IPv4 in 2015 session transcript

IPv4 in 2015

Tuesday, August 26 2008

0900-1030

CULTURAL REPRESENTATIVE:

I just want to welcome you all on to this land, which has many, many different names. The most common name means the place of the sacred green stone, the stone that we all of this place wear with us as our guardian. The sacred green stone itself travelled here before people crossed the oceans, from a land called Hawaiki, and this put itself in the rivers on the west coast of the Island and when our ancestors came here on the many waka, the many canoe which followed, they sought it out to gain the permission to live on this land. The next name that I called about was Te Waka a Maui. Maui, when we give that name, it is a coded name that refers to all of the greatest navigators that existed in Polynesia. And we talk of Maui which stood on this land and fished up the North Island which is Te Waka, so from the people who talk that story, I also welcome.

The third name which is the oldest name is a name we refer to as Moke which means "to come close to the side", because when the first ancestors landed, came here, they didn't land, they circumnavigated the Island to make sure that they weren't going to be treading on someone else's territory, so they stayed very close to the Island.

The name of this Island, Moke, is the same as the Island in the Cook Islands called Moke and it is a very old name. My ancestors are from the oldest people through to the most recent tribes which have come to the South Island. And so, I welcome you, not just to the South Island, but to Moke and Te Waka and to initiate your conference.

It's very important, I want to speak more English because most of you understand it, but it is important for us to acknowledge the people who welcomed you before my words and the first is our young warrior who came out. What he was doing was testing your chiefs over here to make sure that they had the metal and the bottle to handle coming into our land. We're a warrior people, so therefore we always challenge each other in a warrior manner, so for those who think that that might be a wee bit rude, it's not rude in our culture. In fact, it's the most high respect you can give is to challenge the guest. So that they can prove that they are worthy of coming. The second person to acknowledge is our first spoken voice that you hear who calls out to you. The position is a very sacred position because that person is the most vulnerable person in the welcome ceremony, so we acknowledge her, and lastly of course, we acknowledge our people, the ones who have come to support you at this welcome.

Feel welcome, not once, not twice, but three times.

PAUL WILSON:

TRANSLATION: Greetings to the people

Greetings to our many dead

Farewell our departed. Go to the true home. Go to the real home of our father. Farewell, farewell, farewell. Thank you for your greetings to us this day. I am happy and pleased that we have gathered together today. Therefore, I greet you one and all. Greetings to you all. Thank you very much.

NEW SPEAKER:

OK, before we leave you today, our group are just going to offer you three more songs, two or three more songs, just for entertainment. The formal part is now over, but just before we do that, we're going to invite our chiefs to come over and hongi. To us, hongi is very sacred and a breath of life and in the south, we do it by pressing our nose and foreheads together twice. So we'll just do this and then we'll hear the songs.

Again everybody, we're now going to hand back to Paul who is going to organize you all for the rest of your day. And I believe that they're going to need five or ten minutes to set things up on here, but I'll hand things over to Paul.

*APPLAUSE*

PAUL WILSON:

Thank you so much for that wonderful entertainment. So, what more needs to be said. Welcome everyone. Thank you for being here and such great numbers and being here for the APNIC meeting which is about to kick off, so as Mick said, we need to get started here, so if the first panel can start getting ready for the preparations to undertake. So, Jonny, are you ready. We're going to go from something which is quite historic and cultural and ancient to the first session of the meeting today which is something rather less real, but still an immediate and future issue from the past to the future, so we're going to be hearing now a hypothetical session about the future of addressing and IPv4 and IPv6 and we've got a panel of very interesting individuals from hypothetical places which will be led by Jonny Martin. Jonny, are you ready?

JONNY MARTIN:

I'm Jonny Martin for those who don't know me, I'm with InternetNZ, who are hosting this depleting here. So, hosting this meeting here. So before we get started, we have a couple of house notes. So housekeeping notes for today, the Helpdesk is located in meeting room 7. So, if you have any queries, do go and visit them. The prayer room is located in the Star Room which is just off to the back left of you beside the hall. Lunch is provided and is located at the Crowne Plaza Victoria Cafe, that's the main eating area in the lobby of the Crowne Plaza. For those of you that aren't staying at the Crowne Plaza, there's a walkway which goes across to the Town Hall, you walk around to the right a little bit and another walkway that takes you across to the Crowne Plaza.

If you get lost, just follow everyone else.

Hostmaster consultation is available for booking via the Helpdesk. Tonight we have the Cisco Opening Reception at 7:00 in the Town Hall. So Cisco welcome all meeting delegates to join in this fascinating experience. The Town Hall here in Christchurch will be playing host to a cocktail reception and a pipe organ reception, one of the largest pipe organs in the southern hemisphere.

Then on Friday night, we have an informal dinner. We're going to be heading to a restaurant with the best of Canterbury ingredients, not sure what they are while taking in the extensive panoramic views over Christchurch. Transport is provided and there is a cost though, it is $70. So please register before Friday at the registration desk if you wish to attend that. So, that brings us on to today. Firstly, I'd like to thank our sponsors for today, DotAsia, Catalyst and our supporting sponsor, v6. Now, on the wireless network, some of you might have noticed it doesn't perform that great in this venue. There's been a number of issues in the 2.4 gig spectrum, the piece of spectrum that standard wireless LAN cards use.

Those of you who have 80211 capable cards, you'll see a couple of SIDs there, A-APNIC, an extra "A" at the start. If you see that, you'll get a better experience. That said, if there's any wireless gurus out there, there seems to be a bit of interference, have a look around and I'm sure you can make it worth your while if you can come up with a solution or at least a way we can mitigate it a little bit. So, we're very sorry about that but it seems to be an environmental problem.

The program for today, we have v6 - It Works For You. You've seen the news coverage about moving to IPv6' come here to experiment with... oh, that's not even today, that's Wednesday! That's tomorrow. So, we can come here and hear about IPv6 and more importantly, experiment with it so we'll have the chance to work with it and we'll turn off IPv4.

Then we have APOPS where you can learn about hot topics like the Asia-Pacific region, followed by tomorrow night, the Vocus Cultural Night. For that, we can experience a might of Maori culture with a traditional performance and a hangi. For those of you who don't know about a hangi, it's a feast with a hole in the ground and transport will be provided and I'm sure there will be more details on that.

I think we'll get started and we'll need the panel up here to get started. They should all be here.

OK, for start, we'll introduce the panel and first we'll have a bit of IPv6 and then IPv4. And the people we have up here, let's start. We have from the fictional country, Pacland, David Woodgate over there. He's our ISP operator from a large developed country with high Internet penetration.

The other character we have from Pacland is Frank March on the end, a Government regulator.

Next country we have is a place called Asidonia, a high Internet world with a low penetration. We have Gaurab who is our ISP operator and Tulika our Government regulator, so a scenario there, just a different sort of country.

We then have Anime, a small Internet world with high penetration, we have Akinori, our Government regulator there.

We then have Olympia which is a large developing country with lots of addresses but we need more to meet the needs of the billion inhabitants there, so from Olympia we have Mr Lee and then, Geoff Huston who has recently moved to the OECD.

So, a bit of an introduction. For starters, this morning and the next session is supposed to be light-hearted. We're after a little bit of light-heartedness up here.

To set the scene, back in the beginning, this is going back, you know not so many years ago now, a few geeks linked in together and they were playing on Facebook and Bebo and MySpace and all of the other exciting things which I seem to keep getting e-mails about joining. They created an Internet Protocol with four billion addresses which seemed like a limitless amount of addresses back at the stage it was developed. Then of course, we had other people who wanted to become our friends which was pretty cool for a while - it made us feel like we had lots of friends! But then, so many other people started getting linked in that we were starting to run out of v4 addresses to give to the friends and the friends of those friends and the friends of each of those friends and I guess you get the picture!

Now we're starting to realize that perhaps we can't get the whole world linked in via IP version 4. So we could all move over to another social network, you know and we could use IPv6 for that. Now, I've heard that's going to be massive, amazing, you know! Awesome! But the trouble is that when I try it, there's no-one there! It's a pretty lonely place to be is IPv6. So, here we find ourselves in 2008, approximately three years away from the end of the IANA free pool, trying to figure out how industry users and governments can work together through all of this and find ourselves linked in with IP version 6. So as we heard just before, we have four countries here who are aware of the v4 depletion issue and they're going to be discussing this with us today. You might not have heard about the countries but I can assure you, - well, I can't assure you they're real, because they're not! But they might be a little bit familiar to you anyway.

Back to Pacland for starters; this is the large country with a lot of wealth and a very high Internet penetration rate. The Pacwoman and Pacman, they all loyally subscribe to the largest ISP in the country, BigPac. So we've got the general manager here of BigPac who is going to make their position, David Woodgate, and of course, helping to advise the Government of Pacland is the chief regulator, Frank March. Welcome to both of you.

