W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 2012

Re: on DNS records

From: Adrien W. de Croy <adrien@qbik.com>
Date: Wed, 14 Nov 2012 20:59:11 +0000
To: "Eliot Lear" <lear@cisco.com>, "Martin Thomson" <martin.thomson@gmail.com>
Cc: "ietf-http-wg@w3.org Group" <ietf-http-wg@w3.org>
Message-Id: <emb30f9c49-b2c2-4e2c-821f-b6e332d9b244@bombed>

------ Original Message ------
From: "Eliot Lear" <lear@cisco.com>
>On 11/14/12 7:59 PM, Martin Thomson wrote:
>
>
>>
>>  Having multiple records of the same type at the same node is
>>already possible for SRV for load balancing and failover reasons; this
>>would add one more reason: versioning.  Two axes is already a lot, a
>>third is bad enough.
>>
>
>
>That's a fair point.  Question: how would you handle SRV with
>http://www.example.com:49080?
>

in the case of SRV, then 49080 would override port advertised in SRV 
record.

In the case of new RR type, you'd have an unambiguous way to advertise 
in the record to use default or URI-specified port.

But really, in most cases, by allowing advertisement of port for a 
site, people would possibly use a new site / host name instead of 
specifying a new port (except for backward compatibility issues I guess)

I don't know if such a thing exists already in SRV - what happens if 
you say port 0, does that mean just use the default port for the 
protocol?

>
>
>Eliot
>
>
>
>>>
>>>Use of SRV of any form with regard to TLS would require a substantial change
>>>in how we clients validate hostnames.  I tell you from personal experience
>>>that having a new SAN "Other" type is not an easy thing to ask of CAs.
>>>
>>
>>I don't see how you would conclude that.  If you are seeking
>>'example.com', then that is what you should look for in the
>>certificate.  It doesn't matter what you had to query the DNS for to
>>get an IP and port to get there.
>>
>>This happens all the time already with CNAME records - the browser
>>still uses the *input* name to validate the certificate, not some
>>intermediate gunk.
>>
>>
>
>
>I'm saying that TLS processing of hosts with SRV records works
>*sometimes* using other protocols, but at least in this circumstance in
>thinking about it there is no need for SAN Others.
>
>Eliot
>
>
>
>
Received on Wednesday, 14 November 2012 20:59:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 14 November 2012 20:59:28 GMT