Transcript - APNIC Plenary

Transcript - APNIC Plenary


While every effort is made to capture a live speaker's words, it is possible at times that the transcript contains some errors or mistranslations. APNIC apologies for any inconvenience but accepts no liability for any event or action resulting from the transcripts

Sunny Chendi: Good morning and I know it's ... I do apologise for starting this session a couple of minutes late.

Before we begin, I would like to run through some administrative matters, so have a Conference here in Delhi and we take the full benefit of this session from today onwards.

I'm Sunny Chendi, I manage the APNIC Conference program and I would like to acknowledge to our sponsors, for supporting this event and making it a success and for all those who are joining remotely, to take part in the Conference, as well as those who are here in this hall, as well as the speakers and moderators who voluntarily agree to chair the sessions.

Just some reminders, sessions will start at 9 o'clock, but there was just a little bit delay. But we take a morning tea break, depending on when the session starts, 10.30 to 11 and lunch break 12.30 to 2, afternoon tea break, 3.30 to 4. And I hope you have been here since Monday. If you arrived last night or this morning, the toilets are outside the conference hall, either left or right, you see signs from this room.

We don't expect to have any emergencies happening during our sessions, but if it happens, please stay calm and use the exit doors right at the back of the hall or the side doors here. You go straight to the car park if you are exiting this way or if you are exiting that way, you go towards the restaurant and into the lounge.

Our APNIC staff will come and meet you this. This side, we have APNIC staff sitting all over the hall here, so they will take care of you after that.

We have a member services lounge next door if you have any queries, APNIC services, you can visit our member services desk and you can enquire about that or you can speak to one of the APNIC staff located here in the hall.

We do like to follow the eco-APNIC initiative that we initiated in our Secretariat, so if you would like, New Delhi, we will provide a bin outside the hall to recycle those so when the Conference finish, you can take your name tag and drop the badge when you finish.

May I request everyone to turn your mobile phones to silent mode so it won't disturb the webcast, it won't disturb our stenos sitting here and it won't disturb the session as well. I'll give two minutes, please take your mobile phones out and put them in silent mode, please. Even I forgot to do mine as well, so I'll do it at the same time.

There was a lot of feedback from our sessions which captured in the slide here. Do respect that we are different, we have different cultures, different cultural backgrounds in the audience here as well as the APNIC staff. Please be polite when you are making comments. Focus on the subject, not on the person who is making comments. That will enhance the discussions and that will be a positive outcome as well.

If you are approaching the microphone, please don't interrupt the other speakers who are already speaking.

Stay close to the microphone and state your name and affiliation before you start speaking, so that we can capture who you are and what comments you are making, as well as it would be helpful for the stenographics sitting in the front here, you can see that they are captioning every word I speak, so it would help to read and understand what you are saying, what your comments are.

I'm not going to read all the words on the slide here. But just referring to acronyms, we have a lot of first time delegates here who will not understand what the term means, so try to explain what it is so they can have participation in the Conference ^.

We are broadcasting this session, so there will be remote participants joining and asking questions or making comments. Our staff are monitoring the Jabber chatroom. If there are any questions, they will relay it to the hall and they're equally important to us, as you are here, so do respect them as well and their views. Make them be part of this Conference.

So with that, if you have any questions regarding this sessions or regarding the Conference, you can talk to any APNIC staff or myself or you can visit our member services desk.

^ I would like to invite Paul Wilson to come up.

Thank you.

Paul Wilson: Thank you very much, Sunny. Welcome, everyone, ladies and gentlemen, boys and girls.

I say that because I guess quite a number of you might not have been around 20 years ago, at least not in a professional capacity, but 20 years ago, Internet access meant voice phone lines and slow modems and many of us got used to that, but the difficulty in many parts of the world is actually that it was very hard to hear voice line and you could literally wait for a year or two before you got a voice line. In fact, it was kind of a development in some sense the length of the queue to get a voice line. The line was longer in Nairobi or Phnom Penh, you got an idea of what was happening in those places.

What happened quickly after that, in the 1990s was mobile voice and across the world, mobile voice leapfrogged well clear, way over the fixed voice network, to the extent that these days, there are actually more than 6 billion phone subscriptions in the world and fixed lines which are hovering around 1 billion are dropping.

So that's pretty incredible and we actually have an exoct parallel today with broadband.

You only have to look at the world figures and you realise that mobile data is is pretty hot and fixed lines are not so hot.

We have something like 1.2 billion known subscriptions active in the world at the moment and that means that with mobile broadband, that means mobile Internet, of course, and as we know, any growth of the Internet these days means IPv6.

So that brings me to the topic today, which is IPv6 and mobility. It's a critical topic. IPv6 alone is something that APNIC is is focusing on very much in 2012 and for the next couple of years to do our bit to make sure that the transition happens and it goes as smoothly as possible.

APRICOT and APNIC meetings are all about real world experiences, the real and tangible experiences that we all have that we can share, that we can learn about here in these Conferences.

So that, I hope, is what we're going to be hearing from our panelists today, the real world realities of IPv6 mobile deployment.

The message from APNIC itself, from APNIC as an organization, led by the Executive Council, is really loud and clear these days, that mobile operators and manufacturers need to get IPv6 running on their services and their devices, that governments need to continue to provide leadership, more leadership in the movement to IPv6 and awareness to IPv6 and we all need to be working together and rather urgently on that, actually.

So with that, I would like to welcome the moderator to this session, a great friend of APNIC, Mr Martin Levy, the director of IPv6 strategy from Hurricane Electric.

Martin has been a long time supporter of IPv6 in the Asia Pacific region, he's gone well beyond the call of duty, I think, as a vendor and a supporter, a supporter of these events in particular. So I would welcome Martin and ask him to introduce the panelists and take us through. Thanks very much.


Martin Levy: Well, good morning. Wise words about what's going on in the mobile space, but let's focus on the morning first, because I have a few personal comments to make, which will really tell you how wonderful it is for me to be here.

I started the day at -- well, I woke up a little early due to jet lag, we'll ignore that. But I started the day with an hour of yoga downstairs with some other dear friends from the community here and a few that got roped in at the last minute. Why don't you try this yoga thing. It was wonder fortunately. It's a great start to the day. It's a great part of, well, I get to do yoga in India. How cool is that.

To me, it's just case of let's settle things and make sure we understand them.

The topic of the day is v6. You can go through a little bit of a history. We're heard some of it before. But I'll tell you a little bit of mine and show you why it's relevant to that growth in mobile and then where that goes with v6.

I bought my first airline ticket to India in the mid-1980s, much younger, longer hair, backpack, just roaming the world.

It was a different world. Communication was a matter of knowing that someone was going to send you a letter and you would pick it up at the American Express office because you have to travel with an American Express card, that was cool. The thing is that that level of communication, that disconnect, that sending postcards, that hopefully that letter will get to where you are at the right time, is a whole different ball game from where we are now, wandering down the corridor here or in a street somewhere and tweeting where you are, the global roaming, you have a cell phone, you can get a text from family and friends and it's a whole different level of communication and that level of communication is really where we start focusing on on why v6 is so important.

We've heard some of this before, but let's go try a couple of interesting issues out.