In Asidonia, near Pacland is a large country, more and more of the population is becoming middle class and they want access to the Internet, so Gaurab here is trying to juggle all of that as the manager of the largest ISP, Startupnet. Overseeing the IT developments in this country is Tulika Pandey and she advises the IT minister on all Internet-related issues that are critical for the country in the future. Thank you both for taking the time to come and talk to us.

Then, a little bit further north in Anime, we have Akinori, a small and well- developed country. They made a lot of money on documentaries about their everyday life and exported these to the US and Europe in cartoon format. That surpassed Hollywood in value! So, Anime is currently growing very rich on the proceeds and Akinori here oversees the NIR there and he's worked for various ISPs there. And the last country we have is Olympia, and we have Xing Li from there. He's from the Olympia country and it is a large country with more people connected to the Internet than any other country. The starting issue is more people who aren't connected to the Internet, and our final guest today on the panel is Geoff Huston. He used to work for APNIC and he's now the chief economist at the OECD on Internet issues.

We'll start with Geoff and get an idea of what you're all thinking about v4 depletion. So Geoff, before you joined the OECD, you worked in the Internet world. When did people start to realize that v4 was going to run out?

GEOFF HUSTON:

Gee! You know, some people here weren't born when that happened. To memory, I think we figured out that we were going to be victims of success in about 1989.

The first time I was aware of this was an IETF meeting in Santa Fe in 1991 where there was a road group routing and addressing because even back then, when there were about three cats and a dog connected to the Internet, it was actually - the way we were managing the address plan, we were going to run out of class Bs, remember class Bs. Someone must! Andy, class Bs, remember! We were going to run out of them by 1995 which seemed a bit grim and at the same time, we had just passed the magic number of 20,000 entries in the routing table, which made all the Millnet routers collapse so they turned off the rest of the world and only spoke to America which was fine for them. But we figured that that wasn't really the way to go.

It was almost 20 years ago now that we figured out that there was this problem coming and that we should do something about it. So, if the real question was, how long have we been working on that - a lifetime would be about the right time.

JONNY MARTIN:

All right, so why is the OECD so concerned or starting to get worried... maybe "concerned" is a harsh word.

GEOFF HUSTON:

There was a time when AT&T employed half a million people. The communications sector is possibly the most valuable sector on the planet. These days, you can't make a dollar or lose a dollar without a network being involved. Communication systems underpin the global economy. And quite frankly, if we stuff this one up and we're back to smoke signals, you can say that for a whole lot of other things as well. So, for us, trying to make sure that the communication industry is stable, that it actually manages to underpin business and governance, that it actually works from day-to-day is perhaps one of the most critical things that we see out there.

So, when we see some degree of uncertainty in this business, when we see extremely large enterprises, starting to wonder what's going to happen in two years time, then naturally, when someone walks up and down the corridors of the OECD saying, are we having a crisis, some are saying, maybe we are and maybe this is not a good place to be.

So, you know, yes, we are deeply interested in this because communications underpin everything these days.

JONNY MARTIN:

David, you've been working at BigPac for some time. As a long time network operator, you've been aware of v4 depletion for 10 years or so. What do the suits think of v4 depletion? Are they concerned?

DAVID WOODGATE:

Well, they've been noticing the press announcements coming out of the OECD and various articles in newspapers, and you know, starting to see governments and agencies and things starting to ask for some response to the v4 run out. So, some attention is starting to be paid for these things.

JONNY MARTIN:

And businesses obviously tend to plan for the next quarter - quite a short-term focus. When is BigPac going to look at this as a serious issue over the medium to long-term?

DAVID WOODGATE:

I think BigPac is certainly considering all these matters quite seriously now, but obviously, there's a question of priorities, you know, with Pacland having the national game of multi-user Pacman! And Government tender rolled out to connect everybody to that game, you know, it's a question of priorities to be considered in that scheme.

JONNY MARTIN:

OK. So, Frank, we hear that the people of Pacland can't even spell IPv4, let alone IPv6 which is quite a bit bigger. IPv4 depletion is an issue though, isn't it? Well, is it? From what we've issued here?

FRANK MARCH:

Well, as a Government IT advisor, it's one thing to advise Government - it's something else for a Government on a three-year electoral cycle to take notice of the advice they receive from officials. On the scale of things, health, transport, education tends to be regarded as critical issues - something which is slightly over the horizon is a bit difficult for politicians frankly to grasp. As you said, IPv4 is not exactly - it doesn't exactly resonate in the public mind and understand why IPv6 would be there. So, our Government has taken the view that in the - "It's another fine mess you've got us into" and leave it to industry to struggle its way out. There is though a dawning realization that if Government does actually have quite a strong influence on the way they behave, not just in the way of the big sticks, there has been a move being undertaken below the political encouraging leadership.

David in fact alerted to the fact that some Government departments go together with the IPv6 addresses and it is very slow. I can tell you that we've tried to put some of our networks to get things moving but we've come into some problems, because oddly enough, although it seems easy at a network level to understand IPv6, in fact at the applications level, there seem to be all sorts of various obstructions, so getting IPv6 working across the network is a difficult task. So we need to plot out a road map frankly and we're beginning to work on that. But it is slow and we're beginning to understand the nature of the road ahead in terms of the difficulties it's going to provide.

So, we're going to be talking to David at great length about what we want to encourage him to do, but we're a long way away from forcing anybody to do anything.

JONNY MARTIN:

So, Gaurab, you've been hard at work building the ISP in Asidonia and naturally you would want to continue that growth and sustain that growth in the next few years. I guess the problem for you; what if there is no space within three years?

GAURAB RAJ UPADHAYA:

We'll think about it after three years.

JONNY MARTIN:

What does the Asidonian Government think?

TULIKA PANDEY:

We are waiting to see the time before we make a decision on that.

JONNY MARTIN:

How long has the Government been aware of this issue?

TULIKA PANDEY:

Well, the Government has been talking to the advisors for quite some time, a little more than three years now. Each time that we have put up our situations to the decision-makers, they've pointed back to being comfortable using that so then we bring them back to the applications tomorrow or the depletion of v4 would put us in trouble and move industry to demand more policy change from the decision-makers. That's how we see that.

JONNY MARTIN:

So, coming back to Frank's point, obviously we've got priorities with health and that sort of thing, hospitals, fridges, water, electricity. I guess the priorities at Asidonia are somewhat different at other countries. Whereabouts would the v6 fit on your priority list?

TULIKA PANDEY:

I'm sorry?

JONNY MARTIN:

So compared to health and roading, water, electricity, obviously v6 has to fit in the priority list, but I suspect it is not very near the top for you?

TULIKA PANDEY:

When we are looking at it on reaching out to more social issues like education, that's an important factor where we can convince the IPv6 would be good. Then the larger v4 spectrum with 3G and other demands will affect v6 demand in the future.

JONNY MARTIN:

So you see connectivity here as a social issue.

TULIKA PANDEY:

It is a social issue, yes.

JONNY MARTIN:

So, Xing Li from Olympia is pushing for v6 as well. The reason, to show case the v6 activities at the spectacular Bird's Nest Games. Is Olympia ready to give up v4 completely at this stage?

XING LI:

Olympia is a lot of v4, but potentially there is not enough for v4 for the entire Olympia population, so actually, as education and research are going, we can realize that the v4 depletion is actually back to even before the funding of the APNIC, actually back to 1993 while working on Olympia. Like in we tried to get the Class B address but we cannot because that's the time that no more IP class B address were there. Were assigned and we realize that we try to get it back to 1998 so that's a big, big effort and recently, we did some trials for the Games in NAT using IPv6. The mixture of the good news and the bad news. The good news is that we can actually use it. I don't know if you realize actually, the official Olympic website is IPv6! The first time. That's the good news. And the bad news is because we are hosting that, and that the traffic is very, very, very slow compared with IPv4, so that's a very mixed feeling.

JONNY MARTIN:

OK, so it appears here that most of the people are aware of the depletion of v4, which is pretty great. But, it doesn't seem to be the top priority for anyone which is I suppose quite obvious when you're running a country. So let's see what the panellists are thinking and what they're planning for in the period after v4 exhaustion? We'll start with David here who is in a lucky position I think. His ISP started a long time ago, BigPac and got a big chunk of space from IANA back in the early days. You seem to have the addresses you need and the IP market in your country is pretty saturated. So the question here is what you're planning to do after v4 depletion and relating, are you ready to move to v6?

DAVID WOODGATE:

In terms of planning what to do after v4 depletion, obviously it's a question of managing the resources that we have firstly and moving towards common industry position. Clearly, we think that the best thing to do is to head towards IPv6 and incorporate plans for that going forward as part of that we have to consider all of the options and business opportunities that are associated with that.

Sorry, what was the second question?

JONNY MARTIN:

Are you ready to move to v6.

GEOFF HUSTON:

What was the first answer!

