weightwelloffΚινητά – Ασύρματες Τεχνολογίες

12 Δεκ 2013 (πριν από 4 χρόνια και 7 μήνες)

95 εμφανίσεις





16:30 CET




This text is being provided through CART, Communication Access
Realtime Translation. It is intended to facilitate
communication accessibility and may not be a totally verbatim
record of the proceedings.


Hisham Ibrahim
: Ladies and gentlemen.
If you will be
seated, we will start momentarily. Thank you for coming to this
late session on moving to IPv6 organized by the coordinating
body for the five Internet registries that managed and
distribute IPv4 and 6 and system numbers globally.

I would
like to invite our first speaker to set the mood and
give us an overview on how things are on the Internet. Jeff,
please introduce yourself.


JEFF HOUSTON: I'm Jeff Houston. If you aren't tuned to
Channel 3, you should be; otherwise, you won't be hea
ring me.
So everybody, Channel 3.

I would like to set the scene to place this effort about
IPv6 into some context. I often think the period we're living
through in terms of the change being introduced by the massive
change in the escalation of the
computing industry together with
a complete rebuilding of telecoms is so fundamental, it will be
our kids or grandkids who will really put it into perspective.

Because as you live through it, some things it that turn out
to be incredibly significant or on
ly significant years after the
event itself. To set the scene, I would like to go back a bit.
I would like to actually fly back 20 years. As you might
recall, 20 years ago, it was 1992.

If you had gone to the world telecommunications whatever it
was in

1988, you would feel happy you did some regulatory
measurement about the telephone companies, but the computing
word was amazing because 20 years ago everything you are
currently holding wasn't even thought of.

Laptops weren't there. Back in 1992, porta
ble computing was
an Apple Mac which didn't fit in your pocket, or even a
briefcase. Real computers were like this wonderful computer
that was meant to be water
cooled that consumed 20 kilowatts of
electricity, hummed like nothing on earth.

That was 20 y
ears ago. So when you think of the time, the
industry for the last ten years was trying to build bigger,
better, more stylish mainframes. Actually, it was a crap
machine. They failed; it really is ugly.

But that was the idea behind it, building bigger
pieces of
ironwork, but what killed it was the first stage in a fantastic
journey. The personal computer wasn't an instrument of
scientists. It wasn't something bought by universities and
corporates. It wasn't something proudly displayed in a
room. It wasn't something where you had clusters of
computer operators who knelt down and brushed its shiny buttons.

This was consumer electronics. All the sudden everything we
thought about computing was thrown on its head because now we
were in a diff
erent world, a world of mass numbers where if you
didn't ship millions of computers, you were out of business.

So this is one made by HP. They are a survivor. This is
2002, ten years ago. You should never get engineers to design
computers. Isn't it ug
ly? Your desk would fall apart if it was
on it; it looks awful. But this was ten years ago, big thing
that dominated most of your desktop. Ugly.

Look at the next device. This is now, from March of this
year. Computer. Fashion item. It is elegant,

stylish, the
corners are rounded slightly. The white is just a bit
white. Everything about it screams fashion, not computing.

What is the size of it? It has the same power as the huge
dinosaur of 20 years ago. It has a camera. It has everything
you would expect in a computer, but it only has two button, not
300. That is 20 years of evolution.

So if you think about it, and there are a few of you with
your laptop out, what you are tapping on is also a dinosaur
because the world is even moving on.

We really, really are in a
pc world. The next ten years, computing with shrink and
move onward. I want to sort of think about where the protocols
sit inside of all this.

Let's look at the typical geek hide away, the way the
Internet used to be a
destination. I used to go to work to log
on. I used to have a place in any house to log on, a dedicated
place where I used the network with me large screens, dedicated
lighting, reliable power, the bandwidth came through on wires,
my chair was comfy.

is isn't my office, but it could have been because that is
the way we thought about computing. Now the change is just
dramatic, massive. Because on the train on your way to work,
would you have seen this, folks just playing with their i phones
and voices
, possibly doing e
mail, listening to a song or
reading a book.

All of the sudden, the Internet is literally everywhere.
And using it is no longer a destination; it is incidental. It
is just part of the fabric of our lives.

And the more this continues,

the more it is going to
disappear. I suppose these things will be using ip in the next
year or so, but that's not all. All of this is susceptible
because the Internet is down to being hand
sized and enormous
computing now fits in my hand. It is a radio,

something you do
when you have nothing else to do or whatever.

What is a workplace anymore? Are those folks going to work
or are they there? You can't tell, and I can't tell. They are
probably doing their e
mail. All of the sudden we have changed
enormous amount of the world. And then when we start
counting the number, these are not insignificant numbers, 2.5
billion Internet users is a dramatic change.

But the next one is kind of easy, because 5 billion folks
have mobile phones, and the differen
ce between mobile phone and
Internet mobile device is the computing power you pack in.

The next one, there are currently 750 folks using these
mobile handsets. Now how old is yours? I can guarantee that
none of them are older than six years ago because
that is when
Apple first introduced this.

All of those 750 million have happened in six years. Took
us forever to get the first 750 million people on the Internet;
it has taken not very long to get the rest on mobile.

What is tomorrow's usage and user l
ike? All of these
productions are insane. Apple at one point was the largest
valued stock company in America because most of us can see the
inevitability of what is going on.

These things, fashion item though they r cost less than a
hundred dollars. Wi
th the strength of buying Apple has,
probably $50. They are very cheap and incredibly ubiquitous.
So if you look at these numbers in perspective and you think
where is it going, things get frightening.

Five years out from now, let's sort of think about

what that
means. Already last year we shipped 270 million of these mobile
units. The cost is really low, less than $50, and Apple doesn't
have a monopoly

Google and Android is right behind them. The competitive
market is keeping costs down. But beyon
d that, you don't see
proprietary technology; you actually see open technology.

