[arin-ppml] IPv6 Heretic thoughts

Owen DeLong owen at delong.com
Fri Sep 5 14:24:29 EDT 2008


On Sep 5, 2008, at 10:15 AM, Milton L Mueller wrote:

>
>> -----Original Message-----
>> From: Cort Buffington [mailto:cort at kanren.net]
>>
>> I "translated" the concept of moving from IPv4 to IPv6 with my two- 
>> way
>> friend by comparing it to moving from 25kHz channels to 12.5kHz
>> channels in analog FM (for example),
>
> Yes, but this kind of change in fact WAS made repeatedly in the 2-way
> radio space, within the same spectrum. The situation was quite  
> analogous
> to running out of IPv4 addresses. Instead of expanding the amount of
> spectrum used, specialized mobile radio operators made equipment and
> standards changes to enable more intensive use of existing spectrum.  
> It
> is very similar to using NAT or some other reorg of existing address
> space.
>
Actually, it's a lot closer to adding bits to the size of the address.
Literally, they turned what was one channel into two channels
by creating greater precision in the channel numbering.  The
shift from 25Khz spacing to 12.5Khz spacing was analogous
to adding a single bit to IPv4.

It doubles the number of channels available and requires everyone
to upgrade their equipment.  Yes, these changes have happened
repeatedly.  Usually each change takes 3-5 years and goes along
the lines of:

1.	New channel spacing is allowed, not required.  Old radios still
	work, but, new radios supporting narrower spacing are able
	to be deployed and will interoperate with older radios.

2.	0-2 years later, Announcement: In 3 years, old channel
	spacing will no longer be allowed.

3.	Old channel spacing is made illegal.

It would be interesting to see what would happen if the FCC were
to announce today that IPv6 is the new standard for IP addressing
in the US and IPv4 would no longer be legal to use in 3 years.
I think that the conversion effort required here is a bit more expensive
than that required to swap out two-way radio equipment, so, I think
it would cause substantial problems.

I think IPv6 is more akin to the move from NTSC to ATSC.  You simply
can't receive NTSC signals with an ATSC tuner or ATSC signals with
an NTSC tuner.  New televisions are, however, dual-stack capable
and come with both ATSC and NTSC tuners.

> Transfers or no transfers, people will be faced with economic  
> incentives
> of that sort.
>
> So, don't assume that if we just shut off IPv4 then everyone will
> migrate to IPv6. Herrin and others have been making the much more
> realistic point that that may not happen. People may just make more
> intensive use of IPv4 via NAT and other techniques.
>
I think that both solutions will end up deployed in the real world.
However, there is a limit to how far you can go with more intensive
use of NAT.  I think that limit actually comes up far shorter than
Herrin and some others do, but, I've yet to see NAT work flawlessly,
so, perhaps I'm too much of a purist and a skeptic.


Owen




More information about the ARIN-PPML mailing list