk about one factor that we are exploring.

Some of our content providers stated to implement IPv6 because of the risk of not attaining very well their customers.  So it ‑‑ it was kind of history.  We told them ‑‑ we told them that if they do not implement IPv6, they could not lose customers but their customers, because they were behind meant ores, they could have a worst experience.  I think this convinced one of the content providers to deploy IPv6.

So even the case where the customers, the hosting providers started to ask for the hosting providers to implement IPv6, what ‑‑ it seems that it's very different of ‑‑ of this plan.

The other thing I would like to talk is about problems with the IPv6.  I would like you on the panel if possible to talk about it a bit.

And it addresses in the ISPs and the telecom providers in general, they are doing this.  I think it's true for Brazil and almost all the world ISPs but they are out on this.

And one the main problems we have today is with some vendors of people like video games and Wi‑Fi routers and things like that.  And, for example, there's only one of our operators ask me if we had some contact with the people from Sony, because they were use v4, and there are an increasing number of customers complaints.  If one customer is black listed then a lot of other customers are black listed and they look for information with Sony to deploy IPv6 in their network and they are giving IPv6 to their customers and so top down, the ‑‑ the ‑‑ they plan to implement IPv6 by 2020 or something like that.

How to address the problems with Wi‑Fi router, video games, TVs and things like that, that are ‑‑ I don't know.  A big problem today.

Okay, it's actually a good question.

>> MARCO HOGEWONING: And I'm going to give the panelists a bit of time to think about it, as we heading to wrap up.  1.5 hours is shorter than you think.

Before, that I want to quickly go back to the decision, but I asked, how can governments help you.  That was kind of exactly the question that you have do this.  Please supply us with your information.  Maybe you can talk about that.

>> LISE FUHR: Well, thank you.  We are asking for input from the public so they can help us understand better.  Bust I think that one way that governments can help is as Lise mentioned earlier, is to lead by example.  So within the US, we do have a federal IPv6 task force that is active and actually harkening back to last year's Best Practice Forum, we heard from a representative from the German government, who described their work on profiles.  So IPv6 profiles within procurement, which outline those standards and requirements and specifications that help guide vendors through the procurement process within the US, it's the national institution for standards and technology that is leading this work.  So if you are interested to understand where we are, certainly visit the NIST page.

In the RFCs, we fielded a lot of recommendations from the commenters and we had a diverse array of commenters all different types and it was great regional Internet registry, or enterprises, also ICT service providers and some of these suggestions that we got from the community were related to procurement.  They encouraged us to help raise awareness, help bring the different bodies that are existing within this community together more often, so they can come together and share information on the network.

One thing, in particular, that based upon ‑‑ so we ran this RC to ‑‑ for two reasons.  First to help inform the Best Practice Forum, and the content that is submitted, but also to help inform NTIA's efforts going forward and helping to promote IPv6 adoption, and one thing that we are considering are developing multistakeholder conversations around certain topics within IPv6.  So some of the suggestions that we received were dealing with security in the environments and the privacy implementation of developing and deploying IPv6 and mobile IPv6 routing.

So I guess I throw the question back to everybody else, and if you also have ideas for those types of conversations be very eager to receive them.

>> MARCO HOGEWONING: So let's ‑‑ let's be fair.  In total, you have five minutes to share your final thoughts and maybe I will just hand the Mic to Paul and then maybe you can address Antonio within that short time, we also have Antonio question.

>> PAUL WILSON: Wow, that's quite a challenge in a short space of time.  The request for a killer application for IPv6 was kind of abandoned a long time ago, I think, but I always thought that the killer app was the Internet itself.  That's what we are all enabling by deploying IPv6, and unfortunately, I think the early promotion of IPv6 was over done.  Some people have said, well, why are we still talking about IPv6?  I heard it was needed ten years ago and it wasn't.  The fact is that it actually is needed today.  So at the time, the times have definitely changes.  I said earlier that some ‑‑ that there's not much the governments can do, but I think it was thanks to some of the early promotion, that countries like Japan, in my region, actually, play a very strong leadership role at the government and in the industry, and ‑‑ and technical community in promoting IPv6 and actually contributing greatly to IPv6 standards to IPv6 technology and developments.  So I wanted to mention that in particular.

So at the moment, times have changed and we are seeing rapid deployments.  It seems ‑‑ it does seem strange after we have been talking about IPv6 for 10 or 15 years to say that today there is actually an earlier adopter advantage as we also said at the beginning, that an ISP that deploys IPv6 now will see 50% of their traffic nighing over IPv6 from day one and that takes a huge load off the carrier grade and we haven't really covered it here today, that otherwise cost a great deal of money and sink a lot of ISP resources.

So again, times have changed and the time for many people is now.  I don't want to oversell it and say that everyone must ‑‑ must do this today.  But I do suggest that we all need to make an informed judgment about when and how we need to address IPv6 in our own use and reliance on the Internet.

Thank you.