DAVID WOODGATE:

I think the first answer, Geoff, was that it is a question of - once v4 runs out and it depends when you're talking about it, the industry or the company level, it's planning to manage that run out and we're planning to adopt and move it to the next phase of the Internet.

GEOFF HUSTON:

Are you saying that you manage a car crash?

DAVID WOODGATE:

Well, I hope this is not going to be exactly a car crash! If this ends up being a car crash, then I think we've all done very badly in this room, not any particular provider!

Clearly, we can see it coming and even if it may not be the delivery issues of tomorrow, then it is something for the road maps at this point. We need to head forward and you know, make sure that we're considering all the various options and contingencies that arise.

JONNY MARTIN:

So, how important are network address translations in your future?

DAVID WOODGATE:

Well, as I had said, we're considering all the options and obviously that's one of the contingency possibilities. You know, I think it would be great if there were a magical switch in the sky which, when we all puts our hands on it and put it together, it would turn in and magically make the Internet IPv6 overnight.

Naturally we have to consider how the management of our own address spaces, you know, what impact that's going to have in terms of the existing v4 address space and can we extend that usage through implementation of NATS to some extent while getting to the end goal of IPv6 deployment.

JONNY MARTIN:

Have you got a sense of additional budget you're going to need for this work?

DAVID WOODGATE:

Clearly it is going to be a lot of money, but it is all part of the planning really. At the end of the day, I think we all acknowledge that the reality is that IPv6 is a life cycle upgrade. We've reached the end of life of IPv4 and when you've reached the end of life of something, you've got to plan for the money to move to the next bit of equipment. I think we're at that stage with IPv6 and I think the money that goes along with that has to be considered as part of the standard operational planning.

JONNY MARTIN:

So, it's all considerations here? There's a lot about planning for these v6 and a lot of talking about v6, but is it actually going to happen?

DAVID WOODGATE:

I think for the sake of industry, I very much hope it is.

JONNY MARTIN:

OK! So, Frank, from the Government of Pacland's perspective, you sat down and summarized 20,000 posts on the telecom's gossip blog, WorldPac. And the forums seem to be concerned that BigPac is talking a lot but there doesn't seem to be a lot happening at least from the outside, so the minister of IT in Pacland wants to respond to this. Given you're a very deregulated telecom industry, what can you, the Government do to convince some of these players that they should perhaps be doing a bit more?

FRANK MARCH:

First of all, I've been taking copious notes from David's comments here, so I've taken due note of where our major telco is heading with this. The advice to Government would have to be that we're not very impressed. That our major telcos in fact were prepared for what is clearly a rather, a looming, not disastrous but problematic position.

I got a comment on e-mail the other day saying, "Is this just another Y2K scare?" and it is an interesting thing to add up. My answer is no, it's not. First of all, it's real. It's not a midnight on the 31st of December 1999 issue - it's a slow crash. It may be a car crash, but it is a slow crash and we do have opportunity to plan. The opportunity is running out but the comparison with Y2K is not appropriate.

Getting back on to the script again, we've been taking note for a long time, the experiment in the deregulating telecommunications environment has not been an overwhelming success and we've taken rigorous steps recently to ensure that the dominant telco in fact starts to put itself in a position for open competition and align with the needs of our growing economy in the way we want the economy to go.

So, we're getting used to the idea that regulating telecommunications actually isn't such a bad idea. The only question in theory is, would regulation actually assist with re-allocation of scarce IPv4 resources? Governments have to be careful when talking about regulation and we took careful note of the fact. Our minister in particular took careful note of the fact that advice from the OECD is that regulation is inappropriate, at least at this stage. Leadership and encouragement is more appropriate

GEOFF HUSTON:

They're very glad that at least someone wasn't in the bars in Korea. Thank you!

FRANK MARCH:

Our minister took careful note of that. So, where does that leave us? I think the major issue actually is not so much dealing with our economy, but dealing with the depleted situation of IPv4 addresses because that will probably tend to work itself out in one way or another. Just as it has done for the last 30 years which Geoff mentioned - 20 years ago when they had to move away from the allocation of B class addresses.

The real problem I think for our economy is going to be the fact that the opportunity cost of NAT and other work around interfering with the innovation that can occur or should be occurring on the Internet and isn't because of the way that that depletion of addresses, the work around are interfering with the Internet model. So, that's our major concern, is that we move to IPv6, not by regulation, but by leadership. Leave industry to exhaust the resource issue, but because we need to have a vibrant and developing Internet, ensure that we move to IPv6. That's the optimistic outcome.

JONNY MARTIN:

So, you're not just making disapproving noises here but contemplating regulation? Sorry, yes, or no?

FRANK MARCH:

We have lost our fears of regulating in telecommunications.

GEOFF HUSTON:

So, David, you've just heard the Government say that they wish to regulate in order to make sure that the Internet is able to innovate. So far, the last ten years of innovation have rocked your company of around, in Big Pac, about $10 billion a year in revenue and you're being asked to spend even more money to pander to the likes of Google and others, to pander right over to the top of your revenue stream. How do you feel about that?

DAVID WOODGATE:

Obviously on behalf of the shareholders, I have some concern over whether they're getting value for money!

FRANK MARCH:

On behalf of the Government, although we recognize the need for industry investment, we also recognize the need for an innovative and competitive telecommunications market in order to support an Internet economy even more.

JONNY MARTIN:

Akinori, back in your country, Anime, your country has taken a number of IPv6 options with ISPs. What have these told you and how long do you think it will take them to move to IPv6?

AKINORI MAEMURA:

I think Anime is known as the frontier for IPv6 technology, but even in Anime, the ratio of IPv6 is not so high. Only a couple of ISPs, the biggest ISPs just started, the IPv6 Internet service, so it is not so much an easy question to answer when and where we will be ready for the IPv6, but we are now seeing some movement towards IPv4 depletion which is happening in a couple of years.

The move is really, really quick if the industry, the whole industry are aware of the IPv4 depletion, then - my answer to your question is, I don't answer when we can get ready, but we are now having some, you know hint from the industry movement. So, I mean, we will see, we will see the readiness of that for the IPv6 very soon. Have I answered the question?

JONNY MARTIN:

So, David, a quick answer here, so one sentence! If BigPac was to decide tomorrow that it was going to adopt IPv6 across all of your products, how long would it take to achieve?

DAVID WOODGATE:

You think you're going to get an answer to that into one sentence or less! The reality is that it will obviously take several years for any large provider to move towards v6.

GEOFF HUSTON:

To the closest millennium.

DAVID WOODGATE:

Naturally, look, some products will be able to move to it in a faster time frame than others. You know, certainly we hope that we would be making substantial progress in the five-year time frame.

GEOFF HUSTON:

Substantial progress in five years. So the question was, when does it all finish, David?

DAVID WOODGATE:

That's all got to be considered against the other priorities around.

GEOFF HUSTON:

More than five years?

DAVID WOODGATE:

Ask the general manager.

JONNY MARTIN:

There seems to be a general consensus that it will take a while and that's a long while that's not really known when it is going to be.

FRANK MARCH:

If I could make a comment on the thinking about this, that clearly careful planning is needed and close cooperation between private sector in a broader sense as well as telecommunications. So it needs a lot of coordination right across, but one of the things that our neighbour - we have an even bigger neighbour, we call it the West Island, who held a forum about three years ago which I managed to get to. The comment was made that we want to move from Islands of IPv6 and a sea of IPv4 to Islands of IPv4 and a sea of IPv6. And I think that transition is the way to look at it. So, there obviously needs to be an awful lot of planning and that planning is right now.

There shouldn't ever be a time that you need to throw a switch and move from one state to another.

JONNY MARTIN:

So, with other countries, we hope that you get a bit of good IPv6 news. It seems obviously for these more established nations that deploying v6 across all of the existing and old products is going to take years, and they're quite obviously talking a time scale that goes far beyond the life of the remaining v4 addresses. In Asidonia, do you think you could do it differently? Could the local ISPs deploy the v6 networks before v4 has been exhausted?

GAURAB RAJ UPADHAYA:

That's a nice thing to think about. I don't think that we will be able to get to the point where we'll have v6 before the v4 gets exhausted. And quite a few reasons for that. First things first, we have not looked into v4 depletion seriously, because we're busy building companies and expanding networks and so far, we haven't actually felt the things I'm hearing here on the panel today. But it looks like our engineers have been working on v6 as an option and we see issues with v6 on our networks. It is a good idea, there's plenty of address space on v6 compared with v4. So, at some point in time, we'll probably have to look at v6 adoption because 4b addresses of v4 is not enough for the world and not enough for our network and our people, so then v6 is unavoidable. But while there is v4, let the party go on.

And when we're done, then we'll probably have to look at intermediary issues; how we use a scale NAT, optimizing user registering and I think in our case, we're pretty much maximized the use of v4 addresses so we might not get into the trouble that David from Pacland might get into b/c of the wastage of v4 addresses in their network. They say, I never heard that term from Geoff earlier (Class B)!

