Re: [DNSOP] Draft for dynamic discovery of secure resolvers

[Doug Barton <[e-mail]>]

What conflicting information?

On 08/21/2018 08:11 PM, Ted Lemon wrote:
> We aren’t even talking about the same thing. I’m talking about figuring 
> out whether we need to offer guidance for how a host implementation 
> would handle conflicting information and, if so, what guidance to 
> offer.  You are talking about one of a number of different ways of 
> configuring DoT.
> 
> On Tue, Aug 21, 2018 at 11:04 PM Doug Barton <[e-mail] 
> <mailto:[e-mail]>> wrote:
> 
>     On 08/21/2018 05:48 AM, Ted Lemon wrote:
>      > On Tue, Aug 21, 2018 at 12:59 AM, Doug Barton
>     <[e-mail] <mailto:[e-mail]>
>      > <mailto:[e-mail] <mailto:[e-mail]>>> wrote:
>      >
>      >     You, like Ted, are looking at the problem the wrong way 'round.
>      >
>      > And this, in a nutshell, is why this discussion has gone on so long.
>      >   If you just caricature what the people you're conversing with say,
>      > then it's inevitably going to go like this:
> 
>     [ Snipped a bunch of arguments I didn't make ]
> 
>      > This is why discussions balloon in the IETF.   So now I have the
>     choice
>      > of either being silenced, or continuing to be Person A in this
>     charade.
>      >   I think I've spoken my peace.   If you want to proceed with
>     this work,
>      > please do not be surprised if, when the call for adoption comes,
>     I come
>      > in and say "I raised substantive objections to this, which were not
>      > addressed, so please do not take this on as a working group item."
> 
>     Ted,
> 
>     While I'm not concerned about the issues you raised in your caricature,
>     I feel that I have tried to engage you in your discussion of different
>     security models. My understanding is that your models devolve down to
>     two. Either the user configures a resolver themselves (whether it's
>     DOH/DOT or not), and user doesn't configure a resolver themselves. I
>     recognize the distinction you made between your models 1 and 3, and
>     further recognize that it's extremely important to some people. My
>     point
>     is that *from the standpoint of a DHCP option for DOH/DOT* it's not
>     relevant.
> 
>       From our discussion, it seems that you're in agreement with me
>     that if
>     a user isn't configuring a resolver explicitly that they are no worse
>     off with DOH/DOT than they are without it. Am I right so far?
> 
>     Meanwhile, you've also voiced an opinion that the presence of a DHCP
>     option implies some sort of endorsement by the IETF. I (and others)
>     replied that we've never heard of this, and disagree strongly with your
>     position.
> 
>     So other than the fact that we disagree on the endorsement issue, what
>     am I missing here?
> 
>     Doug
> 
> 
>     _______________________________________________
>     DNSOP mailing list
>     [e-mail] <mailto:[e-mail]>
>     https://www.ietf.org/mailman/listinfo/dnsop
>