>> LISE FUHR: Thank you.  I think that many of the numbers we have heard today is actually underlying what you also said, Paul, is that the deployment and the demand is growing and I believe governments can play a crucial role by being the leading example, I think they should ‑‑ by doing any regulation in this area, because we see there is different ways to do different places but it's seen as a competitive advantage now, and many of ETNO's members are looking into this perspective, and it will also be the driver and so absolutely no need for regulation.

"Carlos Martinez:  Since we have enough IP for me, then we have no need for IPv6.  This is based on a misunderstanding of how Internet works.  Actually, in order to have connectivity, you need to be able to sort of speak the same language, the same protocols at the other end, so this actually makes the Internet a place where it's somebody else's problem.  Everybody's problem is everybody's problem.  So if there is a problem with IPv4 that is my problem as well if I want to be connected, right?

I said ‑‑ I also was mentioning that the panel ‑‑ IPv6 will start to become the first choice for applying new networks and IPv6 will become, you know, the thing you do on the side.  And this will sort of move the issues that people see on IPv6 to IPv4.  They think what is that turning point?  I'm pretty sure that they are very close to the turning point today, and we had a very interesting conversation to this one.  For example, in this model, a lot of small numbers in the Caribbean.  The thing is they need to realize that they are relying on a service, on a protocol that will soon become, you know, the little thing that you do on the side, and then, you know, the service quality will suffer and the connectivity will suffer.

So one final thought, again, the Internet is all about connectivity.  And it's about our interconnection.  My problem is everybody's problem.  So ‑‑ and this applies not only to IPv6, it applies to many things, security, you know, attacks, many things.  And this is something we, everybody is going to cooperate and need to have in mind.  Thank you very much.

>> AARON HUGHES: I think my closing comments can address some of Antonio comments.  As an individual, an organization, a government is to encourage people with your wallets and in every purchasing decision going forward, whether that's an IP enabled light bulb or a gaming system or a washing machine or looking for a service providor for connectivity or as a service model, even if your infrastructure doesn't support v4 or you don't have v6 at home, when you make those purchasing decisions, make sure to ask about v6, where it's possible you should buy products and services that have it enabled or have a plan, such that you are not in a position to have to make that purchase again because your device is now considered legacy.  So in a system that is a gaming system, because it would block CGN.  You already had that discussion and I would encourage you to think about that, as purchasing authorities within the organizations, and the governments for the contracts for the RFPs.  Thanks.


>> IZUMI OKUTANI: Thank you to our panelists and moderators.  I would like to do a final wrap‑up of IPv6.  I'm not sure if we are able to see the slides.  Can we switch to showing the slides?  So takeaways for policymakers and business decision makers.  So I think earlier we heard that people are against regulation, but there's actually a request from one of the contributors that we all ‑‑ that they want governments to request vendors to support v6 in certain products because there's actually no inventive for the vendors to do it on their own and for ISPs to ‑‑ or like corporate networks to deploy.  So in future, such as intrusion detection, security, and it doesn't support v6.  This is more encouragement from government is needed.

So that's one point.  And awareness raising of consumers.  So this is related to the point where Antonio was raised that CDN Reuters.  So customers need to be aware and choose the product that supports v6.  And this is an area where they are able to identify what are the products that is available and supports this.  And customers can choose and raise awareness.  And third point, reach out to business RT decision makers in the industry.  Sometimes not just pressure from customers or the technical community but if customer, not in regulation, like pressuring ‑‑ but just to give it an encouragement.  Then we'll actually keep it in the mind of business decision makers.  This is something that our government was saying.

And then lastly training of engineers, of medium and small scale business where they can't afford to have the training of the old staff.  This is where APNIC is collaborating with ITU and promoting trainings and this is where we can cooperate with the private sectors and providing trains that they can do their own staff training.  And so for the vendors, I will make sure that the product supports v6, and for service providers by starting deployment now, you actually have the large investments that is to be ‑‑ you know, that we need to ‑‑ rather than ‑‑ you know, you have to invest a large amount, and then you start thinking, oh, I better be ready for v6.  You do it gradually now, and then you will actually save a large chunk of investment.

Another thing I think it was earlier mentioned by the panelists wherever you upgrade equipment or deploying the network, make sure you choose the equipment that supports v6 so you don't have to actually build a whole new v6 network but gradually do it when the equipment, or our network.

And lastly, when you deploy v6 commercially, turn it all by default.  So I think this is relevant to the point, especially if you are an ISP, I can provide it.  And some may be scared, oh, it will deteriorate, but then we are hearing the reverse case and I think there are a couple of commercial companies that have already put on v6 by default and we are not hearing that they are facing any issues and complaints, and these are the things that business decision makers can actually keep in mind.

So the last slide, the summary.  I'm not so good at clicking.

So if you are in the Internet industry, you should deploy v6 for long‑term business sustainability.  That's the key message we got open the panel and also from our document.  And growing trend you can observe in v6, so it's not just like insurance, and let's just do our part in what we can to be ready.  So that's pretty much it.  Thanks again for panelists, the moderators and would is listening and those who contributed.  To thank you so much.  And our comment, the document is available for comment until this Friday.  So feel free to come up the line or themselves Sumon who has been helping us so much in developing the document.  So thanks, everyone.


(End of session 4:28 Central Time.)