On 10 Dec 2014, at 15:18, Phillip Hallam-Baker wrote:
On Wed, Dec 10, 2014 at 3:48 PM, Ted Lemon <Ted(_dot_)Lemon(_at_)nominum(_dot_)com>
wrote:
On Dec 10, 2014, at 3:07 PM, Lee Howard <Lee(_at_)asgard(_dot_)org> wrote:
My opinion on this Last Call: it's about IPv4, and I don't care
about
IPv4
anymore. We shouldn't be bothering with it in the IETF.
This is why I was so surprised by the controversy. Sigh
Unfortunately it seems that a bunch of folk early on decided that the
best
way to motivate the transition from IPv4 to IPv6 was to make IPv6
'better'
and to sabotage any attempts to mitigate the consequences of IPv4
shortage.
So we had the campaign against NAT, even though it was obviously
benefiting
people economically. With 80 nodes on my internal net, I would be
paying
several thousand dollars a year to have static IPs for each (not to
mention
depriving others of Internet access). In fact my ISP now requires me
to run
NAT.
The amount of work-arounds that NAT has caused in various protocols, and
"interesting" behaviors in various applications that have always
expected E2E (because that was the model we told them) is
"entertaining." Am I reading your statement correctly that you would
have preferred the IETF to completely toss the base architecture of E2E
out? The "campaign against NAT" was, for many of us, not philosophical,
but a campaign against application breakage, and state scale in large
networks. NAT has been a necessary evil, but it shouldn't be
encouraged.
In hindsight 32 bits was exactly the wrong size. If IPv4 had been 16
bits
we would have run out of address space long, long ago when the cost of
transition was not so prohibitive - there would only be 65K nodes to
change(!).
Really? Interesting approach. I would have just preferred a larger
space to begin with, rather than breaking things at either the 16bit or
32bit exhaustion point - call me boring, but I would prefer to NOT
mutate base architecture after it is deployed. Of course, that decision
would have required a certain amount of foresight.
The way to achieve transition is to do the exact opposite of the old
strategy. Instead of making IPv6 different, we have to make it exactly
the
same so that the transition cost is minimal.
I can't speak for everyone, but I can assert that, given what we know
now, I would have preferred some different paths in the IPv6
architecture, but again, that would have required what has turned out to
be 20 years of foresight. We can't get everything correct, and if we
wait until we do, we'll never do anything. IPv6 is what it is, and it
does work. It is (way past) time to stop squawking (except at the bar,
where it is allowed) and do what is necessary to get it deployed.
Christopher
--
李柯睿
Avt tace, avt loqvere meliora silentio
Check my PGP key here: http://www.asgaard.org/cdl/cdl.asc
Current vCard here: http://www.asgaard.org/cdl/cdl.vcf