This is Unix. So are Androids, an operating system
developed in the early 1970s. The libraries are all open
libraries. That is why these phones get hacked all the time
use the code is basically open code.

Almost every part of the kernel of the system is open
software. What about the apps? There is no great specialized
environment for these particular devices. So all the sudden
there is this shift away from computing
being something special,
to being something incidental. And the numbers are still

Look at Apple and the iphone. Two years ago, they shipped
8.4 million iphones. A year later in one quarter, 20 million,
even those massive ipads were shippin
g like crazy. Then you
updated a bit more, and the fourth quarter was bigger, doing $46
billion in sale.

So how do we fit all that together, all of those massive
numbers of 37 million iphones in a quarter and 15 million ipads
and $13 billion in profit.
There is one thing that is pretty
clear, it won't fit in before. If you look at the red curve,
that is in millions of units per year, and that is a very
conservative view of the demand for public addresses. It
doesn't even track what's behind it.

tly 700 million new devices connected per year,
translating into the demand for 200 million addresses for V4.
We don't have that. In 2011 we were slightly below demand.
This year the demand will be for 300 million addresses, and we
only have 100.


won't stop, and Google won't stop. How do we cope
with this massive overhang of demand? Well, we just ran out.
Most of those curves plot the amount of addresses we have, just
all go to the bottom. There are no more V4 addresses.

Theoretically in fi
ve years out, we would have dealt with
this because something will have to happen, but we're not the
telephone industry; we're not a "sure bet" anymore. We are now
considered to be very high
risk as an industry because we have
no idea what we're doing.

nd a lot of the activities we're now doing have a very high
risk of failure. There is now an after
market in V4 because
there are no more addresses in parts of the globe and there will
always be buyers and sellers.

If the price goes through the roof we b
ust the network
because that is acute instability, lying cheating stealing
addresses, say goodbye to the network. What if the price is
highly volatile? Why would you invest in a industry with so
much uncertainty, detracting from meaningful investment?

So we're now not the old, solid telephone companies of the
past. We're no longer a rock solid conservative industry.
Oddly enough, we're high risk. This is not good, and I suspect
down below this is a very fundamental choice the industry has to
make not

in 10 years or 25 years I heard this morning; that's

It will happen in five years. The choice is stark with a
lot of implications. We either about serious about V6 and do
it, or try to keep doing what we do for the past couple years,
g and pushing the network into unnatural places.
Everything gets squashed into carrier
grade nets, but we have no
idea what path the Internet will take.

That is because these days, this is a deregulated industry
with no one in control. The determining f
actor is actually
market forces, and the market forces that are being played out
are quite large. We have seen a massive transformation of the
shift of money. Google and Apple and Facebook and Amazon never
existed before, and they are all quite rich com
panies in their
own way.

Content on the Internet is now an extremely valuable
activity, and carriage is not; it is a mere commodity industry
where all they're doing is pushing the bits. The value is in
the applications and services that sit on these comp
uters, not
in the industry of bit

If we make the network a constriction point, we will change
the balance and the carriage industry will see opportunities of
blackmailing content and you will see further struggles between

and content as to where the money will go.

If we do that, we will badly break the Internet. Because at
that point when you leave control with carriage, you are back to
what the telephone companies did in the 1970's, which was
nothing. They were very, v
ery good at doing absolutely nothing.

Their only innovation in 70 years was the fax machine, the
only innovative product over the phone. Why? Because if you
are a monopoly, there is no competition and there is no
alternative. We can rebuild it, squash
everything back down to
a very basic model with no flexibility, or you can look further
out to let's say ten years.

Let's just assume that within ten years, because you don't
have longer than five. So within ten years, we're running 6.
If that is the ca
se, this whole market in V4 has a very short
lifetime, a very temporary phase. The whole thing about
grade nets is temporary, and no aftermarket. How many
folks run SNA or Apple tool? When no one uses it, it just
doesn't happen.

So if V6 is goin
g to come, it is going to come within five
years. And within ten years, it is over, there is no more V4
left on the planet. Because quite frankly, if we don't do that,
we're going to build a network which is going to be very, very
different because all t
hose carrier grade application networks
will start vulcanizing the Internet and breaking it apart
because we cannot take the next ten years of the silicone
industry and keep producing one common platform, one common
address space, one name

Do you r
ealize how brittle it is? Trying to do the same
thing all over the world? What is the voltage of electricity in
this country? 110 or 240 volts? Why are they different? Does
it use two vertical spades or two rounded prongs for

Is having t
he switch up "off" or "on" and how did we get
electricity so wrong and how did we manage to never build the
same thing all over the planet. Almost everyone does it

Having one common protocol, one common way of formatting the
bits is actually

quite rare and extremely fragile because
everybody wants to innovate by changing it. And as soon as you
go down that track, the way you do your bits and the way you do
your bits, we're not talking anymore.

Quite frankly if we can't get to V6 in ten year
s, we will be
back to where we were in the '80s with strange networks that
don't talk to each other. That would be a shame because I think
there are better fights to have, much better fights that are
much more interesting because I really like this device

I actually like this one too, but the thing is, there is no
wires, brilliant, just sitting in my pocket and wherever I go,
the Internet is with me. But for this to work, I need a whole
heap of radio spectrum. When I have two and three and four of
e devices, I will need a whole heap more.

The problem with radio spectrum, water absorbs a lot of it.
There isn't a lot of radio spectrum out there, and the
competition for spectrum will be huge.

And the more we want this everywhere, the more we have to

really think about how we're going to utilize that space. I
would rather we actually spend the time over the next ten years
not obsessing about why V6 is necessary. Let's get over that,
and let's actually start talking about something more
interesting a
s to how I can truly get gig bits through the air
through this device, because that is exciting stuff.

But I started 20 years ago. Let's look 20 years to the
future, way out on the horizon. I find that is almost
impossible. I don't think in 1992 you c
ould have sat there and
said you know, this whole thing is going to get to devices like
this in your pocket.

