Re: [keyassure] The draft and subj alt names

["Richard L. Barnes" <[e-mail]>]

Let me phrase it differently.  If you give DANE specific name matching requirements, then you not only have to update the appropriate application RFCs, you have to change all the corresponding code.  If you look at HTTPS libraries, for example: 
-- Most of them will only check the cert against the domain name in the HTTPS URI (with no user-configurability)
-- Some can be configured to accept a list of names
-- None that I'm aware of have a "don't check the name" option.  
(Some of them don't ever check names, but they're non-compliant anyway.)

This is not to mention that if you put a cert with the wrong domain name in TLS, you're going to break non-DANE clients.  

I don't think that all this trouble is worth the small optimization, and (not to be a process weenie) I don't think that working at the application layer is within the scope of this working group.

--Richard


On Apr 1, 2011, at 5:15 AM, Jay Daley wrote:

> On 1/04/2011, at 3:06 PM, Richard L. Barnes wrote:
> 
>> The only choice that avoids updating several RFCs is 5.
> 
> Is that your only reason for not wanting to tackle this?  Do you have any view on the nature of trust that DNSSEC provides to applications or not?
> 
>> Just leave it alone!
> 
> No.
> 
> warmest regards
> Jay
> 
>> --Richard
>> 
>> 
>> On Mar 31, 2011, at 10:41 PM, Jay Daley wrote:
>> 
>>> On 1/04/2011, at 12:12 AM, Richard L. Barnes wrote:
>>> 
>>>> As I've said in other threads, this WG should be silent on requirements for certificate names, since different applications use TLS certificate names in different ways.  See for example the difference between RFC 2818 and RFC 6125.  The text you reference in the draft is technically harmless ("must" not "MUST"), but should probably be deleted anyway.
>>> 
>>> There are two separate issues here:
>>> a) whether we think name matching is still needed with DANE
>>> b) what the draft should say about that
>>> 
>>> On a) I would make the case that name matching is should not be done because
>>> - DNSSEC provides sufficient trust
>>> - name matching forces certain operational decisions onto end user (e.g. multiple certs, cert reuse) that are unreasonable
>>> 
>>> On b) There are six options for the draft:
>>> 
>>> 1.  State that name matching must be done
>>> 2.  State that name matching must not be done
>>> 3.  Leave it to applications with a recommendation to match
>>> 4.  Leave it to applications with a recommendation not to match
>>> 5.  Leave it to applications with no recommendation
>>> 6.  Add a flag into TLSA that specifies whether name matching is to be done or not
>>> 
>>> My choice, in order of preference would be 2, 6, 4.  
>>> 
>>> Jay
>>> 
>>>> 
>>>> --Richard
>>>> 
>>>> 
>>>> 
>>>> On Mar 31, 2011, at 12:05 PM, Olle E. Johansson wrote:
>>>> 
>>>>> I have read the draft and reacted to one thing. Sorry if this already have been discussed on the list.
>>>>> 
>>>>> The draft -06 says
>>>>> "The end entity certificate from TLS, regardless of whether it was
>>>>> matched with a TLSA type 1 certificate or chained to a TLSA type 2 CA
>>>>> certificate, must have at least one identifier in the subject or
>>>>> subjectAltName field of the matched certificates matches the expected
>>>>> identifier for the TLS server. "
>>>>> 
>>>>> The new RFC 6125 tells us that we should:
>>>>> "Move toward including and checking even more specific
>>>>>  subjectAlternativeName extensions where appropriate for using the
>>>>>  protocol (e.g., uniformResourceIdentifier and the otherName form
>>>>>  SRVName)."
>>>>> 
>>>>> I feel we need to be more specific in the draft.  What type of subjectAltName should we match with?
>>>>> dnsName?
>>>>> 
>>>>> And if we're using SIP, and have URI's in the certificate - should that be part of this draft
>>>>> or should we open for protocol-specific specifications for other types of subject Alt Names?
>>>>> 
>>>>> /O
>>>>> _______________________________________________
>>>>> keyassure mailing list
>>>>> [e-mail]
>>>>> https://www.ietf.org/mailman/listinfo/keyassure
>>>> 
>>>> _______________________________________________
>>>> keyassure mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/keyassure
>>> 
>>> 
>>> -- 
>>> Jay Daley
>>> Chief Executive
>>> .nz Registry Services (New Zealand Domain Name Registry Limited)
>>> desk: [phone]
>>> mobile: [phone]
>>> 
>> 
> 
> 
> -- 
> Jay Daley
> Chief Executive
> .nz Registry Services (New Zealand Domain Name Registry Limited)
> desk: [phone]
> mobile: [phone]
>