GEOFF HUSTON:

Oh, you children!

GAURAB RAJ UPADHAYA:

New ISPs so Class B's good idea. We got some good address space from APNIC and APNIC haven't yet denied address requests. We realize that it has to happen, but for now, our priorities are how we expand our network and get people connected, because they say it is social issue, it is getting people connected and then we can think about other ideals like end-to-end connectivity and known NAT addresses and things like that.

JONNY MARTIN:

So a fair summary is no, you can't do it differently but you're hoping that you can last a bit longer before it bites you.

GAURAB RAJ UPADHAYA:

Yes, more or less and we can go and buy companies with address space!

JONNY MARTIN:

Xing Li, do you have any idea how long it will take Olympia to move to IPv6?

XING LI:

Actually, we have some findings based on experience of host structure and operation in IPv4 and IPv6. The first conclusion is actually connectivity is very, very important. So, if IPv6 and IPv4 can talk to each other, then there will be a long time of coexistence, no hurry to push for transition. And the second conclusion is that dual stack is more expensive than running a single stack. So, our current strategy will be for the network, just use IPv4. OK, when you want upgrade, then make it dual-stack ready - then turn on IPv4 not IPv6. However, our new services and new networks, for example, 3G and all those kinds of things. Turn on IPv6, only do not rely on IPv4, otherwise it is more expensive in the use of IPv4, not IPv6.

The new generation of NATs between IPv4 and IPv6 is very important; as long as we can have a good translator, then the old network will use IPv4 and the new network will use IPv6 and help each other and we strongly believe this is the right way to go. So probably, after 10 or 20 years, some portion of the Internet will still run IPv4. However, the IPv6 and the IPv4 can talk to each other.

JONNY MARTIN:

So, Geoff, you've rapidly risen up the ranks of the OECD. We've been asking about the latest prediction on the world of IPv6. So in a short answer how long do you think it will take before v6 is widely deployed?

GEOFF HUSTON:

Well, you know, after hearing what I've heard here, I would have to say that if the Y2K date had have been negotiable, this industry would have figured out that it had another 10 years to go! I'm very, very depressed. So far, prevarication and invocations to magic that this thing would somehow work. I have heard economic predictions in the last few years, in fact even last year that it would take as long as 70 years, and before today, I thought it was kind of outlandish, but looking around the panel this morning of this very important meeting of industry and Government, I would have to say that you folks have successfully managed to carve a bright and shining career for yourselves in doing this transition, and you need not look for another job thereafter, because this one will work just fine.

I suspect that it is now going to look like about 20 years or more, and that truly means that someone has got to be horrendously inventive about IPv4, because to make IPv4 work for another 20 years and there is no magic unfortunately, you're going to make the addresses sing and dance like you've never seen them sing and dance before.

JONNY MARTIN:

Just doing a little bit of maths down here.

GEOFF HUSTON:

Always a worry, 2 plus 2!

JONNY MARTIN:

20 years is in fact greater than three years! So, if it takes us 20 years to implement IPv6 and we have three years of IPv4 addresses left, we have a problem. Am I allowed to use the word "crisis" yet? Perhaps not quite yet?

GAURAB RAJ UPADHAYA:

It is nearly a crisis.

DAVID WOODGATE:

This panel is based on the network community and it is fair to say that network operators will need to take a fair degree of leadership about the introduction of v6, that's true. But, by ourselves, we cannot do everything ourselves.

There is a whole end-to-end consideration for vendor quality, hosting content, can you access the same material on a v6 Internet as you can on a v4 Internet? And essentially, we are recreating the Internet again, given that the original Internet developed as an incremental or anarchic entity, expecting that a change to v6 would happen in an extremely organized fashion is perhaps fairly wishful. We can try to look after our own sections as best we may, but obviously it takes the entire industry to be pushing towards making that commitment and heading towards v6 to make it work.

GEOFF HUSTON:

We are a mere victim of our vendors. Is this the excuse you're using?

DAVID WOODGATE:

Not at all, I'm saying that we all have a part to play.

JONNY MARTIN:

So, Xing Li, getting back to your comment about network address translations, what do you think about NATS? Do you think they offer a long-term solution?

XING LI:

For NAT, if it is between two v4s clouds, no, it is a short-term solution. Or, if the NAT between IPv4 and IPv6, I believe it could be a long-term solution.

JONNY MARTIN:

And Gaurab, you mentioned earlier you're hoping to ignore the whole IPv6 problem for a while and make do with what you've got at the moment. You've been using NATS for a while now. Do you think you'll do any better? Do you think you'll have any luck running everything through NATS?

GAURAB RAJ UPADHAYA:

We've been using NATS for a while now, and there might be scaling issues but with current allocations, it might seem feasible. We're going to go on using that. But just during the panel, I got an SMS from the marketing people, you know they're watching the stream and they said, "Oh no, we're v6 ready now", we should say that to all of the operators. Our network is ready to deploy v6, I can provide it. Please talk to the marketing people! Interestingly, I got an SMS from my tech guys saying, "Don't believe the marketing people"! So, I think NATS is the way to go from what I can see, at least for the near future. But long-term definitely, we will have to look at other solutions.

JONNY MARTIN:

So, David, to date, you've largely made the BigPac customers purchase NATS. How do you feel about carrier-grade NATS?

DAVID WOODGATE:

At the risk of repeating myself, we're considering every technology option that helps us to manage across v4 exhaustion and ultimately towards v6.

JONNY MARTIN:

So, have you looked at how many v4 addresses you're going to need to support the scale of NATS that you're looking at?

DAVID WOODGATE:

We're obviously managing v4 resources fairly carefully as we come up to this point.

JONNY MARTIN:

So for a carrier-grade at NATS, how many addresses will you need? 100,000? 1 million? 10 million?

DAVID WOODGATE:

I've forgotten the size of the BigPac customer base at the moment!

GEOFF HUSTON:

It's big - number in the millions!

DAVID WOODGATE:

Obviously it's a question of what we expect in terms of ratios in terms of usage in the NATS.

JONNY MARTIN:

So you could probably control the customers to control the required addresses? OK, would you - so, you're going to need a stack load of addresses for any sort of NAT on that level. Would you be willing to use one of the many /8 address spaces you have for that purpose? Making the assumption that every company has many /8s?

DAVID WOODGATE:

If the company already has many /8s, funnily enough, I'd be trying to maximize the efficiency of usage into the future and work out if that can be implementation of NAT could do it. If that can be achieved through implementation of NAT, then obviously that would be a good way to handle it. I can't remember any /8s in the current resources though.

JONNY MARTIN:

OK, so Akinori, you've been working on v6 for some time now. We heard earlier from Xing Li that dual-stack was obviously more expensive to operate than a single stack. Do you think dual-stack is the way forward?

AKINORI MAEMURA:

Yeah, I think the dual-stack is the way for us to overcome the depletion, and in the case that you know, you mentioned about the carrier-grade NAT and dual stack, it will evolve this way. V6, there's a way to go for the v4 depletion, then yes, we already have some program which is overcoming the dual-stack situation. With the IPv6, if the one host name doesn't have a record, then they need to wait to go over to IPv4, then it takes a long time for example. And we have some other programs already, but I heard that Anime people are trying to solve that problem, I believe that it will be solved very soon.

GAURAB RAJ UPADHAYA:

To add to this, I'm trying to answer the question you asked David. A few weeks ago I was in Anime and talking with a few fellows there, and with the current state of technology, a single IP address can be used for a maximum 200 NAT sessions or thereabouts, so you can use that as a bench mark that at the present state, I probably need 10 million addresses for my network. If we are able to scale of NATS of say 2,000 users which is not saying anything, probably need a million addresses for a network, for my network. So the whole talk about carrier-grade NAT, while it seems interesting is not feasible because I can't get a million addresses anyway, and my users won't be so happy as when I was in Anime, there was some quite interesting questions about missing patches of the Google map on the screen, which I hope Akinori-San can solve some time during the day.

What happens when you use NAT! And definitely, given our new consumers, young people, I don't want to go out of business, so I don't want to have unfinished sessions there. So yes, if somebody can give me a million addresses today, I think I can run on NAT for maybe another 10 years but not beyond that.

AKINORI MAEMURA:

The answer is no, we cannot solve that with the Google Map. There is a reason we need to move forward to the IPv6.

JONNY MARTIN:

So, I guess we can raise the anchor! And say bravely in the future, now that we have quite a bit of juggling to do, I would say in the next few years but I would say longer than that, I guess we'll keep v4 working as best we can using NATS to paper over the cracks we have here with the lack of addresses and we'll look to IPv6 as our ultimate destination. Ultimate, yes. A little bit depressed actually, because we seem to have a little bit of an issue, but on the upside, I think after the break, we've got some - a couple of potential solutions to the dilemma. Not necessarily solutions we'll all like, but are potential candidates all the time. We have a couple of minutes especially if we have any particular questions from the audience at all.