Everybody would have thought you were dreaming because it
just wasn't conceivable. From where I sit now, I am pretty sure
nothing I can conceive
will actually happen in 20 years. I just
lack the amazing imagination.

Because what we've been able to do in the last ten years, by
basically using the foundations of open technologies and open
standards is to build on each other's work, to stand upon ea
other's shoulders and develop amazing products within months by
simply building on other people's work.

And that idea of technology innovation through very small
evolutionary steps has proved a very powerful paradigm. The
world has changed for ten yea
rs because the technology world
figured out the corporate closed proprietary static systems
don't work.

And if you really want a driving, continuous power of
innovation, openness lies right at the heart and soul, and open
standards sit right firmly there.

If we could just figure that out and move on, we would be so
much better off. Because 20 years is hard. And interestingly
if you look back at the last hundred years

and I like this
graph. It is worth looking at.

This is the rate of uptake of techn
ology in U.S. households
across the 20th century. Electricity took 60 years to be
articulated across the country. Even the humble fridge took 30
years. The clothes washer, even now there aren't that many. So
the thing in the early part of the century t
ook up to 40
years, but look at the last part.

How long did it take for the Internet to cover most of the
planet? 20 years? Really 10. VCRs much the same, even cell
phones and computers. So we're actually finding this faster and
faster uptake of te
chnology. So when I look at the next few
years, there is a different dynamic.

20 years ago, what was the future of computing? Ask IBM,
Digital. What is the next 20 years? I wouldn't ask them. I
wouldn't ask the telephone companies; they haven't got a

I wouldn't even ask Apple. I will ask you because we're now in
an environment where what is actually shaping the future is this
idea that computing and communications is just part of consumer
electronics, consumer devices.

All of the sudden this
truly is a post
pc world where
storage and communications is abundant, everywhere. So now it
is all about innovation in a consumer market and that up focus
is where we're heading, and the innovation is all about constant
technology refinement.

But the
assumption behind that dramatic shift is that we can
continue to use open, standard technologies. If we shift the
Internet to V6, you can expect a future where a profound change
will happen, that the Internet will disappear like an
electricity generator.

It will just be part of the fabric of the way we do things
because the observation is true that the most profound
technologies disappear from view. You just assume they're
always there all the time, and I would like to think that's
where we need to go, b
ut I suspect the next five years is really
critical in getting there. If we can't kick this industry into
V6, it is not.

Some of you come from the public sector and the regulatory
sector. You should pay close attention to this because
realistically we'r
e facing the potential for market failure. If
the deregulated market does not choose this path, history will
then judge us as to whether or not we acted appropriately in
preserving a long
term public interest.

If we give back the keys to our future to ba
incumbents who are hell
bent on carriage over content, you will
have given away your future, and that is a market failure. The
only way we can make this work is to make sure that addresses
like computation and communication are abundant.

hat communication becomes not regulated by the scarcity of
addresses, but is enabled by the massive abundance of addresses.
If you can do that, the software folks will do everything else,
and they are just waiting to exercise a universe of connectivity
at even now we have yet to even touch a fraction of.

What we can do in software on massively powerful computers
will boggle the minds of your children, let alone yours. Thank


Hisham Ibrahim
: Thank you, Jeff. To add additional
perspective on th
is, since we manage IPv4 and 6 numbers, that I
can say is over the past two years, I was going over the

November 1st of 2010 to November 1st of 2012, you can see a
dramatic increase of all case of IPv6 space. If you look at the
routing tables an
d IPVC runs user
friendly graphic interface to
show you per region or country the percentage of ASNs announcing
IPv6, you can see those numbers doubled or tripled depending on
the region you look at.

Also a survey was conduct this year which it has been
unning for the past three years. Some of the statistics that
came out of it showed that over a thousand plus ISBs globally,
more than 50% of them had IPv6 internally and externally, and
20% of them had only externally, and 8.5% of them had it only

Martin, could you probably give your feedback on what you
see in the real world?


MARTIN LEVY: Yes, I can. I am Martin Levy from a
company called Hurricane Electric, an Internet service provider
backbone providing bits and connectivity around

the world.

We are actually an Internet service provider that is
considered to be a wholesale backbone. We don't provide
connectivity to an end user on a cable, modem or DSL line or
over a 3
G network to a mobile device.

We provide connectivity to the p
layers that do provide that,
so we sit in this concept of the core of the Internet. And as a
company, although we're about 18 or 19 years old, we decided
about 11 years ago that this V6 thing was important, and we
stood out as really being in a handful of

players in the
commercial side or the actual business of moving bits that want
to look at V6.

It already existed in the ITF world, it already existed as
approved standards. But now we got on to the task of actually
moving bits. So what I am going to ta
lk about, the actual
practicality and show some of the numbers.

Absolutely everything Jeff said, I will reference one thing
out of it, but absolutely everything Jeff said is front and
center. So I will take you from that area to, as I said, the
part of m
oving bits. I want to hit two humorous parts. First,
I am in the commodity industry in the carriage of bits, but
that's okay. It is okay to be in a commodity business.

If you know how to run your business plan, whether you are
in a commodity or in a hi
margin world, that is your choice.
The other one, I was actually at Bell Labs during the invention
of the fax machine.

I wish I had a picture of the first prototype. Ugly beast
that took up this much of a desk and didn't work very well. I
have conso
lidated everything to one slide with five slides on
it. This is how politicians speak, I suppose.

What I have here is essentially five graphs taken out of
world data. Each graph has the identical up and to the
right picture. They are all graphs sh
owing different forms of
measurement of V6 in the real world over short periods of time,
over long periods of time, and I will run through them.

At some point the slides will be available, but these were
actually taken from the web. They are pretty much
available as
both raw and processed data.

Each one will tell a story that is somewhat useful to know
that V6 is real in the real world. That is in fact my message,
that we are in fact running in the real world with real V6 bits,
and that we are making th
e steps.

Unfortunately the percentages are not quite where they want
to be. The first graph, if we start in the top left, simply it
is a measure of every ASN, every separate autonomous network
around the world that has turned on IPv6 in some form.

Doesn't mean they're delivering it to an end customer, but
they have taken the first step of getting V6 into their network.
Sometimes it is a quick period of time before they start
delivering it to a customer or hopefully to their customers,

r times they may implement it internally and not go
further, but it is a measure. The graph meandered for many
years. Prior to 2008, it is pretty sad. But in 2011 in
February of that year, IANA announced it was out of V4 space and
none of the IRRs after

the last allocation would get any more V4

Something magical happened. It want the technical press or
IOR meetings discussed the run
out of V4. Arms showed up in the
Wall Street journal or Bloomberg or the Economist read by the
general public.

t said something on the Internet is to change. For all the
years Jeff spent in front of a podium, microphone and laptop and
for all the years I had done it, we got the message out, but not
quite the way it happened.

The run out of V4 got people have had
in V6. People just
waited until the last minute. The next event that happened,
World V6 day in the middle of June of 2011 when a set of
visibility websites like Facebook and Yahoo and Google got
together and said for 24 hours we're going to turn on
V6 and
measure if there is a problem.

Because hopefully there isn't and all naysayers will be
silenced. That is important because actually people said we
can't do this, there will be problem the. 24 hours test, no
problems. I will show you that graph.

The third event on that graph is World V6 leverage, June of
this year. Will talk about that in a moment. But that graph,
the amount of networks that got interested in V6 and did
something, they made their choices based upon those events, and
they got in

After those event, they waned, the classic problem in life.
You have to show people a deadline, a reason to do something,
you have to show people a lack resources.

So the first graph has been going up. It sits at about
14.8% of network around

the world today. No a great percentage,
but actually a big chunk of them are pretty important.

The second graph, up and to the right. Simply this is a
further view to the routing table. We see more and more people
announcing V6 routes. As a percentag
e of V4, it is tiny. But
equating it, we still get down to 14
15% number.

The people who are easy to convince, they know we have to
take the must edge out further. The bottom graphs aren't about
the routing table, we are about actual Internet traffic.
I talk about the graphs, there are actually two key points about
traffic and IV6 traffic this are important.

Every time a connection is made over V6 versus V4, that is
the movement of bit from one protocol to the other, it isn't
additional traffic;

it is simply the movement of traffic.

The second point, if you have moved bits from one protocol
to another, it means you better look after those bits. It
doesn't matter whether it is a millionth of a percentage or 1%
or a hundred percent, but every bit

is important.

So it is important to make sure as V6 is deployed even at
these low percentage we saw years ago, every bit was deployed
and worked properly. Because we wanted to make sure everybody's
experience with V6 was a good one, because if it wasn't
, there
would be one more naysayer in the world, which wasn't in the
plan to have happen.

Bottom graphs, left to right. The left one shows what
happened on World V6 day. If the graph is hard to read, I will
tell you. At zero UTC, graph went up by a fac
tor of 3
half instantaneously, turning on of the DS records by
Google, Yahoo, Microsoft and a bunch of other players.

They ran for 24 hours, and then they turned off those V6
address records, and the traffic went down again. But during
that time,
the traffic peaked 5
X where it had been. The
traffic graphed similar to what we expect to see humans using
the Internet look like

They walk up, use the Internet. After work they really use
the Internet. And about midnight they finally stop. That is
hen the teenagers go to sleep. They don't go to sleep early or
wake up early. The graph went up and down and we were certain
it was humans, but more importantly in that 24
hour period, very
few complaints. In fact, it was very hard to find people who
d issues connecting to the Internet.

Google and Facebook and Yahoo and Microsoft all still
worked. Google decided the You
Tube video could stay as V6
deliverable. So the graph didn't go back to where are it was,
it stayed up to two and
half to three t
imes the bandwidth,
giving us real traffic to look at from that point on.

This was a good thing, showing faith in V6, zeroing in on
the naysayers saying your time has come we're moving real
traffic. Of course, this is a very small percentage of the rest
of the Internet traffic, but we were looking after it, taking
care of it; and it was a good thing.

Middle graph is a two
year view. And the detail on this is
actually quite interesting. First of all two years before this
graph was done in about July of
this year of 2012, two years ago
the only thing that was really running of any substance as come with the images behind maps turned on.

But it generated traffic. World V6 day showed the jump.
Then for the next 12 month as World V6 launch wa
s marketed to
the rest of the Internet community, what happened was the launch
was a different event than day.

It said: Be ready by this day, don't turn it on by this
day, just be ready two or three months beforehand because we
have proven V6 will work i
n the real world for this.

So as the day approached, we saw the graph go up. It goes
up in step functions as more bits of Google or Netflicks or
Microsoft or Facebook turn into V6 to one data center after
another, one CDM portion to another, bits of Akam
ai or Lime line
turning up for V6.

By the time we get to World V6 launch in July of this last
summer, we are happy with everything going along here. Quite
frankly, there aren't any complaints out there, and this is

We really took a big step forwar
d as a community where we
couldn't have had the story a year ago, we absolutely could not
have had the story three or four years because. We're in a
different position.

The final graph, insight into some of the up from a
structure that operate the Intern
et. It shows a problem and a
solution, and that is actually rather good.

But again the graph up to the right, the yellow graph is the
German Internet exchange in Frankfort, steadily growing from
before July

this is very much zoomed
out multiyear
everything gets averaged out.

But at the beginning of June there is a little peak just
before world V6 leverage where 20 gigabytes of V6 traffic shows
up in Frankfort. And we say oh, this is good, this means there
is V6 traffic out there. Then it

disappears. When traffic goes
away, it worries engineers more than when it shows up.

The question is why did it go away? Turned out, it was some
route that go had not been finally tuned. There was traffic
going from a large content provider into a par
ticular provider
in Rumania via Frankfort, a perfectly acceptable path.

