Re: [v4v6interim] Single namespace

[Andrew Sullivan <[e-mail]>]

On Thu, Oct 02, 2008 at 04:35:47PM +1000, Geoff Huston wrote:
> But we all knew this Andrew, didn't we? So I for one am not sure what point 
> you are making here that wasn't part and parcel of any of these hybrid 
> translation approaches. Am I missing something here in your post that 
> advances something that was not well understood already?

I don't think I'm advancing something that was not understood, but I
am arguing that it is not plain the benefits are worth the cost.  I am
pretty sure there are going to be DNS protocol enthusiasts who think
it's _not_ worth the cost.  Certainly we're going to have to have a
better argument than "We know we're introducing this breakage, so it's
ok."

I can imagine someone arguing, for instance, that if you're an
IPv4-only operator and you want people on v6 networks to be able to
use your services, put a AAAA record for a v4-compatible (or some new,
improved v4-compatible) address in the DNS.  If the response is, "But
v4-only operators might not do that, and v6 users will suffer," then
we're begging the question, because what we're debating is whether
altering DNS answers for some users is ok given the benefit thereby
achieved.  Such a person might also argue that supporting v4 users'
use of v6-only systems is just not a practical goal given the costs,
and if they want to use those services, that will give them an
excellent incentive to add IPv6 support; so for practical reasons, we
shouldn't try to solve that scenario.

Note that there's an important disanalogy between the current
proposals and the corporate network, split-brain case.  If I run a
split-brain DNS installation, I am taking the risk of leaks &c.  It's
my network, and if I want to do funky things inside it, that's
arguably my privilege.  But unless I'm missing something important,
most of the current proposals will change the DNS answers without the
operator of the source DNS answer needing to know anything about it.
The potential for increases in support costs and surprising results is
real, and that will cause people to object.

A

-- 
Andrew Sullivan
[e-mail]
[phone] x104
http://www.commandprompt.com/
_______________________________________________
v4v6interim mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/v4v6interim