First off, you look around at the amount of mobile use, both, well, inside the Conference hall, so just let's say here in India or back in your home country, and you start seeing that it becomes nearly a de facto method of communication for some people. Well, add to that that we are about a year into v4 exhaustion, a topic that is very relevant for this region, then we think about customer growth. If you think about customer growth and the pluses and minuses of capex expenditure, but the reality is that that growth also means addresses -- as fundamental as that.

The reality is that that's where v6 shines, it shines brilliantly in the mobile space. It shines also because let's say you're an operator today and I'll try a couple of practical conversations that I've had in the last few days here. You've got an allocation from RIR, in this case APNIC, you've got an allocation that's a /32 and when I talk to people, I say, OK, what's your addressing plan? We're doing this and that, it's all good addressing plan and you ask an interesting question. Whereas in previous APNICs we would have talked about quite a lot about get your first allocation, you then have that Conference with an operator and say, you do realise that if you are successfully fill up address space and you have obviously the substantiation which in this case the Conference was absolutely clear, the RIR is there to give you more space.

All of a sudden, in the mobile world, even in the broadband world, but definitely in the mobile world, you start realising there isn't a barrier that existed in the v4 timeline, as far as space and therefore as far as customer growth and things change and that bulb goes off and the operator goes, wow, I really can grow. V6, not only from the technical point of view, but something as fundamental as IP allocation, still works and it works really well.

The numbers are huge. We have obviously heard that at previous meetings.

The bottom line is there is a winning combination between mobile and v6. We are going to hear some talks. I want to bring up the panelists and get the day going.

I have three people today, Geoff Huston, who is going to join us, and talk about today's mobile Internet. We have Ramesh Chandra who's going to talk about IPv6 adoption for operators and then we have Hon Kit Lam from Tata, who will be joining us and talking about their v6 deployment, including what they're doing in a mobile sense.

Hon Kit is doing what I did yesterday and it's a very brave thing to do, to not only present, but also to do a live demo. So, you know, hats off for doing that.

Anyway, if you would like to join me on stage and we have seats.


We are going to start off and for those people that want to know, we're doing it in seat order here, just in case you want to know who's up next. We're going to start off with Geoff talking about mobile Internet. Pay attention. I have seen only a few slides. Pay attention.

Geoff Huston: Thanks, Martin. Good morning, everyone. My name is Geoff Huston. I'm with APNIC.

Today I would actually like to talk to you about something that actually made an experience for me, in terms of looking behind this -- the mobile Internet. How many of you have one of these? Oh, come on, everyone, put your hands up. You all have one. Thank you. Good morning. Glad to see you're awake. Isn't it amazing. Go back even five years and you kind of go, who has devices that you can do more than talk to?

We all have mobile phone, very few of you would actually have one of these devices. Go back 10 years and none of you would.

Things change in this world and sometimes this change is so profound that it just disappears and you don't even notice. The classic example there is that water tap. Think about it for a second and the amount of technology required to deliver water down a tap.

Not only is there massive amount of metallurgy and that's a nicely designed tap, but how to capture water, store it, articulate it and move it so that this tap that you turn on, water just comes out.

Did you think about that this morning when you turned on the tap? Of course you didn't.

The only time you would think about it is if you turned on the tap and nothing happened. It's so profound that we're all habituated that the technology just doesn't matter any more.

That's technology. That's really elegant technology.

So let's take that idea, the most profound technologies, the ones that are truly, truly absolutely changing in terms of society are the ones that we actually don't even notice any more. They just disappear and become part of the way we live our lives.

So you all have these little things in your pocket. Interestingly, you also have, most of you -- I think the hit rate is about 95 per cent, you're all staring at your laptops. So why? Isn't one good enough? What's going on here?

Are these mobile devices merely just the latest cool thing? Does anyone still have a Palm Pilot? Well done, sir. Only one left in the room.

Because quite frankly, it was a fad. It died. Does anyone still have a working Apple Newton? Give that man a prize. Give him another Newton.

Again, it was just a fad.

Because sometimes these things aren't that profound. You buy one one year and throw them away the next. The kind of question that's in my head is: Are these iPhones, are these Androids, are these little things in your pocket going to disappear in the next five years and be replaced by something cooler? Or are they something more persistent and profound? Are they honestly going to bed down and simply in five years time, this is part of the way you live your life.

So I would like to put this into context. Don't you love this photo? I sort of was burrowing around for sort of images of every day live and all of a sudden, I found this image of a city in somehow in both Chinese and English, there's this shopfront that sells digital technology computer and communications, evidently that's where we should all be.

I would like to look at this broader question of exactly where we have come from in terms of this industry and this technology and look at the evolution of computing and communications technology and try and put these sort of mobile devices into some context.

So as an industry, as an IT industry, where do we start? What was kind of one of those seminal moments of technology that started to change this?

1946, this is one of the very, very earliest things that we actually called computers. ENIAC. That was actually developed in both America and the UK at the time. These are ripper devices, aren't they? Discrete technology. Those big cables aren't data, they're power. This things consumed enormous amounts of power and was only good for doing numeric calculations. Programming it was a matter of changing some of the wiring as it went along. It was all attended by little men normally wearing lab coats, but he's very well addressed here, attending these machines.

The computer was enormous. It was so expensive that companies could not afford them. The only folk that really bought these and played with them were universities typically with the luxury of the brand from of all places the local military, because these things cost millions upon millions of your local currency unit.

Computing was big stuff. Only very big boys could play.

Let's go forward 10 years. The mid-1950s. Things were actually getting a little bit interesting. This is a picture of Sage. Again, this is still a research machine and you can see that front panel there is a whole bunch of valves and behind that is a thing that some of you might recall but most of us came up against, core memory. This is the real reason why we call memory core, because those thing thills were little magnetic cores wrapped around with copper wire.

The first thing you notice is that the design guys had moved in. This is now sleek and sexy and has little bits of silver all through it and the core member is yellow because it's really important. All of a sudden, the computer was packing a device, something that you desired, something that had design elements imposed on it.

It wasn't in programmable calculator of sorts, the programming it was really quite tough. I think this actually pre-dates punch cards. So in 1954, computing was still a case of get your soldering iron out and wire stuff up.

Let's move forward another 10 years because the next 10 years were dramatic. Not only in clothing style, because that's a very natty check suit, but you'll notice that now he's talking to the computer using a typewriter.

This is a great leap forward. But boot, belts and suspenders, not only does he have a typewriter, but when all else fails, in front of him are 64 dip switches and if all else fails, he can program that manually word by word.

... there are a mountain of them. But what's important about this photo? Is that all of a sudden this guy doesn't work for the government. He doesn't even work for a research institute. He probably works for a company, because this IBM 360 was sold to companies. All of a sudden, computing became part of the business landscape and it became a resource that businesses used and all of the business data was sitting on those silly tape drives. God, why did we ever bother?

Anyway, remarkable achievement. Still costs millions upon millions of dollars and it didn't have its own dedicated building, but it certainly had its own dedicated facility. Air conditioning, power, lighting, everything was built around the computer.

That poor guy worked at the computer probably from 9 to 5, that was his job, to sit there.

So we all attended these computers and did their bidding.

Another 10 years, we really perfected the mainframe and seeing created one of the best jobs ever. This is a picture of the CRAY-1 from 1976. That looks like a really comfy seat. No, that's the water supply system, because these suckers were so fast they needed water to cool them.

Why is it a semicircle? Because this was still discrete components and they still need to make the wires between the boards, but if you laid it all out in long line, the speed of light -- the speed of electricity down the copper took Tong, so we made it into a circle to reduce the amount of wiring.

This is a really elegant photo, marvellous design, crisp, clean 1970s colours, did anyone remember that inside of these machines, kilometre after kilometre of wiring, all inside two cubic metres. It was solid copper behind there.

So, yes, it still looked elegant, but we are still dealing with the same old technology. That machine has about as much power as probably a couple of those.

It certainly after we got on a bit, showed its age.

But 1976 was a very interesting year for another reason. This was the reason. While the mainstream computer market was trying to go bigger and faster, while CR Cray and Gene Arndale chk were trying to make machines that outperformed the IBM360 and went into millions of instructions per second, there was also another model out there. The folk who used to meet at Stanford and other places who experimented with kits and I think it was a Z80, it was certainly one of the early small integrated small circuit manufacturers and these folks put it into a package. It has all the elements of the old IBM360. There is the keyboard, the computer behind that wooden box, oh, what design and of course there is the screen to look at.

But this was something different. Because now it wasn't an asset of the company, it wasn't a business tool. The Apple 2 was sold to you and me.

All of a sudden, computing became a consumer device. All of a sudden, design got -- God, that was an awful design -- was really important. Wood didn't cut it. It's a nice idea, but this one was crap. The next one, by 1984, this guys cracked it. This was an earth shattering changing device.

The keyboard is actually less important than that screen. Because in that screen, is a new way of interacting with computing. Most of the compute power of this machine actually drove that graphics engine, drove that graphics interface and the machine didn't spit back this arcane crap we used to call JCL, job control language, $$?. This machine said hello. This machine talked to you in your terms.

You didn't have to learn about computers. The Mac was learning about you.

This was a major, major change. From then on, the next few years were exciting. They were exciting actually because they also ultimately radically changed the underlying silicon industry.

Because these devices had to do everything in a tiny form factor. You couldn't build warehouses of sort of single blades and chuck them all together. Everything had to be small. The guys who really perfected small, fast and amazing were Intel. I actually think the Pentium processor was again another earth shatteringly good device. All of a sudden on that one integrated circuit and by 1993, I think there were approximately 7 to 10 million transistors on this device, was the entire computer on one chip.

There was a central processing unit, the floating point unit, the buses, all of the interface peripheral drivers, the cache, the lot. The only thing this particular device was missing was RAM on board. But that was it. All of a sudden, that warehouse of dedicated device was rammed into a chip. The chip could operate at temperatures of about 0 degrees Celsius after some of you used to fry them at about 50 degrees Celsius. Wherever you went, the chip could come with you and still compute.

It didn't need a dedicated environment. All of a sudden, at the underlying technology, computing was a portable commodity that just went with you.

So some clever person packaged it up and said, here, stuff it in your pocket. The iPhone was indeed revolutionary. Because it put all of that accumulated technology.

Let's go visual. Let's drop the keyboard. There is no keyboard on this. Quite frankly, any of you trying to write a book on an iPhone would appreciate that the keyboard is not really very functional at all. The entire technology is so small that it's actually designed to be driven by one thumb.

That's the interface.

All of a sudden, it's changing stuff. Because if you look behind it, too, there's almost nothing in an iPhone.

I think it's about 10 chips. It's got all of the functionality and indeed most of the capability of those laptops you're looking at. It's a lot cuter, a lot sexier and it does a whole punch more and the form function is tiny.

It's a full blown Internet device. It's got an absolute general purpose browser, unlike the crappy old mobile phone you have where the telephone companies decided that HTML was difficult and they would give you that shit called WAP. You don't need that. HTML is what you needed. Most of these ^ with almost everything, except if you have an iPhone, for some reason they don't like Flash, the bastards. Get over it. Flash is good. Apart from that, everything on that iPhone is on your laptop. It's a fully functional thing, down to 10 chips. That's this year.

It's only going to get different.

But it's not just the technology that's changing, it's the way we interact with that technology. Because this is truly transformational.

This change is massive.

Because I think across the next five years, society's interaction with technology is going to be massively changed by this.

Most of you work at home and at an office and most of your environments are very similar to this. Your computer sits in its own dedicated spot, an altar of computing. You have dedicated lighting. You have its own power supply. You have your own comfy chair to work at. I'm sure most of you, geeks that you are, surround yourself with as many large screens as you can possibly get on one desk.

What you actually have is your own environment and when you wish to use the Internet, you go to that environment.

I think that's actually going to be a generational definition, that you and I are now getting old. The world that we grew up in was a world that was very physical. You wanted to meet and chat with friends? You went down and had a coffee with them. You want some information? You go to the library.

You need some food and bread and milk? You go to the shops. The activities are physical. We even made using the the Internet physical. You go to your workplace and sit at a dedicated environment and do your work.

When you don't want to do your work any more, you get up and move around. You do something else.

What are these people doing? Working? Going to work? Playing? Can you tell?

They're on the Internet. They're sitting there on a train, they've got radio connectivity. We don't have reliable power, they have battery power. The lady on the left is operating a device, not with both hands, but with a thumb.

She's eminently distractable, because she's on a train. When the train comes to the right stop, she'll stop using this divide immediately and get off and use it again. The Internet is incidental at times, like the gentleman there on the other side, is totally engrossed in it. I'm sure it's some kind of game.

But she is not so engrossed. It's just something to do. So the Internet is now different. It's any time, anywhere. Anything you are doing, it's just there.

And it's trivial. You don't have to go somewhere and sit down and turn on the light and boot up the machine. It's just there in your pocket. Is the taxi driver taking me to the right hotel? Oh, dear, maybe I should look up Google maps right now and find out where I am.

How many of you do that? All of you do now. Because it's trivial, it's commonplace and it solves your problem.

All of a sudden, the Internet is now ... way I think I'll use the Internet now. It's just a part of everything else we do.

So if you think about that massive transformation in our lives, in the way we do things, and then map that back into how many of the world are like us? Are we special? We use the Internet all the day. There are 2 billion others on this planet that sit there exactly the same way. Of those 2 billion, well, in the developed world, around 70-odd per cent of the developed world populations use the Internet all the time. On the developing world, it's a little lower, but as you look at this graph over the next 10 year or so, it's growing as well. 2 billion Internet users.

Interestingly, mobile phones have really penetrated the world since around the year 2000 and in the developed world, a lot of you have two phones. In other words, per 100 population, there are now about 114 mobile phone services.

Even in the developing world, it's up around 78 out of 100, 78 per cent. There are 5 billion of us that actually use mobile phones.

Mobile Internet users, that's amazing. Out of those 2 billion, one-third of them, 630 million, use mobile Internet devices.

That's been a dramatic shift that's occurred over the last four or five years.

This has snuck up on us by surprise.

It's a different kind of usage. These machines are not intensive. They're incidental. They're trivial. They're little things. They're Twitter. If we count the bytes used by the device in your pocket versus the device on your desk, on average, what's in your pocket only uses 1/10th of the data. They demand very high availability, but they don't demand intensity of communication. They're slight. It's a different way of actually looking at the world.

The production numbers are intense. Anyone remember the Digital Equipment Corporation? Long since dead. In their heyday, they boasted to their customers that after 15 years of selling computers, all over the world, they had sold their 100,000th computer. Much celebration.