But instead of going over private fiber connections between
two providers, it was going over a public platform, which
actually meant it was completely measurable. So we got to see
. This is a lot of traffic. It turns out these guys in
Rumania had done the other half of the problem of content.

About a year or so ago they said we have millions of
eyeballs in Rumania and we on all of the customer premise
equipment, the routers that
sit at the premise for their
subscribers, we will throw V6 into the software mix and add V6
to our offering.

By the way they didn't tell the millions of users. They
opportunity didn't need to. We knew that from World V6 day. By
this point they rolled o
ut over a million users, more than any
other provider in Europe or Asia. AT&T has done the same thing
in the U.S., shipping more traffic to their customer base as
they have rolled out new software.

That final point may be the most interesting thing about

Because if we're lucky, what we're going to do with our work
both at the bit level and at the higher levels of V6 is deploy
the rest of the IPv6 to the rest of the user community, and they
may never notice.

It isn't up to my mom to know what IPv6 is
, but the reality
is that the end user just wants access to the Internet. There
is sort of a full
stop at the end of the word "the Internet"
whether V4 or V6, they just want their computers, access, ipads
or TVs or whatever to access it.

This shows the i
nternal industry of moving bits around the
world, the commodity industry as Jeff points out, is kind of
doing its bit and will obviously do a lot more. The graphs do
continue to go up to the right. That is what I wanted to share
today. It is a good, pos
itive story, and that is where I will
stop. Thank you very much.


Hisham Ibrahim
: Thank you. Also I would like to add
some more facts of what we can see globally. When it comes to
CCTLDs, out of the 248 CCTLDS out there, 152 have IPv6 on them.

of the 20 current deployed 15 have IPv6. New LTD
process by ICANN, if you attended the session this morning, IPv6
is a requirement in applying for one of those new GTLDs. 9 out
of the 13 main servers also support IPv6. You can find all this
and more from the surveys that I am mentioning on
the NRL website.

I would like to mention another bit of what we found on the
survey before going to my third speaker. The question was what
are the biggest hurdles that you face while deploying IPv6? In
other words, it was vendor supports, knowledge or know
within staff, costs, business case for nontechs and information

I hope Demi can actually expand more from the case he sees
in Brazil to give us more information.


ou very much for being at this very
exciting panel. It will be very hard to make any presentation
after the two brilliant presentations we just had.

I have some slides. I am Demi Getko. I am an obsolete
engineer, but I am following the process of deplo
ying and
fostering IPv6 in Brazil. Just briefly, we get our money with
registering domains dot PR and we use the money for activities
like statistics and running the Brazilian search and having the
points, there are 22 change
points around the whol
country. It is quite a big country.

One of the activities is to have training and dissemination
of IPv6 in some way. It is in the bottom of this light.
Repeating some points we heard brilliantly this afternoon, we
have to have in mind that the Inter
net is based in end

If we can preserve this future, it will be very good for the
Internet and for us. This is just a preamble that we were
saying something about, but it someway it can hurt the principle
of end
end communication.

The second important principle that is important to keep
strength, it can evolve because it is simple in its core. The
other just deliver packets of Internet from one side to another.

We can't burden the Internet with a lot of strenuous thing
and compl
exities. We have to keep the complexities at the
borders of the net, the outer part. This is another thing we
need to keep in mind all the time.

Just to reaffirm this IPV6 is an old thing, it began in
1994, the first complete was description in 1994. T
hen we have
14 years of IPv6. The argument that it is a new thing that I am
not aware of, it is clearly a false argument.

After a short transition time, all users on all the devices
in the Internet that transitioned from IPv4 to IPv6 and then we
can calm
ly, peacefully switch of a the IPv4 and keep all the
things running in IPv6. But unfortunately the thing didn't
happen this way. Maybe the people staying in their comfort zone
and no one took care to make the beginning of the pleasure of

Now we ha
ve a much harder situation because we tonight have
IPv4 in quantity to make this calm, peaceful transition to IPv6,
and we have to think of some kind of transition techniques that
can allow us to make the transition.

All these techniques have good and
bad sides, as we can try
to show to you. From the user perspective, as Martin clearly
said, the user doesn't mind if they are using IPv4 or 6. If a
user comes with IPv4 and accesses services in IPv4, it is all
right, working well, no problems at.

If it
comes from IPB like You
Tube and Facebook, no well,
working with no problem the. But if we are running out of
number and we need to give IPv6 to new commerce, the new users,
they can use dual
stack services, but they cannot access
services that have not m
igrated to dual stack or are only in
IPv4 length.

This is the problem with the red arrow in the slide. Right
now we have to convince the service providers who are running
IPv4 services that they have to use the dual
stack approach.

Otherwise, this big part of the Internet will not be
available for the new users that came with IPv6. Other
challenges are more or less along the same line. Not all
equipment supports IPv6. We have to check different transition

The dual sta
ck is the cleaner, most adequate, but there are
some problems on that. A controversial technique the carriers
are trying to use, the double net, and there are other
techniques. Of course many of them use IPv4 sharing.

I will not go into detail of what
is the double
net, but
maybe you can see in this slide. If a user comes from the left
side using IPv6, you have to convert IPv4 and give another fake
IPv4 before it will go to the world. With IPv6, aiming to go to
IPv6, Internet, no problems at all. But

maybe you have to face
this dual net of translation inside your provider.

What is the problem with this double
net? In some way, it
foster IPv6 deployment. They say okay, we are okay with this,
we have sufficient IPv4 with port numbers a thousand time
number we need, we can keep this going on. In some way the
double knot breaks the end
end model as well as the simple
core principle, and also probably delivers a bad user

Other aspect related to the double net, it requires some
ment, equipment and hardware. The high investment tends
to perpetuate itself, and then it will be a way to delay the
IPv6 deployment, a bad way.