I think the panellists have talked enough for now. I'm not sure which way to look. Let's start with Randy.

RANDY BUSH:

Randy Bush IIJ. Frank, you deployed NATS. I assure you, you don't deploy NATS any more than I do. There's a little question with IPv6, it is incompatible on the wire with IPv4. We're not going to have a flag day. The last flag day we had, I didn't have grey hair. The NATS are what's going to happen? There's a real tension about doing NATS without - with breaking as little as possible - what's being referred to as carrier-grade NATS is NAT in the middle of the network. That smart core stupid edges, that's telco thinking. That breaks the Internet, breaks the users and breaks and stifles innovation, which should be of great concern to you and of less concern to the big carriers because stifling innovation is exactly what they're going to be about in the next years as they try to control things and keep things as much as a stack to themselves.

I believe that there are alternatives to carrier-grade NATS that really need to be considered and I'm glad we discussed that.

JONNY MARTIN:

I think we're down to about 0 minutes now so you're going to have to be quick.

ANDY LINTON:

I didn't hear anything from anybody on the panel - I heard about about equipment, vendors, Telcos and so on, but I didn't hear anything about where the people who are going to do all of the work from v6 are going to come from. A lot of people talked about the policy stuff here, but I'm not hearing anything about how we're actually going to educate people to work and implement this.

It is pretty hard to find people who understand v4, and I'm not seeing or hearing anything on the panel with views on that.

JONNY MARTIN:

We probably have time... sorry, do you want to respond to that?

GAURAB RAJ UPADHAYA:

In Asidonia, we have extensive training programs for people to learn IPv4 and IPv6 and we're doing Forums where a lot of people have been teaching IPv6, plus v4 actually at the same time to make the next generation comfortable with v6. But, I think our academic courses we are still quite a bit behind, and I mean, when I was at school, I was still learning Class A and Class B. Sorry Geoff, it's none of my business, but the process is lagging behind quite a bit. I think all over the world, and I think that needs to be connected fairly quickly. I mean, professional trainings are there, but as you said, I think in university, we are still teaching students Class A and Class B, I agree with Andy. But there are professional trainings that train people on v6.

JONNY MARTIN:

In response to that, I have part of the answer for you. There's a v6 workshop on here in Christchurch at the end of the meeting, Saturday through to Wednesday. There's six or seven spots available, so if you want to hang around, come and see myself or someone else from InternetNZ to get you on the course. That's followed by a 5-day course up in Auckland the following week. There's one spot left for that so come and see me if you want to get on. But that's obviously not the entire solution.

On that, I think we'll have some coffee and tea and biscuits and that sort of thing. When we come back after the break which is at 11:00, we'll see if we can't sort this mess out. Enjoy.

*APPLAUSE*.

(End of session)

IPv4 in 2015.

Tuesday, August 26, 2008.

11:00-12:00.

JONNY MARTIN:

If everyone can make their way in now, we will get started. So, we will start the second session of this hypothetical. Before we start, we have one hour for this second part of the session. Lunch is early today at, what time? At 12 o'clock.

Because we have such a limited amount of time, I'm going to ask the panellists to keep it short and snappy. So that you get a chance to get through the content, and more importantly, get a chance for questions from the audience at the end, which I suspect is going to be the very interesting bits. We will run through this half an hour. If you have questions, store them up for the end and stand up and propose them. If you want to ask a question anonymously, just hunt down one of the APNIC staff, George or Louise down the back, and they can ask. If you are not confident to ask or you don't want that question to be associated with yourself, we can do it anonymously.

We are at January 2010, and APNIC has considered way to distribute the IPv4 addresses. Have had a look at allowing IPv6 transfers, legitimizing an IPv6 markets. Gaurab submitted a proposal for those in Asidonia for a large chunk to be reserved for the country and the region. But there was division on the effects of the proposals and couldn't agree on any of them. The policies in effect in 2008 are still in effect in 2010. It is much the same situation in other regions also. So, IP address allocation rates have been increasing steadily over the past few years.

In 2008, we thought the free pool would be exhausted in around about 2011 but we now find ourselves in a position where it has already run out. The RIRs had address space left but APNIC has 8 months left of v4 addresses left. Geoff from the OECD is predicting the RIR pools will out just short of 2010, at 11 minutes past 3 and 22 seconds. Actually, as it turns out, Geoff got it wrong, the last IPv4 allocation from the remaining unallocated pool of RIR resources was made on Tuesday, 2010, 28th October, Saint Jude Day, being the Catholic patron saint of lost causes.

Does the end of the RIR free pool on Saint Jude's Day mean there is a higher power saying there is no IPv4, get the IPv6, maybe not? Gaurab's power in Asidonia is going and he is putting pressure on Tulika. All around the developing world the situation is the same. In some parts of the world, the ISPs take the situation in their own hands. It leads to the start of a black market. APNIC has run out of IPv4 addresses but we all still need the addresses. Gaurab, where are you going to get them from?

GAURAB RAJ UPADHAYA:

I'm looking for smaller companies to buy up IPv4 addresses. That has been a proven working formula.

JONNY MARTIN:

When you look in Pacland's vaults, there is a pool. You have Gaurab, who is desperate, he would be happy with a /24. He has more customers but he is willing to use NATs to stretch it as far as we can. Gaurab has $1,000 in his pocket and wants to take it for a class C. Would you sell it? $1,000?

DAVID WOODGATE:

In US dollars?

JONNY MARTIN:

The sale is in Canadian.

DAVID WOODGATE:

It makes it more interesting. The obvious questions that come up in such circumstances are what have we worked out? We may have found the /16 but we have only worked out that.

JONNY MARTIN:

$1,000, $10,000? $64,000? Name a price - $12 million?

DAVID WOODGATE:

I hear the chief economist of the OECD is running a black market.

JONNY MARTIN:

You would be getting half a million for the one /24.

GAURAB RAJ UPADHAYA:

$500,000.

JONNY MARTIN:

$1 million - if you are offered a million?

DAVID WOODGATE:

The obvious answer to all of this is the, we are trying to enable - connection in the Internet. In one way or another, we are trying to pursue sensible business objectives that are legal and ethical.

JONNY MARTIN:

Social good is going to win out?

DAVID WOODGATE:

No, it is a combination of working out business interactions in all of those constraints, legal, ethical, business drivers and what our own internal requirements are.

GAURAB RAJ UPADHAYA:

And also shareholders, that is pretty good cash.

DAVID WOODGATE:

It is part of the business consideration.

GAURAB RAJ UPADHAYA:

Sell it.

DAVID WOODGATE:

How much do you think it would be worth?

JONNY MARTIN:

Well, you have just been offered $1 million. Anyone else want to sell?

DAVID WOODGATE:

Hands up who would sell for $1 million? Anybody who wouldn't sell out there for $1 million?

GEOFF HUSTON:

You.

DAVID WOODGATE:

We are selling.

JONNY MARTIN:

You can debate the number, but there is a point.

TONY HAIN:

Tony Hain from CISCO, why have a single transaction when I can get the money every month?

JONNY MARTIN:

So, I think we can assume that most businesses are going to be up...

GAURAB RAJ UPADHAYA:

If I had it in my pocket, I will give it to you, a /24.

GEOFF HUSTON:

Come on, we know what it is worth! Stop trying to buy it cheap.

JONNY MARTIN:

There will be a price for a willing buyer and an equally willing seller. Obviously, we are going to run into a situation where, in this case, Gaurab has bought the /24 he dearly needs, buy it off David, he is willing to announce and Gaurab gives a letter that says it is all OK for advertising the stream. So it brings us to the next question, Frank, today's daily NANOG news headline is Pacland has been caught dealing addresses on the black market. Is it time for an inquiry? Will you refocus his carrier license?

FRANK MARCH:

We have had a change of government and looked at it differently. We are pleased a market philosophy is prevailing. But more seriously, what we are impressed about is the fact we heard all the doom and gloom for years, there is no business case. But now there is beginning to be a clear business case emerging. We are confident the v6 world will persuade everyone to move. Particularly as the government has decreed all of the IPv6 will create pressure. The market for IPv4 may be temporary.

JONNY MARTIN:

Gaurab, you have been caught by the routing police; your addresses have been declared black.

GAURAB RAJ UPADHAYA:

We have been ignoring them since 1980.

JONNY MARTIN:

Akinori, given there has been an uproar about what has happened here, your members are starting to ask you whether routing addresses should be removed from your registry, are you going to remove them? Your members have been asking you to remove that particular /24 from the routing registries. Are you going to remove them? Leave them there? What sort of discussions are you likely to have with your members?

AKINORI MAEMURA:

Sorry, I didn't catch the question.

JONNY MARTIN:

The black addresses from Gaurab, will you remove them from your route registry?

