[arin-ppml] New Entrants shut out? (Was: ARIN-2011-5: ... - Last Call

[Owen DeLong]

On Apr 30, 2011, at 12:50 PM, Jeffrey Lyon wrote:

> On Sat, Apr 30, 2011 at 3:28 PM, Owen DeLong <owen at delong.com> wrote:
>> 
>> 
>> Jimmy,
>> 
>> At the moment we have a system that favors small players at the
>> expense of commerce. It also fails to create economic incentives to
>> migrate to IPv6. Note that C-Squad execs speak dollars, not value to
>> the community.
>> 
>> I would argue that the current system appears to be creating quite a few
>> incentives to add IPv6 capabilities if you look at the current uptick in v6
>> statistics since Feb. 3.
>> 
>> So long as we continue to squeeze blood out of the IPv4 turnip,
>> companies will continue to delay IPv6. The choices become the Lyon
>> strategy of letting the market set the price and encourage natural
>> migration, or the Owen strategy of taking IPv4 off life support.
>> 
>> I don't think it is on life support and I think a natural evolution is
>> occurring. Some organizations which are unable to look beyond short term
>> dollars will have faster and more disruptive migration processes while
>> others with better vision have been planning and executing their migration
>> strategies for years.
>> The longer you wait, the more rushed, expensive, and disruptive your
>> inevitable transition will be.
>> Owen
>> 
> 
> Owen,
> 
> I agree that there are some incentives to migrate to IPv6 and that
> companies who wait will suffer. My point is that these incentives are
> not economic in nature which is what will be necessary to motivate
> companies to act. Companies are the driving force behind either
> creating or removing roadblocks to adoption (eg. carrier support,
> vendor support, and so forth).
> 
I will say that not ALL incentives are economic in nature.

However, the economy of failing to implement IPv6 is a pretty
strong incentive if you fully understand the results.

Becoming increasingly disconnected from more and more of your
potential customer base should, for any rational business, create
a strong economic incentive to do something different. That is the
inevitable result of remaining IPv4-only.

Further, the costs at the carrier and the subscriber level of continuing
to attempt to maintain some level of connectivity to a growing
internet incapable of producing additional IPv4 addresses for those
new connections will significantly increase the costs of remaining
on IPv4 and keeping it functional. These costs are only increased
by failing to add IPv6 capabilities.

> I fundamentally disagree that we should taking the position of
> "migrate now or suffer later," rather create the economic incentive
> for a natural progression to IPv6 without having to twist any arms or
> cause any suffering.
> 
Huh?

1.	I think that migrate now is a misnomer. What I am saying is
	"Add IPv6 capabilities to your network now, or, you will,
	inevitably suffer later."

	That's not to say I am trying to cause that suffering or that
	anyone is out to inflict additional suffering on those who
	fail to add IPv6 capabilities. The NATURAL CONSEQUENCE
	of failing to adapt to a changing and growing internet is
	suffering. It will happen to those that do not add IPv6
	capabilities no matter what anyone else does or does not
	do.

2.	There is already a great deal of economic incentive there
	for anyone who takes the time to understand the economics
	of the situation. IPv4 will inevitably become increasingly
	more costly.

	This policy is intended to allow a single /10 to be used
	for purposes that will otherwise require multiple providers
	to collectively use much more than one /10 and will ultimately
	result in accelerating the time at which IPv4 becomes
	more expensive and accelerating and increasing the
	resulting suffering.

> Very rarely are technologies widely adopted on account of a decree.
> Successful adoption occurs when migrating to a new technology is an
> all around attractive prospect. I would be willing to hypothesize that
> there are about equal numbers of people who want immediate adoption of
> IPv6 and those who want to see IPv4 continue to survive for a number
> of years, at least until IPv6 can gain a more stable footing.
> 
I think I can say that almost everyone wishes that IPv4 could continue to survive
for many more years. I wish that were true.

I think, instead, you can say that there is some disagreement among professionals
as to whether this is actually a viable strategy or not.

I don't know of anyone among the IPv6 cheerleading crowd (and I'm pretty
sure I have about as much of an IPv6 cheerleader perspective as anyone)
who wouldn't love to see a way for IPv4 to provide a real solution to continued
growth in the internet without the disruption or inconvenience or costs of
deploying a new protocol. The difference comes in the subtlety of what
people are willing to accept as a "real solution".

Some of us want to see the internet continue to be able to provide at least
the services and capabilities that exist today. Some are willing to accept
a greatly reduced subset. Some of us want to see expansion and continued
innovation.

For the first camp (existing capabilities), IPv4 has a very limited lifespan,
but, until the RIRs all run out and maybe even for a few months thereafter,
we can kind of limp along and keep that somewhat functional.

For the second camp, there is NAT444 and if you are willing to reduce the
internet to basic HTTP/HTTPs and SMTP and little else, then, you can
probably keep IPv4 somewhat serviceable for a few more years.

Finally, for the third camp, there really is no remaining viability in IPv4.
The end-to-end model failed with the introduction of NAT. The peer to
peer nature has been reduced to the consumer/provider model as
a result and services which people want (remote access to resources
at home, for example) have to involve rendezvous servers hosted by
other providers to be made possible. There are so many cool things
that we already know how to do and have the technology for that
cannot be implemented for lack of end-to-end connections that I
think there is additional economic incentive available here as well.

Indeed, the question is not whether or not IPv6 has economic
incentives. The question is how to best explain and convey those
economic incentives to the CxOs of the world.

In the meantime, this policy seeks to provide a way that many providers can
coordinate their use of a very limited amount of IPv4 space for a
specific purpose rather than requiring each of them to get their own
distinct space for this purpose.

Owen