If you go to double net, all the hotel operators will have
your URLs and navigation automatically because the
y have to have
the net inside their premises. Then it can violate the privacy
of the user. In Brazil we try to force IPv6 deployment. The
first way convince operators an vendors to have the option of
IPv6 on their shelf.

The second one is to convince c
ontent providers and e
providers that they have to offer the dual stack thing in IPv4
and 6. The third is with access providers, what will be the
best transition technique we can come up with to avoid the
pitfalls of some of them. We are doing some
free training. We
have trained more than 2000 professional in this area. We have
some free IPv6 transits just to make some experiments.

This tree has a change
point, the biggest in the country; it
is growing. We have more or less 45 system changing tra
ffic in
IPv6 in our big automated systems, not small users. We develop
a site with information, and we also develop a piece of software
that is quite interesting.

There are other versions, but it is only if you want to try.
It is a validator that test
s when you put URL there, it tests
the possibility of being totally compatible with IPv6, or
partially compatible. So then can you try to validate any site
you want, if it is ready for IPv6. Or if it is not already yet
running, available for free on the
site, your use it to test
them. Right now, it is just available in Portuguese and

This is the IPv6 link. We did this world experiment to try
to set up services in a steady way using IPv6, and we really saw
a peak of IPv6 use for internal sta
cks. This is a thing we did
in this year in February. There are a lot of young people
assembled here, and they spend a whole week together.

We put active IPv6, and the people develop it using IPv6.
It was a good way to make the younger hackers aware.
This is
the people

no, not important. Next slide. This resulted
from the event. There is a growth of traffic. Not very big,
but it is important. It is actually related to the events. The
original IPv6 week was the source of one peak, and the IPv6
world launch was the other peak on the graphic.

This is a graph showing the number of websites ending with
IPv6 support, listed like one million top sites. The original
IPv6 week and launch is responsible for the slope on the graph.
This is a table for

location IPv6 in Brazil.

As you see, the slope is to go faster now, but we are
optimistic there will be a good dissemination. Thank you.


Hisham Ibrahim
: Thank you. Unfortunately we had two
cancellations from our panelists today. I will quickly
the two topics they were going to talk about.

They were capacity building and interacting with decision
makers and government mental officials. I will also bring the
perspective. In terms of technical capacity building, the IRRs
are heavy engaged.

We do trainings at face
face meeting, each IR does this.
We are involved with different logs in different regions. We
try to take the experiences and the best practices we reason all
over the globe and try to deliver them in these technical

There are a bunch of them that are roadshowing and have done
as well. All the IARs are heavily involved in getting the
technical message across, and it doesn't necessarily have to be
in person. Some IRRs do webinars.

You can dial in from wherever yo
u are and get the knowledge
of what IPv6, what have topic is being dealt with, whether
technical or name technical.

On decision making part, IPv6 decision maker tract. This is
something we found is very instrumental because a lot of
decision makers addre
ss decisions within their chess or nations.

They focus on some of the development in the ICT world but
still don't see IPv6 as a main course; they still treat it as a
side show.

However when you go to them and discuss, for example, one
topic is there is
no cutting
off date for IPv4 to 6 transition.
However now they are discussing cutoff date for analog to
digital television.

When you ask them how will you build networks IPv4 or 6,
what will you invest now, you find heads continuing and things
checking a
nd making sense. These are some of the messages we
tried to push through these IPv6. Some things we have. Also
there is a lot outreach to not only ISPs, but other networks
that have a presence on the Internet, like mobile Internet, for

In regi
ons like mine in Africa, mobile Internet is the
biggest thing. People are not looking into deploying that much
into fixed infrastructure, going directly mobile.

That is because as Jeff mentioned, it is the next thing.
And in regions like Africa, we have

the lucky of not having
those big legacy networks. We have lectures on LTE over IPv6,
so on and so forth.

These are some of the things we do to try to increase that
knowledge. I would like to now open

one last thing, and I
will have

Mark I was ju
st going to call out on you. One
thing that was mentioned was equipment, vendor support. I know
you ran the IPv6 CPE readiness survey. Let's expand on that
before we on the mic for questions.


AUDIENCE MEMBER: I am Mark, also part of the NRO. I am
a co
chair of the IPv6 working group set forward by the
technical community to basically deal with everything about IPv6

In my previous life as engineer, I am responsible for a
large majority of the IPv6 users in the nether hands. CP was a
oblem, and well, people keep telling me still is. One of the
activities we have undertaken, we did that already in 2009, the
called IPv6 survey.

We are basically looking, and that to does started in
Europe, so it primarily focuses right now on cable n
etworks and
FTC. The basically of people don't if

in 2009 we proved
there were seem out there. We recently did another run, the
2012 edition is now out.

We are still collecting feedback from other vendors. I
think right now our list contains slightl
y over 75% models from
eight different manufacturers, they are all out of the box and
they all support IPv6 out of the box.

But almost on a daily passes people say I can't do IPv6. At
that point it is no longer that the CP is no longer available,
it is
the willingness to invest in new customer equipment by the
operators that already have an install base.

For Africa, slightly different because you have green
deployment and new users. In the European market where
everything is pretty much saturated
, penetration rates over 90%,
huge investment to replace equipment which is a challenge.

That brings me to the point I was trying to make, and
unfortunately connectivity died so I can't check the real
numbers. But to my knowledge off the top of my head,
we have
about 8,000 members, and slightly over 50% of them have an IVP6

If we start looking at routability and how many networks we
see on the Internet, get much less. We are looking at 15
17% of
the networks that support IPv6. If you look one

step further,
that is the point I think Jeff was making, that is the core of
the network. Deploying the core of the network is in the
enough. You need network that deploy IPv6.

But with what is really important and what is driving it in
the end, the ey
eballs. As local as your device in your hand, if
you get new effort phone or android they do support IPv6. That
brings me to the question to the panel, who whose response birth
is that to move forward in terms of regulation, should there be
more pressure

on the network operators to push that down through
the market?