Last year we shipped 270 million mobile devices. 270 million.

Why? They're cheap. I don't know how much you paid for your mobile Internet device, your iPhone or your Android, but it cost a lot less to make it. One can build an Android based mobile device for under $100 and even retailing costing, the components of the iPhone cost a little bit over $200. They sell for $700 normally. The margins are amazing. They're rich.

You don't even have to build your own software because Android is free. So all you need is the chips in Android and you're there.

You don't even need to generate your own content. Nokia did get it wrong. You don't need to build your own environment of APNIC apps because the web is your app ^. Android doesn't do anything special, it's all just HTML and Flash and JavaScript and all the other things that are on the Internet already.

You don't need to build anything around from it.

What about if I'm making chips? If I'm today's Intel? What should I build next year? A hundred doing great high powered massive desktop chip or a refined version of a processor that uses even let's power and packs into a smaller form function? You guessed it, the volumes are down here. All of a sudden, the entire industry is focusing its major production capability into these lightweight small capable go-anywhere devices.

270 million units shipped is so compelling this industry is now addicted to mobiles.

If you think desktops have a future, think again.

If you think laptops have a future, think again. Because I don't think a bit like the mainframe computer, they're going to be around for much longer, because these things are what we want.

Who's playing? There are four major players out there that are transforming the world. Android and the Android operating system is amazing. This is the UNIX and it is indeed UNIX platform. This is UNIX revived for this century. It's free.

If I want to make a device and I'm a device manufacturer, I can put Android in it for free. Kindle these days had its own operating system, now going Android. Samsung, galaxy lines, HTC, all of those vendors use Android. 90% of users in ... 2011 were Android and the number is going up. Way, way up.

Android, of course, is good enough it's extending into the tablet world and possibly even into large screens, Android doesn't care. You want to run, it will run it. Extremely neat way. I'm not sure about the UNIX vote. Two decades ago, ever thought this is where UNIX would end up, that UNIX would drive the world and it is, more power to it.

On the other side is Apple's iPhone and a iPad. Slightly less, 18 per cent of small phone shipments in 201 is and it's probably not going to get any higher. It's sort of a bunch of folk who truly appreciate mind bendingly good design and eye wateringly high prices and they are eye wateringly high because Apple make you pay for design and that's why Apple make more money.

Because they don't ship quite as many units, but their margins are amazing. $27 million in revenue in 2011.

Of course, there was BlackBerry. I say was because while it was the highest revenue margin product and businesses use Blackberrys not consumers and they pay very high prices. You might remember in July last year when the BlackBerry network stopped. The user experience was mind numbingly bad at this point and their service disruption I think is going to cost them an enormous amount.

I'm not sure that they're going to recover easily from that. It has been a very difficult year.

Of course, Nokia. The manufacturer from Finland, the wonder child of economic transformation in Scandinavia. The biggest most amazing phone ^ provider on the planet a mere 10 years ago. Everyone else struggled to be like Nokia. It was the darling of the industry.

Even now, it's still big, 35 per cent of all smartphones in 2011 were made by Nokia. Long-term future? Not looking good.

Why? Software. A combination of not knowing what to do with Symbian and jumping onto Windows way too late and Microsoft themselves not actually understanding what Windows phone is really all about, haven't proved a very successful strategy so far for Nokia. They are certainly behind the 8 ball.

No matter which way you look at it, the sales projections over the next few years are truly amazing.

2011, 270 million units. This year, a third of a billion. Next year, 400 million. The year after, pushing up to 490 million units easily within three years.

Everyone will have one or two or more.

As you see there, this is proving absolutely transformational. This is from the Wall Street Journal in mid year that middle jumped and their shares soared and it wasn't because you all bought Apple laptops. The Macbooks have nothing to do with their profit. They thank you, it's very nice of you, but that's not where the money is. The money is in all those iPads you also have and all those iPhones that are sitting in your pocket.

The numbers even for Apple are impressive in their absolute growth. 2010, one-quarter. They shipped 8.4 million iPhones. One year later, they almost tripled it, 20 million iPhones in one-quarter.

Even the iPads, you guys just love them. 3.3 million in one-quarter in 2010, tripled to 9.2 million iPads in the corresponding quarter in 2011.

What was their profit out of all of this? Just in one-quarter, of the year, $7.3 billion. Even though the US dollar is looking ailing, I'll still have 7.3 billion of them, thank you very much. That is an amazing amount of money.

But it hasn't stopped. Because they're just recorded their fourth quarter of 2011. The third quarter they shipped 20 million iPhones. In the fourth quarter they shipped 37 million. Even iPads shipped 15 million in one-quarter. This was a remarkable year.

Their total revenue, $13 billion. For a small period of time, on the stock market, Apple was larger than Exxon Mobil, the largest listed company by capital value on the planet, because of this technology, because of mobiles, because of the mobile Internet is totally changing our perception of the world.

So rough prediction of the money. The profit that's coming out of this industry, certainly count the money in the hundreds of billion dollars of dollars. And makes everything else pale by comparison. Whether it's the Android market, the Apple market or even Windows, there is a huge amount of money and momentum in this industry.

So that's the numbers and the numbers are compelling. We want them and everybody is included this that "we". All 6.8 billion of us. Sooner or later have aspirations here and this company, the silicon side of the business, essential wants to provide them.

But between the device and you is a communications network. I want to briefly look at that technology for mobility. The transformational change actually came out of Europe and it came out of ETSI and it was the original GSM stand. It was the GPRS subsystem that basically used spare voice channels which is why you get 16 and 32 kilobits with multiples of a single G ... dedicated it into data. The data was chel prt crap. The latency was like going up to a satellite, 6 hub milliseconds. If there were a whole lot of people around you, you didn't work very well because you needed more time slots to work faster. GPRS was kind of technology proofs, but it was only kilobits a second. Little wonder that the telephone companies came up with that dead end technology called WAP because this was the kind of HTML light.

What about 3G? 3G came along and used a different technology that came out of America and came out of ... CDMA and all of a sudden, you start to get interesting capacity. Because this stuff works at megabits, not kilobits. The latencies are right down. It's actually acceptable performance, way better than modems and actually started to make it look like you're on DSL. It's starting to smell like broadband. These days, a whole lot of these 3G providers are moving to an evolved high-speed access, HSPA+, which gives you a 22 megs up uplink. It does this by this massive amount of signal processing so it's going into the phase amplitude modulation space and uses 64QAM to get there but that's sexy stuff. It now is running at tens of megabits and starts to make your WiFi look slightly tired.

But they haven't stopped there. They have gone full colour and now have gone into 4G. This is long-term evolution. Theoretical maximum speed of these networks, geez, 300 megabits per second. I must admit, you have to violate most of the laws of physics to get to 326 megabits a second. It's not going to get there. Probably around 10 to 12 megabits is all I have been able to get out of mine, but you will have had a similar experience. 10 to 12 megabits a second, anywhere you are, in your pocket, is amazing. That truly is communications becoming invisible.

So the technology is now that mobility can work. What's happening in the industry?

So finally, I would like to very quickly look at the numbers from the address registry perspective. What do we see out there? What's this industry doing?

Our biggest year in address allocation wasn't 2011, it was 2010. We got rid of 250 million IPv4 addresses. We would have bettered that in 2011 but as you probably know, APNIC ran out a third of the way through the year.