AKINORI MAEMURA:

The black market, actually, we cannot detect the addresses from the black market, it is quite a difficult question to answer.

JONNY MARTIN:

In the case of this being the first high profile, high-monetary value transaction that people became aware of, would it make any difference? There a lot of noise about it, the route registry can't detect it, but the Internet operators have.

AKINORI MAEMURA:

Um, what can I say?

JONNY MARTIN:

Just too hard to think about?

AKINORI MAEMURA:

Yes, if the market works for the trading then the registry should keep the book of the address holding and in that case, there is the real concerns for bookkeeping but leave the market, just work for had recognition of the IP addresses.

JONNY MARTIN:

Obviously, this is not the Wild West. Black markets and global telecommunications networks are not compatible. It quite likely we won't get to the point where Akinori is going to have to contemplate what to do with the entries on the route registry and Frank won't have to look at revoking licenses or other regulatory intervention. Gaurab and David, the routing police will pardon you and you can carry on. A number of developing countries have put pressure on the UN to hold a crisis meeting; they are demanding a fair distribution of addresses. The UN says to start again with a new registry. They will redistribute from scratch a company called UANA. But getting the whole planet to renumber from scratch is impossible. Even most IPv4 staffed country can see that.

On the first week of January 2012 they hold a meeting in Christchurch and devise a solution. In Christchurch, in front of the world's media and a straggling bunch of NZNOG bloggers, advisors, including Tulika and Frank, hammer out policies. Under the protocol, each country has a per capita IPv4 target. Some countries are well endowed and others are struggling for anything. So countries with more than their allocation of IPv4 addresses have six months to give up the excess addresses to be redistributed to the countries that are starved of IPv4. Xing Li, you stand to receive an additional 20 /8s from the protocol, are you happy with the outcome - from Olympia's perspective?

GEOFF HUSTON:

Overjoyed.

XING LI:

Yes, OK.

JONNY MARTIN:

OK. Frank, the Christchurch protocol set Pacland a target of shedding 25% of the IPv4 address space. It will be difficult to do in the timeframe but other countries like Bushtopia have to shed 75% so we haven't got it so bad. It is now your job to phone up David and ask for 30% of addresses. What exactly are you going to say to him on the phone?

FRANK MARCH:

First of all, I would congratulate him on having discovered a market model after all, and a business plan. Since he would have been consulted about the 25% required to be shed, he would be well aware of the issues involved and I would expect full compliance, especially as he already discovered IPv6 was the better way to go.

JONNY MARTIN:

Tulika, as you know, Asidonia was going to get additional /8 in the redistribution. The /8 is not enough to meet your needs but it is going to help. How would you redistribute these addresses within your country?

TULIKA PANDEY:

My first question as to the policy is how you allocated more to Olympia and less to Asidonia, that is one. I would like to know the policy that has led to that. Now, go ahead and you must have watched me hobnobbing with industry. I was trying to see how much percentage the government can get from them, if we decide a policy as per their liking. I didn't get a very good response so in our country, the scarce resources like spectrum, we have a policy of going for an auction. And that is what I am going to do with these scarce resources that have been allocated to my country.

JONNY MARTIN:

So the auction is going to be your solution?

TULIKA PANDEY:

Yes, because I have more than 100 ISPs running in this country and more numbered users than the resources that I have, so that's my only option.

JONNY MARTIN:

So in the case of having one large telco and 2,000 small ISPs, that is likely to lead to the situation where you have one large ISP and maybe a couple of small ones. You see that as an inevitable outcome of that?

TULIKA PANDEY:

At present, yes. That is how I see it.

JONNY MARTIN:

Gaurab, you are in the lucky position of running the largest ISP in Asidonia. You want to connect a new 100 million more customers. How many addresses are you going to get from that /8?

GAURAB RAJ UPADHAYA:

We will take all of it.

JONNY MARTIN:

Is he going to get all of it, Tulika?

TULIKA PANDEY:

I will have to do a rethink. The auction is off.

JONNY MARTIN:

OK.

GAURAB RAJ UPADHAYA:

To put it in perspective, there is a spectrum allocation policy which actually is not just auctioning off the entire bit but dividing bits into that (I think that's what Tulika meant earlier). Some will probably going to auction and some is probably there for other things. I think that is what the policy says - for a wireless spectrum, at least.

JONNY MARTIN:

Goeff, is our Christchurch protocol going to work? Do you think it a fair solution to the IPv4 problem?

GEOFF HUSTON:

No, in fact, it is probably the worst possible solution. Once you get into the geo-political systems of allocation and rationing you actually hide and suppress the true situation of scarcity so players who receive the undervalued goods actually don't make efficient use of it, others who are trying to make sue of it have to survive on a whole lot less here. What the natural reaction is, is that the beneficiaries of this kind of redistribution system start hoarding and that hoarding further reduces the supply to the market, exacerbating the situation rather than relieving it. We found time and time again over history, any kind of common good that is allocated, irrespective of the economic value of its exploitation is distributed amazingly inefficiently and poorly and it doesn't actually do anyone any good. We turned into auctioning for spectrums as being the way of making the most value out of the radio spectrum which is a scarce displacing good and that the similarities of that to addressing in a situation of scarcity are certainly there.

So in this particular case, we would have to observe that while the Christchurch protocol certainly made the participants feel very good about themselves in solving the problem, in reality, in the cold light of day, they made it worse.

JONNY MARTIN:

Back at the NRO then, we've got a number of ISPs who are complaining at RIR meetings that the Christchurch protocol will kill business. The information superhighway in early 2011 is looking more and more like a train wreck. So in an attempt to kind of solve, partially, the problem, the RIRs fast- track the implementation of policies which are going to recognize address transfers in whatever shape or form they may happen.

To assist in the orderly transition here, to recognize v4 markets, ICANN and the GAC hammer out an address trade agreement under the auspices of the World Trade Organization. The replacement agreement, which replaces the old DOHA agreement, is known as the Dunedin Agreement and at this point we have IPv4 markets. David, you are sitting here trying to figure out how to collect the addresses that Frank asked for a little bit earlier. Now we have a legitimate IPv4 address market, and it's possible you might be able to buy addresses for less that it cost it would cost to re-engineer the network. Are you tempted to buy the addresses Frank wants you to give him on the market and leave the network as it is?

DAVID WOODGATE:

Obviously, if it is what the business studies show, then I would be mad not to consider it.

JONNY MARTIN:

So you're likely to be a buyer?

DAVID WOODGATE:

Likely.

JONNY MARTIN:

Gaurab, you decided IPv6 is the answer and you run NAT to perform the protocol translation. Consequently, you only need a /14 or so to do this. But you manage to obtain the entire /8s for your country under the Christchurch protocol, so you have a whole heap of extra space now. Are you a seller?

GAURAB RAJ UPADHAYA:

At a price.

JONNY MARTIN:

Which takes us back to David. What is your buy price? It is going to cost a lot of money to engineer the network according to your technical people.

DAVID WOODGATE:

Well, I'm hoping it is going to be somewhat less than the $1 million per /24 than Gaurab bought from me, otherwise it would be $16 billion. I think I need a… Bill?

BILL MANNING:

So I have a question.

GAURAB RAJ UPADHAYA:

Bill, who are you?

BILL MANNING:

Al Gore. So, Bill Manning, in this hypothetical scenario, I don't think David could buy the addresses because it would violate the Christchurch Protocol and equitable distribution.

JONNY MARTIN:

David is doing what?

DAVID WOODGATE:

The legitimate market I was buying off.

BILL MANNING:

There is the wrinkle, there is the other thing touched on earlier by Tony, which is the permanence of the IP address assignment. You are talking about markets as a way to shift them around over some longer temporal time, but we already have a technical solution, it is called the DHCP server, and if there is a single ISP in a country, just run the DHCP server with 15 minutes of expiration and there will be plenty of addresses or just do the DHCP server for the Internet. That's one way to do it to make sure there is equitable distribution: people get v4 addresses when they need them and when they are done they stop using them if they want permanent v6 as a choice.

JONNY MARTIN:

We haven't got too long in the session. Save the audience questions for a few minutes time and plenty of time then to debate and ask questions. Got to the stage now where David has thought pretty seriously about buying addresses to satisfy the government's needs to take these addresses off him. Gaurab, you have a potential buyer, is there is a deal here.

GAURAB RAJ UPADHAYA:

As I said, at a price.

JONNY MARTIN:

Tulika, Asidonia went to the UN to fight for the extra IPv4 addresses but now they are being sold off-shore again by your largest ISP?. On the other hand, Gaurab's enterprise is now a very rich enterprise, now, how do you feel about this?

TULIKA PANDEY:

It leads me to think about getting policy where I create an industry to go for IPv6 than continuing with IPv4. It is how I see it today and very soon our government will bring out a policy on this.

GAURAB RAJ UPADHAYA:

Of course, they got the money from the auction.

