On Jul 3, 2011 12:29 AM, "Keith Moore" <moore(_at_)network-heretics(_dot_)com>
wrote:
On Jul 3, 2011, at 3:15 AM, Ray Hunter wrote:
Keith Moore wrote:
On Jul 3, 2011, at 2:23 AM, Ray Hunter wrote:
IMHO Right now, we need services with native IPv6 based interfaces,
with equivalent performance and equivalent features and equivalent price
that we have today with IPv4. Anything that detracts from the roll out of
native IPv6 based service interfaces at this time is a bad move IMVHO and
hastens the day that the Internet fragments into a bunch of CGN zones, that
is dominated by businesses that can afford to buy public IPv4 addresses for
their servers or services, or whose business model relies on NAT traversal
being difficult. I personally don't want that sort of Internet.
Right now, applications developers need to be able to write and ship
code that uses IPv6 and can talk to other application instances using IPv6.
Anything that detracts from the ability of applications to use IPv6 at
this time is a bad move IMHO and decreases the chance that there will ever
be sufficient use of IPv6 (of any kind) to justify widespread deployment of
native IPv6.
Given that development and engineering support time is finite, I'd
much rather that 6to4 was declared historic so that developers and engineers
could spend more time on deployment of native IPv6 service interfaces.
I have a better suggestion: let's declare NAT historic. That would
free up lots of developers and engineers to spend time on both native v6 and
better v6 transition mechanisms. Not only would they not need to engineer
new NATs, applications developers wouldn't need to engineer new workarounds
for new NATs. Everybody would win.
Keith
I'm presuming your second comment was facetious.
Mostly. Though I do think that declaring NAT historic is absolutely as
valid as declaring 6to4 historic. Both 6to4 and NAT are things that are
useful in some cases and cause harm in others. Except that 6to4 doesn't
actually cause harm except in conjunction with other dubious practices
(bogus anycast route advertisements, protocol 41 filtering, use public IPv4
addresses behind LSN) which are outside of 6to4's scope, whereas NAT
inherently causes harm.
Right. Because you are not accountable for growing the internet or customer
experience. The people that say kill 6to4 are. I hope that is clear to iesg.
Please look closely at the motives.
Cb
But it wan't a serious suggestion, just an analogy.
I'm also presuming from your first comment that you will thus oppose the
proposal to turn off 6to4 by default.
Am I correct?
I've already said on several occasions that I agree that 6to4 should be
off by default. It's mostly the Historic label that I have the problem
with. (I have other objections to the document also, but those are just
places where I think the wording is misleading. The label is the big
thing.)
Keith
_______________________________________________
v6ops mailing list
v6ops(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/v6ops
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf