Session Initiation Proposal Investigation (sipping)

Olle E. Johansson | 3 Oct 2011 19:58

RFC 6157 actually updates RFC 3263 too - dual stack DNS lookups

Hi!

I notice that RFC 6157 updates RFC 3264, but does not mention changes to RFC 3263.

RFC 3263 - locating SIP servers - repeatedly says: "the client performs an A or AAAA record lookup of the domainvname". Note the "or"

RFC 6157, section 3.1 says:

Furthermore, there SHOULD exist both IPv6 and IPv4 DNS entries for outbound proxy servers. This allows the user agent to query DNS and obtain an IP address most appropriate for its use (i.e., an IPv4-only user agent will query DNS for A resource records (RRs), an IPv6-only user agent will query DNS for AAAA RRs, and a dual-stack user agent will query DNS for all RRs and choose a specific network.)

The clause about a dual-stack user agent clearly doesn't follow RFC 3263, since it implies "and" instead of "or". There's no MUST, SHOULD or MAY language applied here, so it seems like this is an oversight - not that RFC 6157 is wrong, but that there should have been a more clear update to RFC 3263.

Nitpicking, but I think it's important to clarify the DNS functionality in regards to dual stacks. In addition, I think there's a need for a BCP to explain how a domain can indicate

preference of address family - ipv4 or ipv6 - by using SRV entries.

_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP

headers

Iñaki Baz Castillo | 3 Oct 2011 22:34

Re: RFC 6157 actually updates RFC 3263 too - dual stack DNS lookups

2011/10/3 Olle E. Johansson <oej <at> edvina.net>:
> The clause about a dual-stack user agent clearly doesn't follow RFC 3263,
> since it implies "and" instead of "or". There's no MUST, SHOULD or MAY
> language applied here, so it seems like this is an oversight - not that RFC
> 6157 is wrong, but that there should have been a more clear update to RFC
> 3263.
> Nitpicking, but I think it's important to clarify the DNS functionality in
> regards to dual stacks. In addition, I think there's a need for a BCP to
> explain how a domain can indicate
> preference of address family - ipv4 or ipv6 - by using SRV entries.

I expect that any vendor implementing SIP over IPv4 and IPv6 should
also implement DNS A and AAAA. But I agree that there should be a
mention to it somewhere in a RFC (not a "or" but an "and").

--

-- 
Iñaki Baz Castillo
<ibc <at> aliax.net>
_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP

headers

Vijay K. Gurbani | 5 Oct 2011 16:04

Re: RFC 6157 actually updates RFC 3263 too - dual stack DNS lookups

On 2011/10/3 Olle E. Johansson<oej <at> edvina.net> wrote:
> The clause about a dual-stack user agent clearly doesn't follow RFC
> 3263, since it implies "and" instead of "or". There's no MUST, SHOULD
> or MAY language applied here, so it seems like this is an oversight -
> not that RFC 6157 is wrong, but that there should have been a more
> clear update to RFC 3263.

Interesting reading of the tea leaves.

One thing you may want to do is to file an errata (see "How to Report
Errata" at http://www.rfc-editor.org/how_to_report.html).

Then, the authors and any verifying parties will deliberate on whether
or not this is indeed an errata.  This appears to work for errata in
text, but your claim is that rfc6157 should have updated rfc3263 as
well, so the change is in the masthead of rfc6157 and not in the text
itself.

I suspect that a decision will be made on what to do after an errata
is filed ...

Thanks,

- vijay
--

-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg <at> {bell-labs.com,acm.org} / vijay.gurbani <at> alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP

headers

Olle E. Johansson | 8 Oct 2011 13:44

Re: RFC 6157 actually updates RFC 3263 too - dual stack DNS lookups


5 okt 2011 kl. 16:04 skrev Vijay K. Gurbani:

> On 2011/10/3 Olle E. Johansson<oej <at> edvina.net> wrote:
>> The clause about a dual-stack user agent clearly doesn't follow RFC
>> 3263, since it implies "and" instead of "or". There's no MUST, SHOULD
>> or MAY language applied here, so it seems like this is an oversight -
>> not that RFC 6157 is wrong, but that there should have been a more
>> clear update to RFC 3263.
> 
> Interesting reading of the tea leaves.
Just noted that MSRP has the same issue. Growing to a tea bush 

> 
> One thing you may want to do is to file an errata (see "How to Report
> Errata" at http://www.rfc-editor.org/how_to_report.html).
> 
> Then, the authors and any verifying parties will deliberate on whether
> or not this is indeed an errata.  This appears to work for errata in
> text, but your claim is that rfc6157 should have updated rfc3263 as
> well, so the change is in the masthead of rfc6157 and not in the text
> itself.
> 
> I suspect that a decision will be made on what to do after an errata
> is filed ...
> 
Ok, will do.

/O

_______________________________________________
Sipping mailing list  https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors <at> cs.columbia.edu for questions on current sip
Use sip <at> ietf.org for new developments of core SIP