JONNY MARTIN:

It is a replay of the early '90s when the developing world was faced with the prospect of upgrading the copper telephone network or installing GSM network. But the choice at the time was GSM, and now the right choice would be to take the money and get on with the v6 deployment. It places your country, pretty well in an Internet sense. So, Frank, out of all this, your IT minister got what he wanted, BigPac spent some money but they got the same number of IP addresses that they started with. Do you think the Christchurch protocol was worth the effort?

FRANK MARCH:

It is the classic problem of government regulation that can lead to adverse outcomes. That is always the case, which is why governments need to be very careful about entering into regulatory regime in any circumstances. But the question is whether it would be sufficient concern for the government to intervene at this point, I suspect no - I mean, one could say that somehow or other, by involving itself at this level, the government, inter-governmental discussions probably broke the impasse that was there, even though the outcome wasn't quite as foreseen. So we have got plenty of examples where regulation has resulted in that outcomes, unexpectedly, but nevertheless, relatively beneficial outcomes. So maybe we should think of that.

JONNY MARTIN:

So that was a no? It wasn't worth the effort?

FRANK MARCH:

It was.

GAURAB RAJ UPADHAYA:

Jonny, you conveniently left out with that when the UN meet, they don't focus on one thing, you left out the redistribution of IPv6 and the crisis protocol, my experience with the UN process, whatever, WTO, I don't think they leave it out of the whole thing.

JONNY MARTIN:

It was convenient thing to leave out. Xing Li, some of Olympia's ISPs had surplus addresses allocated to them and they were also selling on the market. Of course, the market price has got too high for other ISPs to bear, and they can't afford it. Does IPv6 start to look attractive at this point?

XING LI:

Yes.

JONNY MARTIN:

David, the market is going and your products are selling but you need another /16. At what price does a /32 in v6 start to look like a better alternative?

DAVID WOODGATE:

The price that it makes sense to be in business.

GEOFF HUSTON:

You are the economist, what makes an efficient market?

GEOFF HUSTON:

What makes it efficient is market information, when prices are open, prices are disclosed; when the identity of buyers and sellers is known, when transactions are visible. What makes an inefficient market is when it fails, when transactions are occluded and buyers and sellers identities are hidden and when the good being transacted, in this case, simple integers are not honestly and openly transacted. Does anybody want to buy a number three today? It is for sale. In essence, what is going on, saying a market in an address as it appears is only part of the problem, historically we found markets need various forms of regulatory control to make sure the worst excesses of behaviour are limited.

JONNY MARTIN:

So an open and transparent market be enough to prevent hording?

GEOFF HUSTON:

It would probably be enough to prevent hoarding in the long run, it is a market that has a very definite future, what happens as price escalates, v6 looks more and more attractive, as ultimately, David and Gaurab have to bear the cost of buying the addresses and putting in all the translation infrastructure and keeping v4 running. At some point the total cost becomes unbearable and v6 looks more affordable. So while the markets might try to push an aging and decrepit v4 past its use-by date, by upping the price because of scarcity, you create the economic signals for v6 that so far apparently have been lacking for the last 10 years. Your continuous repeating of the phrase "it makes business sense" says to me v6 hasn't made business sense to you for the last 10 years because, otherwise, you would have done it. At what point do we kick business sense in here for v6 and market that operate in an open fashion way giving you the necessary business signals you are looking for?

JONNY MARTIN:

It takes us to a good point, and the audience here, it seems we have to make IPv4 run for far longer than it was designed to run, and mechanisms to redistribute these addresses can take the form of black markets, open markets or international protocols. We looked at it briefly and hopefully teased out some scenarios. We want the audience to look at and make statements, what do you think, what might happen? What should happen? How can we make it through the transition and have an Internet that works just fine? We want the broadest participation we can get here. If you are shy about standing up behind the microphone, please talk to George there and give him a question, he can ask it for you. Otherwise, line-up behind the microphone and start for the next 25 minutes or so looking at that.

It is now 31 December, 2014, addresses are being bought and sold all across the world, the RIRs are registering their whois database, the Internet is buzzing along and v6 usage is increasing. Back to the audience here, and starting with George, and some of the anonymous questions.

GEORGE MICHAELSON:

Two questions, one is about the business case. The question essentially, what is the business case for v6 if you are in a public company with a shareholder base? When does it become rational to spend money, mitigating against the long-term risk that is looking at the moment, still like a potential risk rather than using that money for short-term profit? Is there some kind of sweet spot that you can know about, that you can work towards that makes sense to make it flip over in business practice?

DAVID WOODGATE:

I think Geoff has already answered that question to some extent, but I point out even a public company is expected to investigate its options and under the risks and risk planning and take judgment on those risks at times. So it is, it becomes a question of judging what the relative value versus relative risk is. At the moment, it looks like you know, even if there are contingencies, things like NATing etc, in the long term we need to go to IPv6. A lot of concentration on doing that because that is, it does appear to be the ultimate way the technology is going. Now, at the same time, you have just got to keep your finger on the pulse of the market and see what happens.

JONNY MARTIN:

Anyone else have anything to quickly add to that?

GEORGE MICHAELSON:

The second one; New Zealand is essentially a regulated telco economy, the industry is regulated by the government but they have been encouraging competition and there is a very competetive environment here but in the context of v6, the increasing competition has made the likely successful transition harder and these are the kinds of questions it is substantially easier to tackle when you have a large monopoly-based industry because you only have one player you have to re-energize on addressing rather than deal with it on a competitive environment. Is it not a sustainable argument? Maybe?

FRANK MARCH:

The experience in Pacland has been the large monopoly telco is the most difficult thing to move, especially for the whole of society. So in theory, yes - in practice, no.

RANDY BUSH:

You can look at the two parts of Pacland, you do have in New Zealand a competitive economy and in Australia with a single monopoly and look at the actual difference. I have had to sit here and listen to a large monopolistic telco use the words "ethical" and "legal"; I have heard Geoff say he knows the price of a /24 and I'm having a hard time finding reality here.

DAVID WOODGATE:

This is not reality, it is a hypothetical.

JONNY MARTIN:

We can move into reality now.

RANDY BUSH:

Hopefully it is trying to relate to something we are trying to deal with. In the evening, we will go to play for entertainment, thank you. Nobody has mentioned the fact that the introduction of certification for address space will affect whether you are going to continue to have a black market or not, and how that makes it difficult; whether the certification of address space has a problem with routing, the registries on my routing, or routability, whether in a market for this stuff, that is actually going to cause a - cause major players like our friendly monopolist up there to move, or whether it is going to drive the Internet to v4 NATs period and IPv6 is dead. That is my biggest worry, not when David will transition. He will transition when his Board tells him he has the money.

People that plan for the next quarter, it will be late and they will pay the penalty then and for people that plan 20 years ahead, like Akinori-san there'll be IPv6 sooner. My problem is not when they'll transition to IPv6 but if they will transition at all and become an IPv4 NATed world instead.

JONNY MARTIN:

Your first point of the emergence of certification of address space is going to impact the emergence of the black market could be turned around. Would the black market impact on the possibility of introducing certification and that sort of thing. It is a two-way street. Yes?

NARESH AJWANI:

Good morning, everyone. Shall we leave the policy formulation with the stakeholder who has nothing to do with the ultimate person who is getting affected by the black market, or shall we take the policy and action situations at ICANN level, where we shall be in a position to go for something with the perspective of all the stakeholders in the system. My question to Frank and Tulika, would it be right on the part of ICANN to leave this policy decision on you people?

JONNY MARTIN:

Thoughts?

FRANK MARCH:

Yes, that is basically, are we talking hypothetically? The Pacland scenario is a government which is as hands-off as it can be until it is forced into regulatory action. Under those circumstances, decisions are made very pragmatically rather than on a theoretical basis. At the moment, it is working; the government is really not interested in changing it. When it ceases working, the government may or may not be in a position to act in a reasonable and useful and timely manner to change that. And so I think yes, it is reasonable.

NARESH AJWANI:

What should be worried about auctioning and ultimately affecting the end-user, Tulika?

FRANK MARCH:

Going back to what I was saying about the business case and Geoff was saying, it becomes a balance point, a black market operating in some way, or even a white market, will if it is a genuine scarcity and there is an option available, as we are seeing with the liquid fuels market and the gradual emergence of possibly electrically fuelled vehicles as an alternative, it is entirely driven by market forces with guidance from regulatory reform but governments that totally regulate against the tide are doomed to failure. It is true of corporate forces as well. If they ignore the reality of what is going on, eventually they are superseded. I think IPv4 versus v6 is that situation. A business case will inevitably emerge. My concern is that the innovation that has characterized the Internet which is being stifled by the intervention of work around like NAT is, it is not part of the equation at the moment. And it is impossible to put it on a balance sheet, either way. It is an opportunity cost I'm convinced it is there, but have no way of measuring. I think that is the real problem. We are stifling the changes and the possibilities that the Internet offers because we are not moving on. That is a personal view, rather than a Pacland view.

NARESH AJWANI:

The policy right now, immediately, whoever wants the allotment of IPv4 has to take mandatorily the same number of IPv6 and that would be a solution. What Bill has been referring to is exactly the same, somewhere the policy has to originate from ICANN than from any other stakeholder. If we really want to integrate IPv4 and v6, then we have to have some kind of policy formulation where the balancing act starts from the day 1 instead of waiting for 2011 or 2012.

FRANK MARCH:

That is opening up the whole question that you are looking at this afternoon, around the role of Internet governance and the role that governments versus private sector play in that. I personally would not want to comment or step on to that ground.

NARESH AJWANI:

Shall I hear from you, sir?

GEOFF HUSTON:

Oh, I was always of the view that ICANN was a product of the deregulated industry and it was a natural expression of the will and intent of the industry players. It seems a remarkable twist that you are invoking ICANN as an instrument of regulatory direction and it is certainly an indication that I never thought I would live to hear, and I thank you for saying it because I never expected to hear it. I am really completely unsure what ICANN will do with such an invitation to act in a way that is so contrary to its basic principles and modes of operation.

NARESH AJWANI:

Sir, I'm more worried about black marketing than about your value systems and ethics and about your freedom. I'm more worried about the end-user, the stakeholders somewhere we have to understand our role in this policy origination and action has to be from your end and nowhere else. Thank you very much.

DEAN PEMBERTON:

We might need something in terms of an IPv4 vet to take the pulse of the horse and let us know when it is actually dead, because it seems the people on the panel would like to keep flogging the horse for a different amount of time. We had similar things like going from analogue cell phones to digital or analogue broadcast to digital TV, there have been times, especially in the Australian market, where someone needs to take that pulse, put a stake in the sand and say, "By this date, something will happen and the old one will be gone or legacied and the new one will be started." Without do - do you think we would still be running around with analogue handsets?

JONNY MARTIN:

Got a date in mind to turn off v4?

DAVID WOODGATE:

I don't have a specific date in mind but, obviously, some of the dates that have been applied in the Australian market have come through strong regulatory drives as much as anything. Having said that, I did make a comment on the APNIC policy, along your lines that it would be good for the industry to have notional targets, it may not be something that is mandatory but if there are no targets set we won't have anything to work towards.

JAMES SPENCELEY:

It appears we are on a very large ship, there is a iceberg and deck chairs. We continue to want to continue to rearrange the deck chairs. From the panel I have heard hypothetically adjectives, thinking, planning, considering, thinking. It reflects very well but the true reality of the situation is we are talking about planning and talking about considering and talking about implementing, yet we aren't focussing on the doing. So in that, I think the panel has captured very concisely the problem in the industry, we aren't talking about the doing. The doing comes typically on the Internet from the smallest parties upwards, rather than the largest parties down. The smallest parties aren't aware there is a problem on the ship called the "Titanic"; we need to educate the industry far wider than we are. We consistently and often hear about the problems we are facing with IPv6 but there are hundreds of ISPs in Australia who don't believe there is a problem with the address base running out. We need to educate the smaller ISPs who will innovate because they have hundreds or thousands or tens of thousands of users to put market pressure on the larger providers to themselves innovate and adopt IPv6.

FRANK MARCH:

I remember attending the RIPE NNC conference in Amsterdam in April in 2005 when the line of speaker after speaker was for a government regulatory audience. Speaker after speaker assured government there was no problem with IPv4 address space. It would be fair to say there was a fair measure of scepticism but nevertheless, it was the message. The point I'm making the industry itself and a number of authorities and government and increasingly ISPs are changing their view over time. And I suspect what you are saying is less and less true now. This discussion is part of that change. So it is a moving target, I think the sense of alarm is growing, I hope, without a sense of despondency because there is plenty of time to put the pressure on. But we are shooting at a moving target and perceptions are changing a great deal and awareness. Now the area I'm worried about is the non-telco, non-Internet business sector, which is not aware of the problem. They are the ones who are going to be quite difficult to wake up.

AKINORI MAEMURA:

I'm quite surprised with James' comment. In our country, we had some survey for awareness of IPv4 depletion and it was incredibly low awareness, even in Anime, the frontier for IPv6. The case for the ISPs was much better, the case for the hosting provider was horrible; the awareness is almost zero. So I am very aware that we need more and more campaigns to increase awareness of the IPv4 depletion itself. Our country have some a couple of research projects for 2007 and 2008 and they find it has some timetable for towards IPv4 depletion which says let's start and complete the planning for the, you know, for the counter-measure for the IPv4 depletion within the year 2008 to implement it in two years. So it is quite difficult to achieve that. It is what we need for the, to confront the IPv4 depletion.

GRAHAM INGRAM:

I'm really confused. I was at the OECD (Next Generation Network) NGN workshop in Budapest a couple of years ago. I was told the NGN is going to fix the problems. I was wondering why we are worried because apparently the NGN is going to be the replacement for the Internet anyway. Where are these two things meeting up I suppose?

JONNY MARTIN:

You guys are answering.

TULIKA PANDEY:

The NGN has been bought out by the telcos. It is their way of saying, "Well, we are still there and we are going to be there to bring in all the solutions that you seek".

GRAHAM INGRAM:

Exactly, my point.

TULIKA PANDEY:

Whether we here, would totally agree with what they have to offer, or whether they're really clear about the issues that we are talking about today or in the past, I will just mention that here and not any other answer now. OK.

ANDY LINTON:

I was just listening to the last speaker talking about technology fixing the problem, there is a lot of parallels between this debate and the debate on climate change. We have seen a lot of people talk in the past that technology is going to fix it. I think our friend, hoping, the telcos are going to fix it is like getting the oil companies to fix the problem of global warming. They are not going to do it, it is not in their interests to do it and that sort of thing. One of the things we should be thinking about, some of the lessons that have been learned in communicating global climate change are some of the lessons we need to learn here and communicate the problem of v6 exhaustion. It is a similar problem, a finite resource, things are changing, there is going to have to be massive change and it seems pretty clear a lot of scientists are saying, it is inevitable, it doesn't matter which thing you look at, climate change or v4 depletion, so perhaps there are lessons to learn.

GAURAB RAJ UPADHAYA:

What Andy is saying is we need Al Gore for the Internet.

JONNY MARTIN:

Any more questions from the floor at all? Otherwise, probably have a few seconds to lunch.

GEOFF HUSTON:

No, it was a response to Andy. When you talk about inevitability, certainly, the cessation of the current method of distributing IPv4 addresses is going to run out, but to simultaneously assume v6 deployment and its replacement to v4 is inevitable is a major assumption. It is certainly true today a few hundred million handsets are not really on the Internet; IMS actually is an application level gateway system that admits only certain applications and denies all others, and services upwards of a billion end-points. There is an existence of proof out there of something that looks a little bit like the Internet but vests control and economic capability in different hands. Certainly, when one looks at the future, there are a number of possibilities and many options and in assuming the inevitability of v6, I think, is relatively naive and certainly, wonderfully optimistic, but if it is what you want you have really got to work for it. In some places, there are other futures they are working for that actually vest and replace a huge amount of control on the user experience back into the hands of the operator of the network. That may not be a future you particularly like.

ANDY LINTON:

I don't think I'm saying the depletion of v4 means v6 is inevitable, I don't believe that. In the same way the removal of oil as a primary fuel for transport means electric cars are inevitable, it may mean a complete change in the way we do things. Another option, which is perhaps heresy to do here, is we should not give up on the Internet and go home and do other things because we have exhausted the possibilities of it and we could go down that path. I don't think it is a realistic one, it is like the using the horse and cart, I don't think it means v6 is inevitable, I believe there will we change, I agree with you, Geoff.

JONNY MARTIN:

Pretty much out of time now, bring us to a close. We seem to have come to terms with transition. The industry is a little bit comfortable, or, more comfortable with v6 adoption. There is even some confidence the last v4 host will actually be switched off, perhaps as soon as 2015, if not, some date in the near future and we will have a v6 Internet. So success, assume it will happen. Of course, we get to the stage we deem a successful solution, but not long before something strange in Asidonia happens. A combination of a rare form of Silicon, a small AAA battery, and a hen. A researcher manages to generate a strange effect. He managed to receive a data packet even before he turned on his laptop. The new intuitive quantum technology, is better, faster, cheaper and incredibly cooler than the old Internet net. Another boom starts off, start-ups appeared around the world and we head to quantum networking but we need a registry for the quantum addresses and someone asked what the membership model will be. It is where we will leave it today, in a happy place.

Thank you very much.

SAM DICKINSON:

Just a quick reminder, lunch is from 12 to 1:30, the ICANN session starts back here at 1:30.

GAURAB RAJ UPADHAYA:

Today is 31 December 2014, so you are all welcome to party at the Victoria Café - let's party.