But as you see, the industry is addicted to v4. Who got them?

These are the top 10 countries who received IPv4 addresses for the last three years. The number of addresses each of them received.

It should come as no surprise that China is expanding at phenomenal rates, 50, 55 million addresses per year consistently. Don't forget in 2011, we ran out in April.

So just from January to April, 53 million addresses headed into China.

USA has been consistently second, but notice the numbers declining, 38 million, 42 million, 21 million.

Is that demand declining? Or a constraint on supply?

Maybe I should have a look inside and you'll see that there's certainly constraints on supply and it's kind of difficult to tell whether the industry in America is actually getting the addresses it needs and whether this is actually an artificial constraint. Interesting area and probably worth some more examination.

I'd like to also point out in 2011, some very interesting countries that have never been up there before. Indonesia. India is there for two years, 6 million addresses this year, 9 million last year.

Russia. These are, if you will, some of the untraditional centre of the Internet kinds of countries ... many other countries are there. But I would like to zoom in a little bit. And take it now to the level of the carrier, the LIR. The ISP. Because last year, one-third of the addresses, 67 million, went to just 18 countries.

Because there's only 18, it fits on a slide. That's kind of interesting because I have marked with a star the ones that I think are servicing the mobile Internet. NTT Communications I believe has some relationship with Docomo, the mobile Internet folk. China Mobile Communications Corporation, I wonder if they do mobiles?

PT Telekomunikasi Cellular Indonesia. That's a give away. ATT mobility, 4 million addresses. Down there are some interesting ones, Maroc Telecom, Bharti Airtel from India, 2 million addresses.

So the mobile folk are right there inside that entire table of those folk who need addresses. They need them now and they need a lot.

We have run out. We can't service that demand. We stopped doing it in April of last year. If you want addresses, I give IPv6 addresses, APNIC has none other than a /20 (IPv4 ^. 1,000 addresses. How you get that to number billions of new subscribers, I'm not sure. So it's useful to ask where all this goes, where are we headed? Because so far, the mobile Internet has been constructed exclusively using v4.

Even in this region alone, the aggregate demand just in Asia Pacific is for 100 million new addresses each year to support growth. Totally, the world needs a third of a billion.

The last year has been interesting and this industry has actually managed to survive without showing cracks and breaking things largely by eating up consumer stockpiles and starting to deploy NATs.

But this is a long-term plan. This can't work in the future.

I have done a rough guess at demand versus supply across the next four years taking the RIRs number.

Even taking a very conservative view of growth, the red line, the world is going to grow from 300 million up to 400 million addresses per year over the coming years.

APNIC ran out last year. We confidently expect RIPE NCC will run out of addresses in Europe some time around the middle of the year. Probably by next year, ARIN will have the same fate.

The blue line is supply. As you see, supply and demand are tracking. The green line is the cumulative gap.

Within three years, by the end of 2014, we're short by 800 million addresses. At that point, I really start to wonder all the NATs in the world might not get you there. At that point, the credibility of taking short-term solutions and making them span billions size gaps I won't believe you. The rest of the market won't believe you either. You can't do it.

So we all say the answer is 6. We have said that for 10 years. Fascinatingly, those things on your desk in front of you right now took the Cool Aid and believe. All those maps, you are running 6. Anyone running Windows Vista, Windows 7, you have 6. 6 is all over the desktop.

But oddly enough, that's not where it needs to be. It needs to be here. Some of you have Android devices that actually have 6. Some of you. Not all of you. Certainly not all of you, because most of the an droids don't. But run of the really recent stuff actually has v6 in both the ... some of the apps, but by no means all, everyone do something with it.

What about the rest of us on our sexy fashionable iPhones? Surely it does IPv6. Yes, and no. It will certainly do v6 on the wave lab, but it doesn't do v6 on the radio. It will not do v6 over 3G, 4G or anything else G at this point.

It doesn't have it in the radio.

You need it most, it's not there. What about Nokia using Windows phone? It's not there. Not even announced. There is some vague clues that I can see wandering around that well-known authoritative source Wikipedia that it will come in the Apollo release of Windows phone, but it's not here now.

So the desktop devices are ready, but they don't matter. The mobiles aren't ready and they really, really matter.

But it's not just the device. As we have found in the wide world, it's the last mile provider and those CPE devices that aren't doing 6. How about the radio providers? Are they doing 6? How many radio providers in the world? Multiple hundreds. Google, IPv6 mobile providers, I did. I found T-Mobile, there are some pilots under way in the US.

I found Verizon. They made some announcements about their intent -- sound familiar -- to use v6 in LTD, but that's all. I looked around the world. I find Mobitel in Slovenia, I find a network called NWN, I think, in Norway. All the Googling in the world doesn't find me anyone else.

No one. So what about everybody else? What's this industry with growth rates of 300 million devices a year going to do about v6 because NATs aren't going to cut it.

You should get worried.

Thank you.


Martin Levy: I am going to hold questions until all three speakers, for obvious reasons.

I have core memory in my garage. But I also have an iPhone in my pocket.

OK. Let's move on.

Ramesh from Bharti Airtel is going to talk about their v6 experience, adoption, how they have engineered it internally and some interesting statistics, actually, that match some of yours, which is good to know.


Ramesh Chandra: Thank you, Martin.

Good morning, everybody. I'm Ramesh Chandra from Bharti Airtel, responsible for ... transport services. I'm going to talk about IPv6 and strategies and the plans for multi-service operator, like any other operator who is providing the twin services ...

The items we are going to discuss is drivers for IPv6, I think is common for everybody. What are the new service opportunities that have been available to us. What are the present IPv4 ecosystem that we are operating at the moment, what should be the design considerations and the right strategy and plans for IPv6.

What are the options available with the operators and their providers. At the same time, we also support the challenges which are with us while we do IPv6 and also talk about what is Airtel doing in the address space.

The drivers, I think known to everybody, we have a limited IPv4 address spaces available and the depletion of that is a big driver. If we look at the right side, the curves here, the operators have a different side of that. Based on the growth, they're going to deplete in the next two to three years, it is starting delay this year.

This depletion also depends on how many IP addresses they can use, where are they in terms of adopting IPv6 in their network.

Available to network, India has almost 112 million subscribers, Internet users and this is, I think, the third largest economy in terms of that user base.

We have also seen growth in this mobile sector, who are using Internet services, this is almost 89 times, 10 million in 2002, raised to 890 million in 2011. What then means to us, demand in the market, user adopt based ...

Apart from this, there is demand from business customers, so businesses uses IPv4, available, now after IPv4 no more available, they are working aggressively in terms of operators and starting with IPv6 services.

The Government of India has also come out with working groups, different task forces focusing on different things, whether it's architecture, the ... security, there are different working groups and they are focusing on to get the industry ready and start using IPv6. So they are also one of the drivers for IPv6 and of course, there will be some new opportunities where we welcome IPv6.

Some of the report this was published in March 2011, 2,200,000,000 mobile access from the mobile is going to the YouTube, this is very huge number generating like 6 million GB. Similar experience on Facebook and Twitter, the mobile devices are using and accessing these services that are available on-line.

In the new service opportunities, these are available across the verticals. If we look at some of these like ... and there is very good development which is going on and they are doing this year and next year. If we look at this, when the Government of India is doing a focus in developing the help desk services in economy like India and very focus more on the rural sector, this was not and you would agree if other sectors.