Thank you.

Hisham Ibrahim
: Do we have any questions from participants?
How about from the floor?


AUDIENCE MEMBER: I am from the Swiss systems user group.
Part of what we do is advoc
acy lobbying. I have been wondering
for years if we should try to push our national regulator to do
something to promote the transition?

Is it perhaps too early, like if we push a lot now, somehow
the momentum is too hard to build, and if we push a lot n
then maybe later we have already shot our ammunition? In view
of this risk that is obviously there and we have been aware of
for years, where is the point that we should watch for where we
should realize if our government doesn't take action, we are
oing the wrong direction?

Where is the point we really need to mobilize everything we
have to go all out for advocacy?


JEFF HOUSTON: Both of those are actually fascinating
public regulatory questions. They go straight to the heart of
the regulatory


Because what you are really saying, is this a case of market
fall your, is a market
based distribution if you regard the
transition to V6 as a market function, is the market actually
functioning efficiently?

Are there competitive pressures out

there that are driving
providers to provide V6 service? And the experience that we're
living through says this is amazing. Both in Asia Pacific and
now Europe and Middle East, we ran out of IV4 addresses.

We thought the industry would do something but
all we see a
more carrier
grade nap. I am thinking the case can be made it
is market failure. The carriage operators are not acting in the
larger interest and are actually taking a convenient left turn
that satisfies their own interest at the expense of
and everybody else.

What can you do in a public sector form? The U.S.
government, the Australian government, I can probably think of a
few others but I don't track them closely.

They did the first thing, said to all their public agencies
all their functions of government that are performed over
the net, you have to do this in V6, and they set target dates.
Some of them slipped, some are still a few months out.

But interestingly in a lot of these economies, the public
sector is large. If

you include every high school and
university, it is indeed a large buying force. Don't forget the
Internet started in the public sector as a research activity.

The theory is that if you do this, you drive the commercial
providers into providing V6 services, but I'm not sure that is
everything because it is not the problem of the backbones. If I
can get a V6 packet into people's backbones, I can get it around

world; that's not hard.

It is the last mile, the access network in DSL. It is that
G network, they are busy putting out 4
G networks and Verizon
says yes, ours is V6 but so many operators are putting V4 over
the carrier grade LTE.

I kind of think m
obile services are everything in terms of
revenue and margin per user, mobile is where the money is. And
if mobile moves to 6, services and content are just hanging
around with their mouth open because that is where the money is.

And if we actually manag
e to get 4 G networks to move
quickly, I think the rest will follow. Some of the 4
providers are; many are not. If I was sitting in the regular
where I can apply a tiny bet of incentive and pressure? I would
look at 4
G network rollouts and say: C
an you do something?
And what would it take to get you motivated. The handsets, the
over is high; every two years there's a new one. They are
now doing V6 on radio and wi
fi. It is now base stations and

I would apply pressure there and

clean up the wired network
second. That would be my view. Maybe Martin has a different


MARTIN: Every little bit helps. I am always heartened
when I see one more bit of the network move over. But go back
to the core of the question shoul
d you engage the public sector
now? Later? Should you have done it two or four years ago? Or

I will talk about that part and just say take it as a task.
It never hurts to have the conversation, and it never hurts to
have the anticipate be y
es, we know about this, don't they are
us, it is on our schedule.

Because you can ask again or you, if you hear the no I don't
know what you are talking about, that is the perfect
opportunity, the perfect opening. Shouldn't be too complicated
in a place
like Switzerland. Was that a naive statement? Give
it a try.


AUDIENCE MEMBER: I am sorry to disappoint you,
Switzerland. this thing is a Banana Republic. I am on the
committee that makes the rule for the government, and
essentially the government i
s obligated to do it, but they don't
do it.


MARTIN: I suppose one of the things the government can
do, buy equipment and put all the structure capability with
IPv6. Because as we know, the first fix
time cycle for this
equipment is short. I agree th
e problem is the last mile
because that is where there is a lot of numerous small equipment
that is much harder to change.

If we carry the net and solutions between codes, maybe it
can entrench the problem and delay more the transition. This is
a question I put to the table also, and I ask advice for from
our panelists, if there are alternatives to carrier grade net we
can us
e to avoid this kind of pitfall.


JEFF HOUSTON: The attempts by U.S. government and
European Union to pull Microsoft's obvious monopoly apart were
pitiful in their ineffectuality.

What under apart the monopoly of Microsoft? Google. What
was Google?

It was innovation. Google was doing the same thing
different low. And the reason Google was actually able to
thrive as a business, the network was open and needed nobody's
permission. It was a website that just worked.

But what if you have a network
that is closed and that it is
hard to reach unless you are running content distribution, until
you have multiple millions of dollars out there in deployed
infrastructure, until you are already big, you can't get
anywhere, what happens then? Because at som
e point we will look
at the current monopolies and say hang on, you are kind of big,
and there is a regulatory problem going on with your behavior.

Will I face another Microsoft problem? But this time if we
close up the network and say innovation is over
, you will need
permission to run is new application because all the middleware
in the world stuffed in to make VR then.

Otherwise you will find data out there that individual
government will be powerless to deal with, and I suspect even
the EU and combin
ed regional governments will sit there and go
that is a bit of a problem, I wonder what we will do.

We can sit here and bemoan the fact we're not embracing
openness and let the next five years roll on. When you look
back in five years, you will find you
are in such a dark place
already, the folks who are closing up the network have no
incentive to open it up again, and the innovators who could have
done something long ago went into biogenetics because this field
would have closed up shop.

The choice is h
ere and now, it is a big, dramatic choice.
It is actually about what kind of network you want for the next
ten years. Because if you really want openness, innovation, if
you want a good idea to become an application in a week or folks
to think I can make

a career out of this, if you want that, you
need to act now.

