Re: [v6ops] Last Call: <draft-ietf-v6ops-v6-aaaa-whitelisting-implications-08.txt> (Considerations for Transitioning Content to IPv6) to Informational RFC

[Joel jaeggli <[e-mail]>]

On 2/9/12 01:25 , Lorenzo Colitti wrote:
> On Thu, Feb 9, 2012 at 00:36, Joel jaeggli <[e-mail]
> <mailto:[e-mail]>> wrote:
> 
>     Ops is not marketing.
> 
> 
> And if I were looking for a marketing venue, a standards body that
> produces ASCII text documents read by a handful of engineers would not
> be high on my list. This is not about marketing.
>

Sorry for being so droll, I found it hard to restrain myself.

> 
>     If you're saying some flag day makes the contents of the document no
>     longer operationally relevant after a given date, I'll take the point
>     but disagree.
> 
> 
> I think you're missing my point.
> 
> It seems to me that approximately 30% of the non-biolerplate text in
> this draft discusses DNS whitelisting. (And in fact, in its original
> form the draft entirely on DNS whitelisting - hence the filename. The
> rest was added later.)
> 
> Whitelisting is a practice relevant to a few large websites (since
> nobody else is using it). It so happens that the websites that employ
> this practice are going to stop using it, all together. Given the cost
> and implications, I'd say practice is unlikely to be resurrected.

I do not belive that the selective (inclusive) return of A or A + AAAA
records on the basis of source address is likely to end on a particular
day. It may well for you and some others, which is fine, or you may find
it necessary again, or it may become a list of exclusions rather than
inclusions. I belive you're on record indicating as much. In any event
others may find it necessary.

> So, you decide to tell the whole story, and talk about whitelisting
> *and* World IPv6 Launch. Or you can decide that whitelisting will soon
> be irrelevant, and not talk about either whitelisting or World IPv6
> Launch. But you can't talk about whitelisting without talking about
> World IPv6 Launch, because if you do, your document is missing the key
> piece "how do you remove the whitelist", and that's a disservice to its
> readers.
> 
> To be more specific, at least section 5.5 ("it is unclear
> how implementers will judge when the network conditions will have
> changed sufficiently to justify turning off DNS Resolver Whitelisting
> and/or what the process and timing will be for discontinuing this
> practice") is now incorrect. It *is* clear, and it's what those
> implementers are doing as part of World IPv6 Launch.

Invidual service operators like you and I are likely to make decisions
on the basis of our instrumentation, we may well alter their behavior on
a uni or multilateral basis, and some of us may do so for world ipv6
launch. ipv4/v6 Transition is not something with a flag day however, and
I do not believe that the concerns embedded in the draft will be
fundamentally altered on 6/6/12.

> Does that make more sense?

yes, that doesn't imply that we're in concert however.

> Cheers,
> Lorenzo