[arin-ppml] About needs basis in 8.3 transfers

[Matthew Petach]

On Tue, Jun 3, 2014 at 12:31 PM, David Huberman <
David.Huberman at microsoft.com> wrote:

>  We had a discussion today at NANOG in the ARIN PPC about needs-basis in
> 8.3 transfers.
>
>
>
> I’d like to state the following, and then let’s see where the discussion
> takes us:
>
>
>
> My team runs an AS. And yep, we’re a pretty big company.  We rely on IPv4
> today for most of our numbering, and will continue to do so for the next
> couple of years.[1]  In the coming year, when we can’t get space from ARIN
> or other RIRs, we have to turn to the market for our IP address needs.   We
> may choose to buy more than a 2 year supply, because it may make business
> sense for us to do so.   ARIN policy, however, only allows us to take the
> IP addresses we buy and transfer the portion which represents a 2 year
> need.   The rest will remain in the name of whoever sold the IP addresses
> to us.
>
>
>
> Why is this result good for the operator community?  Wouldn’t it be better
> if ARIN rules allowed us to transfer into our name all the IP addresses
> which we now own?
>
>
>
> Regards,
>
> /david
>
>
>
> [1] We’re working on increasing IPv6 presence in our network and our
> products, but large corporations move slowly ;)
>

See, this is why I support maintaining the
needs-based decisionmaking around number
allocations.

Because it's far too easy for a really big company
with a couple of billion dollars in the bank to decide
that IPv6 is just too hard, and it's easier to buy up
large blocks of IPv4 space, and keep their critical
resources on v4 addresses--which, if those resources
are crucial enough, could artificially drive up demand
for IPv4.

As a purely hypothetical exercise, let's consider
the three major smartphone OS vendors; each
of them have over $100B in assets in the bank
at the moment; and each of them have a relative
lock on the mobile phone OS for a given set of
handset hardware.  If they each decided that
IPv6 was too hard, and they could get enough
IPv4 space on the unrestricted market (remember,
with over $100B each in assets, at $12/IP, each
of them could in theory be able to afford every
remaining IP address on the planet, with cash
left over).  If they kept their software update
servers on IPv4, they'd have a lock on over
a *billion* endpoints[1] that they would force a
requirement for IPv4 onto.  That would
generate an ongoing demand for IPv4
resources which they would have an
ironfisted control over.  Any mobile carrier
that wanted to provide service for a
smartphone would have to ensure the
phone could reach the IPv4-only software
update servers.

Yes, this is a bit of hyperbole; I'm not expecting
the big three smartphone OS vendors to go out
and buy up all the IPv4 space remaining, just to
have absolute control over the global portable
communications infrastructure.  But the numbers
do indicate that would be a not-infeasible scenario.

Even Apple alone, with their assets in the bank,
could afford to buy up every remaining IPv4
address at current market prices; requiring just
ios devices alone to speak IPv4 to get software
updates would force pretty much every major
mobile carrier in the world to have to maintain
IPv4 infrastructure for the foreseeable future,
and could potentially put them into a situation
where they would have to lease IPv4 addresses
for their handsets from Apple, in order to be
able to provision their own customers.

The ability for one entity to control both ends
of the connection; at the client device, and on the
server side, and with enough cash and assets
to create a monopoly on the means by which
those two endpoints communicate...see, that's
why I will continue to vote in favour of needs-based
justification for resources; because the temptation
to have that level of absolute control over a
resource is too risky to leave without some level
of community input as well.

Thanks!

Matt
(being paranoid, yes--but also recognizing
that greed can drive anti-social behaviours)
P.S. Apologies to Apple--I used you simply as an example
to highlight that in responding to David, I did not mean to
imply in any way that my concern was only about Microsoft;
any of the dominant smartphone OS vendors has the same
level of capability at this point in time.

[1]
http://www.bloomberg.com/news/2012-10-17/smartphones-in-use-surpass-1-billion-will-double-by-2015.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20140611/91519f0e/attachment.htm>