Similar within the transportation also, when should be available when need. There is a huge demand for IP addresses in this. These are some of the verticals.

Present IPv4 ecosystem, wherein we have segmented this into the five parts, like there is the user devices, which are used, if we look only at the mobile sector, but different 2G or 3G services or home segment and the Indian network structure, we have DSM or ADSM... these are part of it. Responsible for providing with the data services.

Same case with NLD, and the broadband ...

This is the segment where the operator is managing and Internet service ID network along with the ... systems in place. And there is IPv4 Internet and IPv6 Internet connecting to the contents which are available for the users.

In these segments, if we look at every segment has the potential growth in terms of available users and also investment which is required to make these systems ready.

If we look at the LI system, the LI system is not only for on-line system, it can be used ... this current LI system is used in the mobile centre, LI does not use IPv4 or v6 monitoring the users.

For monitoring, when the same system is going to IPv6, but the traditional requirement of a user in historic terms, if there is any activity which has happened, whether any user, and somebody is trying to access some of the site and then there is a complaint because the activity was not ... in that situation, the challenge to the operator to give the use to the relation.

We will talk more in a day what exactly the environment available the challenge.

When we adopt the IPv6, there are certain business considerations which we are on at the moment. Available to any user, mobile user, broadband user, that user is not worried whether he's assigned IPv4 address or IPv6 address. As long as the service is, he is getting a better experience. So he's not very worried about IP address satisfied with the device.

Other factor which is very important with the first factor, is the cost. Every devices that available and IPv6 support is more expensive.

Operator who wish to upgrade their infrastructure is going to cost money. At the same time, it will any user the user is going to pay because he's going to use IPv6 address and it's more of a challenge, more like the operators is less than. Here is a solution must be more cost effective and it should not in fact extend the longer period. When we adopt IPv6, we need to also take care that network must be scalable, must be more reliable and it should have the capability to monitor end-to-end services for the customers.

When we have the services, of course it will enable new service opportunitieses, but at the seem time IPv4 and IPv6 long-term co-exist. Allow IPv6 adoption should be available seamlessly, co-exist.

Some of the IPv6 design considerations, which are very important to consider, when considering the plans and strategies, like IPv6 addressing plan. The way we have used IPv6 addressing and IPv4 is going to change IPv6, not because of the. IPv4 allocation was so far limited space, it was allocated again and again in smaller pools. So when the operators have a different backbone available at different time based on the business.

In the case of IPv6, there is a need because IPv6 we need to plan that at least for a longer perspective. Whatever growth and opportunity everyone has, how many are going from all services at least for the next 10 years and then the addressing plan to meet the requirement, start implementing, yes, there are different IP addressing, a plan needs to be used by mobile services and broadband services so there is a thought for it to plan for a longer perspective.

Selection of the transition strategies, what is the best strategy for the retail services and what are the strategies which are available for the business services.

Interconnection, yes, the operators working to get that working, so will be get ready for IPv6 services, we need to look at how we are ready in upstream providers and how we are ready with the local peering or private peering or public peering.

Here we adopt IPv6, when we are using more IPs, it's more on the user devices, that is the user plan. The IPs which are used are very, very small numbers. So we announce whether you wart to start with the user plane and transport plane, we start on user plane first and once they are ready at that time, we address the transport planes. So there are different options available, greenfield network, mixed platform, impact on both, existing operator, its doesn't make sense if I have to start with a user plane.. We then address on phase 2 the transport plan once we see even the other systems are ready.

LI is also equally most important item to meet with the compliance for the licence information. Services, yes, any product when we made we have to upgrade the existing infrastructure and at the same time, we are also going to use some newer devices. We clear new devices, yes, we need to look at the scalability that when we study, yes, it should meet the requirement down the line.

The services, what we are going to hear start IPv6, definitely any user that's not expect any degradation on the service. So when we make a choice of the right technology, right product, we need to ensure that the product is not going to create any adverse impact on the end user services. So have the right selection and have right technologies is is more scalable, it is meeting the environment, it is also end to end monitoring, so that there is immediate impact to ensure that the end experience is not deteriorated.

The more important here is talk more in detail is that there are a lot of devices which are available in IPv6, a lot of support, but it's more important that as an operator, we must in fact evaluate, we must test our services, we must see what is the impact on the user, involve any transition like this.

In IPv6 allocation, what we have seen since IPv4, small, small pools, but by default APNIC is allocating /32. This with a /32 is bigger depend on the business growth, but you see this number, the regulations in India is also has a recommendation to make that any IP address space which you use in the network, it should be distinguished by different reasonable regulation agencies. What does it mean? Create that IP pool separate from the infrastructure and create a separate pool for the part a of the not part or east or west and whether that region create the pool size for different category of the downstream service providers or the enterprise houses, other services.

They have a very specific implementation and this is going to mean that put the reasonable ... they had in terms of whether the pools ... this is going to similar few the capability. So from the service perspective, which.

We look at the transition strategies. I think we all aware there are dual stack and there are translators and there are tunnels. There are three major options which are available.

In the translation tunnel.. On the operators who are every segment of this system is dedicated IPv6 and that meant start from any user devices and network operator network and the application, in these three major components are IPv6, there is no need for any translation to, there is no need for any tunnel to be used. Then we just stay and focus on the dual stack, use IPv4 and using the devices and operate in any case contents are also on IPv6. Anything which which do in the translation just interpret it and possibly is going to go away and probably in the next 5 or 10 years down the line.

In this dual stack, yes, whether there are available options, one implement IPv6 in the network, run IPv4 and IPv6, use the NTS capability in the network and carry out v6 ... is yet to adopt available and the operator is including has implemented network so that we use existing IPv4 plan and carry IPv6 capability.

I will continue ... number of any devices available supporting IPv6 is very small if numbers.

If these devices does not support IPv6, then the only choice available to you is IPv4.

Probably IPv4 is no more available, the only option left is use private IPv4.

We have to use private IPv4 then we cannot provide at this intermediate stage the translation. Then we have to do the translation to convert from private IPv4 to the public IPv4 for the contents on IPv4.

The last option, which is the tunnels, it means in the translation which are more relevant than for the operators environment is NAT44 and possibly NAT64. We will talk more in detail and going into tunnels is more of intermediate, IPv6 when you carry IPv4 from one segment to another segment, used 64. I think there is no more focus as per the operators, possibly ... dual stack and there is the need after that existing with the contents and user device is nod ready, you can do translation.

Why is that translation required? We did some study from our own network. The 3G handsets available from 22 megs, about 754 models and surprisingly, what we have seen in that in terms of numbers, 48 per cent of that 3G handsets support IPv6. This is as per the information available on the website. Also realise, although when the handset. It is ready for IPv6, but different applications which are available in the handset not application will support IPv6. So more important to consider even the application which are available in the handset, all applications should be ready in IPv6, that is more important than just giving the number, yes, and specific reason for that, supporting IPv6.

It's a good trend that numbers are increasing and there is a lot of focus more on from the top equipment which are available, devices are available. When we look at the trend, major user base, this is like GPRS services.

To the best of our information, none of the models use by GPRS services support IPv6. If we look at the numbers point of view, the operator there are like 30 million subscriber, 30 million like close to 26 million of GPRS users. 3G has just started.

