Re: moving forward on draft-lear-httpbis-svcinfo-rr

[Mark Nottingham <[e-mail]>]

We heard strong interest from at least Patrick, IIRC. I'm sure he'll correct me if I got that wrong.

To be clear - we're NOT planning on implementing this first, nor blocking on it. However, as discussed in Tokyo, it's worth doing in a background thread, because (as Eliot pointed out there), sometimes getting DNS things moving takes a while.

Eliot - a major portion of the conversation was about whether this would be in a new record type vs. in a TXT record. Do you plan to address that in your next draft? 

Cheers,



On 11/02/2013, at 8:56 AM, Mike Belshe <[e-mail]> wrote:

> I am sorry if this sounds too negative.  
> 
> But personally, I don't think this proposal is a good avenue to take at all.  Further, I don't believe there will be any significant implementations of it.  I heard similar comments from others offline, but I'll let them speak up here.
> 
> I would propose:
>    a) Find out if there is really any support for this approach from implementors
>    b) If so, have at it!
>    c) If not, let's table it and move on
> 
> Mike
> 
> 
> 
> On Tue, Feb 5, 2013 at 1:04 AM, Eliot Lear <[e-mail]> wrote:
> Hi everyone,
> 
> At the interim meeting we discussed this draft in several different
> contexts.  Here is my understanding of what people would like to see,
> going forward:
> 
> 1.  Change the InstanceId to be something mnemonic instead of a number,
> so that services can have names.
> 2.  Combine transport protocol and version information into a profile
> 3.  Add a text field that can provide browser hints.
> 
> I have no issues with the first two.  There were two examples given for
> the 3rd: browser hints and BDP.  I think we decided that BDP wouldn't
> work well, so do people agree that browser hints are appropriate for
> DNS?  If so, what's a good example?
> 
> Thanks,
> 
> Eliot
> 
> 

--
Mark Nottingham   http://www.mnot.net/