We are now playing with very big forces and the coincident
of computing, and we need a constant overview that keeps it
open. I can't keep it open in V4, and you can't either. If we
start st
uffing it with application devices and middleware, it
isn't a permission network; it is permission
ridden, which means

The incumbents win, everyone else loses. That's certainly
not why I'm here. I would like to think we can persuade folks
s is what is required. I believe it is a market failure, and
like digital boxes had public intervention and public measures,
you need to look and allow folks to change the CP that does
something else. The boxes only cost $50.

Stop making a mountain out
of what should be relatively
minor molehill. If that's true, there is nothing to worry
about. Just do it now; don't wait for a year. The consumer has
always proven it can vote with its wallet. I live in a town
with a cable TV provider and I have a DSL
provider or now a
fiber to the curb plus DSL provider.

One of the nice things about this industry, you get to know
the individual engineers. You say here is my zip code in the
U.S., don't install a carrier grade in that city. I have said
it. I have an
"in" and I admit it, but it probably has no
effect whatsoever.

But the reality is I don't want to live in that world. I
actually have a cable TV provider with a cable modem offing that
runs V6, so I'm root up there in a pretty happy land, to be

But the consumer will vote. Because if it turns out the
other provider has a about the or service, more importantly if
the teenage son realizes that playing with the X
box 360 is
better at his body's house and they have different providers,
they will turn

around and say mom and dad, we need to swap
because it works better with the other provider. If the issue
is carrier grade, one thing that may happen, the V6 thing isn't
much. I still have faith in the consumer and its wallet.


k Elkins. I run an ISP. For the
last five years, I have been running IPV6, and I have it at
home. I was wondering why I don't have IPv6 at this convention
at the moment.




MARTIN: We have actually tested using a client

testing mechanism embedded in online AD. I can say Rumania is
looking at I think around 80% or so.

Some countries have a lot of V6, and there are some where we
get absolutely no V6 response at all. And we tested quite a
number of folks or networks that

claim to be here, and have yet
to see IPv6 packet. So if this conference did have IPv6, it is
possibly a first.


AUDIENCE MEMBER: Permission to disagree with you,
Martin. I have been traveling the same circus for a couple
years. Essentially the con
sumer will vote with his feet, or
will he? Because by the time he is up to the point that he is
voting with his feet, people already broke the network.

Now in a situation where you still have newcomers and
pioneers capable of doing some inventive work, t
he next
generation Googles, you might have a pioneering network that
provides that or the service, but what if that market is locked
up between an incumbent and two other big market players, and
all three break the network equally hard? That customer does

not have a choice. That is the situation Jeff is talking about.

The big networks, I come from a country you can say the
market is divided by three or four players. If all four install
carrier grade, all break it, neighbor doesn't know because his
rk is equally bad.


MARTIN: I this is a hard conference for me. I am very
much about moving bits and very little about moving policy. But
in that scenario, very few times do I say somebody should phone
the regulator and do something. That is where I

change my tone.

But prior to that, I am still a pretty positive
type of guy
and I still think the consumer has some pull, I think.


Hisham Ibrahim
: Before taking the question, I just want
to run a quick survey in the room. How many people in the roo
have IPv6 in their workplace? Fair enough. How many have it at
home? How many have never heard of IPv6 before? Ask your
question, Paul.


PAUL VIXEY: Paul Vixey. Jeff had a great slide showing
two paths, and somebody beside me was asking the ques
tion of
which path will we take? History says we will take both paths.

And with regard to breaking the network, I have been engaged
in trying to deliver new DNS services, new DNS functionality for
most of the last 20 years.

And I can tell you that non
e of them work in most coffee
shops or hotel rooms because some idiot middle
box vendor got
there ahead of me and decided in advance what a DNS packet would
have to look like.

And it is very difficult to add things like IPv6 support or
DNS security
support to a network where the first mover
advantage was a plastic box manufacturer in Taiwan who decided a
long time ago that what you wanted to do should not work. I
know all of that CGN stuff will get done; people will bet that

We all know that t
hey are going to lose. So we know that in
the end, one path works, one path doesn't. But please let's
remember that the brokenness is happening now, and will happen.
We will all combined the price of the people who bet the wrong
way but I very much echo

what Jeff said here. We need to look
to the future, and the future probably makes all
carriage into a
commodity, not a value
added service. Thank you.


Hisham Ibrahim
: At this point I would like to ask for
closing remarks from my panelists. Jeff, M
artin and then Demi.


JEFF HOUSTON: I'M not sure what else there is to say.
It isn't an argument about technology; it is an argument about
openness. We are finding that the city does not move with one
common motivation.

We are actually many parts.
Folks, make the commitment,
carriage, consumers and consumer devices. We are finding the
motivations for shifting the technology for investment in V6
varies depending on who you are and where you see your own

The amazing fact right now is
that almost every single one
of these PCs that are dinosaurs have V6 in the stack and it is
active. These days as long as you trade in the iphone for a new
within you will have radio on radio and the land. Amazing
achievement, shipping stuff around the w

But that last mile is killing me, and I really don't
understand how to move the last mile carriage industry to simply
get a clue in the mass market and actually just move on and
deploy. That is why I'm thinking maybe a word are the regulator
is n
ow time, like analog to digital set top box issue, we are
now facing a wedging of this market and that may be a market
failure that will require some fort of light intervention that
just focuses their attention slightly differently.


MARTIN: I just don
't like the regulator part of that.


DEMI GETKO: I remember the operators in the '80s,
committed to the stack, and they changed their mind, but not
because of the regulator. Maybe we can convince the market and
the people involved that there are reall
y clean solutions and
you have to go forward keeping the Internet open end
end, and
with the core, it is simple.


Hisham Ibrahim
: Thank you all for attending. Feel free
to get more information about what is happening in your region
or country, or
the NRO in general. Thank you very much.


This text is being provided through CART, Communication
Access Realtime Translation. It is intended to facilitate
communication accessibility and may not be a totally verbatim
record of the proceedin