The majority of like more than 90 or 95 per cent of the devices, 95 per cent devices do not support IPv6, that means we have to ... on IPv6 and we have to take care the translation in the network.

When we do translation, this NAT44 or NAT64 which are more important, but it's very important to consider when we look at the scenario, take care the NAT44, NAT64, NAT64 is a very good option, wherein we can start translating IPv6 when we are looking to the user and trying to access to IPv4, sometimes we take care. But the problem that we see on browser based applications are supported in this NAT64, but many of the Internet based apparitions, whether it is on-line gaming or messaging, the other services which does not work in NAT64. So when we adopt it, it is very important to consider for the application whether access 64. It looks good, but at the same time, we started ... see what is the impact on end user applications.

Operator transition solution all IPv6. Yes, the operators are ready and the contents are there on IPv6, there is to need for any translation of that.

Because this is the graph here, IPv4 pool is available now lately this 2011 where it is exhausted. So whatever IPv4 available, we are going to be used and there is no further growth in this.

The green line shows IPv6 deployment. It's going to just start, already started. But it is very limited at the moment.

Considering what we have on IPv4, we cannot provide at the moment the translation.

This translation, this is the red line they are beginning to start using IPv4 and start the translation which is not very ^ but we have no choice.

The operators offering services for the mobile services, surprise services, broadband segment, what are the services are more important, is it migrate to IPv6? Is it easier to adopt on IPv6? Also, bring the more impact on the user. So when we work on the IPv4, at the same time, we just focus on implementation of IPv6. Maybe possibly the more number of IPs are in the mobile sector, follow broadband services, that should be the most focus and the latter stage, IP services with the dongles that are available for ... IPv6 and start to use those.

Transfer plane, services ... in the private can continue with the mobile.

When we work on this, there are certain challenges, what we have in this IP ecosystem, when number of devices, yes, we have seen in the increasing the operator starting with the IPv6 services, but this is very small in numbers. The market is full of legacy handsets, this is not IPv6. This legacy is further going to grow, because the end user who is making a choice of buying even the handset, very, very few, you just supporting IPv6 or not, but most of them going to what are the features, what is the camera and quality, what is the screen size, so IPv6 is more more criterion selecting impact on handset. If this legacy is available today, it's going to go further. To take care of end users, this translation cannot be provided.

Second is when contents availability on IPv6, yes, we have seen IPv6 contents available and then in that there is a lot of impact.. So good trend has started, but at the same time need to focus so the contents are available at the same time.

We are not IPv4 and IPv6 are private IPv4 very important requirement on regulation point of view address.

In this, since the same public IPv4 is used by multiple customers on different source ports, when it is going to access platform content, if there is need for regulation for the user, unless the old source port is available, operator is not available to identify the specific user. So this is a new change that is going to come and when the availability of the source port from the application is not being supervised, many of the applications they have this option available, but most of them will not availability, so this is also a challenge.

At the moment, there is demand has just started. Achieve ...

In this environment, on the left side IPv4 and right side is when we have the IPv6, it is going to be more complex and have private IP or public. Translation must be when the tunnelling technique and from the user, we have the storage for users. Poablg even more complex ...

Work at Airtel, we are doing we are upgrade it to a dual stack, we are working at this to identify business customers to do the application. We have upgraded the upstream peering on the dual stack. We are also working with local people, NIXI on IPv6 and also with the private operators to have upgrade. Other part is mobile customers and also ... getting ready. CGNAT trialling almost completed for all services. Now with all this platform, also working with the European supporting different.

Supporting platform, so they are offering activity that we engage in and we are trying to do more leading these activities so that we are ready.

What transition is inevitable. We need to collective effort. The operator alone cannot do it unless there is support from the content provider and the user.

Cost of the transition increases with time to start. If you start possibly we can test application and then.

Develop your own priorities and transition mechanism. What is appropriate for one operator may change to the other operator.

Service POC is very important, you must focus and do it in that for impact on user applications. We have seen in IPv6 definitely not Y2K, but impact here is much bigger, that was not Y2K.

Thank you very much. Happy to take any clarification of any of the items.


Martin Levy: Thank you very much. We'll hold questions to the end.

The third speaker, which will be Hon Kit Lam from Tata, and Hon and I crossed paths many years ago, so the v6 focus is as I know pretty solid there.

Let's hear about deployment and mobile side.

Hon Kit Lam: Thank you, you can call me HK and I'm based in Hong Kong, so a lot of people call me Hong Kong Lam, because I'm from Hong Kong.

I know the time is limited, I can do it very quick and but I will skip through some slides, a lot of things you know already, I won't go through that in detail, but I will go through some key messages and I agree with Martin, it's it is believe to do a live demo, but I would like to do a live noble v6 demo, just very short, at the end of the live presentation, and because we need courage to step into the future. That's what I believe.

So I am going to start.

Obviously, I'm not a Mac user.

I know a lot of iPhones here. How many of you know, I still need to ask, how many of you have v6 and v4 in your laptops? Can you put up your hand? How many of you -- keep your hands up. How many of you know how much traffic you use in v6?

OK. So less than 1 per cent of you is more than 10 per cent. So it's very, very tiny numbers. That's the challenge as well.

Here the chart shows the populations against the IP address of in India. It is around 1.22 billion population, but less than 20 million IP address. So less than 2 per cent IP address for the population.

Surprisingly, the wireless population is very large, more than 900 million. So the issues of allocating addresses to each individual device is a challenge.

You can see the mobile penetrations, the ratio for the global you can see from representation is 2 billion to 5 roughly, fixed to mobile.

But in India, it's around like -- sorry, I should say it is about 32 to 900. The wireless doesn't mean mobile, but also includes like WiFi, which maximum access, but all includes it together, it is like more than ratio in the rest of the world. So that's a very key for us, in terms of India, to drive the v6 and I will show in our sister company, as a mobile arm of Tata group, how they do to and what is the status and what is the challenge.

I won't go through detail word by word here.

Here is the challenges of the v6. We all know about well we don't have IPv4 address enough, the transition technology, capex, where is the money coming from and also the NAT solutions don't allow some applications like peer to peer, that we all know.

Making worse is not all the content we all know about that, too, not all the content is supporting the v6 and v4 together. So a lot of translations need to be done and there is no additional web.

I can tell you for Tata Communications, we don't charge premium for the v6. We charge the same for v4 and v6. I think most of the people do the same, so there's no additional and even for the mobile, the v6, we don't charge the premium, but we need to increase the populations of the access to then users.

I don't go over that, we all know well all that. I will show later on our TCL global and also India, the v6 turn up like per customer, the percentage of our network, how much of the customers already use that already.

The v6 drivers for the machine to machine, extended service, these are also the drivers, but I will let you -- you already know well about that.

The mobile machine to machines like remote monitoring, a lot of different applications you need addresses, people talking about every device have IP address, these are the applications in the future.

But we are facing the current situation in the next step further.

What we talk about, how to enable it before we can use it. We have 2.5G and 3G network. Upper part is voice lower part is what we turn on the v6 network for the mobile and then this is the Tata to service and we connect v4 and v6 into TCL and upstream to the Internet.

Tata teleservice, TTSL also have CDMA as well as the GSM. So we can see from here, the P DS M support the dual step, transport network support the dual step and we test already the dongle. That's what I will show here. That is we have the dongle connected to the VC and you can get the v6 address, it is like a native v6. You can find anywhere for the v6 site and I will show later on in the demo.

However, everybody knows the same v6 enabled content is 5 per cent. This is the roughly the traffic ratio that we observe.

Similarly for the GSM, the GGSM also support the transport layer and we support the mobile phone as well as dongle using Windows 7 obviously that is the similar for dongle and also the mobile phone for the GSM can support v6.

Here is the challenges we face for smartphone. We know, like well, that the unique price of smartphone in India, the demand is quite relatively low. So the percentages you can see is even lower in terms of readiness for the v6.

For the CDMA, we also need to specific upgrade for some of the chipsets to enable that the v6 supports. As well, there is content not much content now, it's much better, I can tell the IPv6 improves a lot. A lot of sites have their awareness, but the traffic that I can see from you guys raising your hand, that the population for content and also the driver for the v6 traffic is still very low.

We have dealt with the v6 content, we talked the same, we need the NAT64. There is some issues we can do that, but there's still some issues as we all know about the some application does not support that and making it even difficult, because we also have some regulatory requirement to monitor that for the security reasons. So the NAT64 still need further development to make sure that we can comply with the regulatory environment in terms of ...

Also, because of the applications, it will be only good for the web browsing in terms of 6 to 6 like a version 6 to version 6 web browsing.

So with this as I mentioned, of the current status, India status is like most of the major or big ISPs are in for v6, which means their network is ready for v6 transport, but the penetration is less than 1 per cent.

Key reasons, because as I mentioned, that the readiness of content as well as the also like the infrastructure is not all the access infrastructure is v6 ready, although the transport and the core is v6 ready, but the infrastructure is another key element, so that's why mobile plays a very important role into that, so once you have the dongle, you get the WiFi access or you get the mobile access, then you can access to v6, so that is very important for all of us.

Here is also the IPv6 status. We offer IPv6 service in global network since 2003 and 100 per cent of the pops are v6 ready. More than 47 is like a percentage, you can see the red line. More than 47, 50 per cent of our customers in global network is dual stack turn up.

For India, we offer our service of v6 in 2007. Also, 100 per cent of the pops are v6 ready. We connected to the NIXI over the v6 as well. A lot of our top ISP customers turned on all the v6 as well.

The green line is the India turn up rate. The red line is the global turn up rate. You can see that the trend is slightly increasing. It's quite encouraging, but the traffic still very honest, relatively low. We need to drive that.

I won't go over what we participate in World IPv6 Day.

We also been part of the role together with the India government, the IPv6 Taskforce and there are multiple streams of the work group and TCL, Tata Communications is part of the -- is leading one of the task force which is prepared and also educate the industry together to push the v6 network migrations, technology sharing, in terms of the policy and also the likely environment in terms of within India.

So we have a long way to go, although we have all the infrastructure going to be ready, but we have still have a long way to go, in terms of the content, some of the technologies.

So I think the way that we move forward together, sharing the concept, we are sharing that for counsel, I mean, quite a number of years in terms of the v6, we have to make in terms of instead of showing the PowerPoint. That's why I want to share with demo together with the people here.

I will share the demo over switching to another laptop.


This laptop connected to a Tata teleservice, photon dongle, which is a USB dongle, CDMA.

You can see that the interface and also the -- I want to see the IP address. Yep. So this is a pure v6 interface IP allocated over the Internet. Can you also show the website? I know the reception here is not extremely good, so it may not be very, very quick. The speed may not be extremely fast, but when you also think about you're on the road, travelling on a busy India traffic, stuck in one hour, two hour in the cab, if you have a dongle, you can access the Internet, doing everything you want, then the congestion or the traffic jam may not be too bad.

This is the Tata Communications one. The other one is from Google. This is not from the WiFi in the premises here, it's all from the mobile dongle, so you can access on the road.

So I will finish my presentation here and I will take questions and I also welcome if you have any questions, we can talk later on as well.


Martin Levy: Thank you.


Martin Levy: Thank you very much.

I welcome questions, but as they say, I have the microphone first, so I get to ask mine first.

Actually, I was quite impressed by one point, and I want Geoff to talk about just a little bit. Your data, your graphs, are an absolute staple of v6 presentations these days.

Just as a slight, you know, question, how good is that data been as you have looked back? I know that we have been seeing data from you for quite some time and you have come up with some interesting projections for 2014, you have 800 million shortfall of v4, we have to go back and really work on our business plans using that as an absolute number? What's the spread on that one, based on your past experience?

Geoff Huston: Thanks, Martin. There is this sort of law in statistics called the law of large numbers. And you and I might think as individuals we have this thing called free will and that I can't predict your actions, because you could do anything in the next few seconds.

But, you know, when we get a whole bunch of you together, a few million of you, you're actually really, really predictable. A whole bunch of folk use that ^ as an industry, we are predictable. You actually look at some of the key things about the way we perform, the number of new autonomous systems that hit the Internet every day, it's 14 ASs per day and has been for five years and longer.

Oddly enough, our industry actually works quite predictably and quite mechanically, and that the numbers have been rock solid in terms of consistency for over about 10 years. We have gone from modem based networks to broadband deployment, now to mobiles.

Interestingly, the numbers have been solidly reliable throughout that. There have been very few catastrophic surprises.

So I would actually say as the years go on, I'm more and more confident that those numbers are actually a deep and useful insight into precisely how this industry is behaving and that the only thing I could basically say about the slide that's behind us now, that red line, the growing curve of demand, could be slightly under-projected, but the blue, supply, very clear. We understand that. That shortfall number, within three years, we're basically a billion addresses short, I would actually have a very, very high degree of confidence.

When you think about that billion, think about precisely how good NATs are in address compression. If you think that a NAT transition strategy is going to last longer than a couple of years, maybe you're backing the wrong horse -- because it might not.

At that point, all this money and effort you are just spending with your fantastic NAT based strategy, might not work very long. In fact, you might need a new strategy before you even get the last one deployed.

I'm confidence in those numbers, Martin, because I said as an industry, we now, I think, understand ourselves remarkably well and we are, because we are a big industry, the law of large numbers applies, we are predictable.

Martin Levy: OK. We are predictable.

I have one update that I shared this with Geoff also. I showed him the picture of one of the Verizon phones with v6 running off their LTE was pretty much the day that was the Verizon Thunderbolt. I have been told time is up, but I'm going to continue talking for another 30 seconds.

What I was going to say is that was a working example of v6 on the line.

I have a lot more questions. I have been told I don't have time.

You may have questions. The speakers are out here and let's grab them during the break.

I thank you very much for this morning. It has been a great start to the meeting and keep in mind it keeps on going.

Tea time. Thank you.


Sunny Chendi: Thanks, Martin, and thanks to all the speakers.

We'll have the tea next door in the demo area. We'll take a 20-minute break, instead of 30 minutes. Meanwhile, the APRICOT sessions will start at 11 o'clock, have already started, so if you would like to join those sessions, you are free to go, join the sessions or you can stay here for the next session. Thank you and we'll see you back here in 20 minutes.

Key Info


Ashok Hotel,
New Delhi, India


21 February - 2 March 2012


Open now

Program includes:

Technical workshops, Tutorials, and Conference streams.