On 04/25/2011 03:41 PM, Iñaki Baz Castillo wrote:
> DNS is supposed to have global scope. This is, a domain is supposed
> to exist (or not to exist) regardless o the nameserver you ask. So in
> the case of non existing domain, I still see no benefict in using
> 604 instead of 404. Perhaps I miss something.
Think of it this way: say you are attempting to jaywalk
and are told by a random person not to do so (a 404). You
probably will not listen to that person, or listen
to that person and not jaywalk across that intersection,
but continue jaywalking at other intersections.
Now, instead of a random person, a police officer tells you
not to jaywalk (a 604). Who will you consider has more
authority: a random person or a police officer?
Most probably you will choose the latter and change your
behaviour by not jaywalking --- at least until your fear has
subsided and you are ready to be adventuresome again ;-)
> Yes, but AFAIK a 404 should not trigger a new client transaction in
> an UAC or a proxy as 404 means "destination doesn't exist". What is
> the real difference between receiving a 404 and a 604? From the
> client transaction point of view, it should be the same, am I wrong?
The behaviour on receiving a 404 is different in the proxy.
The 404 allows the proxy to continue spawning new client
branches. The 604 inhibits the proxy from doing so.
> I'm lost now. I missed such section of the RFC. Does it mean that,
> for my initial question (the outbound proxy receives a request or a
> non existing domain), the proxy must reply 408 rather than 404/604?
>
> However, note that RFC 3261 16.6 section 7 says: The proxy does not
> need to place anything in the response context, but otherwise acts as
> if this element of the target set returned a 408 (Request Timeout)
> final response.
>
> But take into account that in case of a non-existing domain, there
> is no target set (as RFC 3263 procedures returned no
> transport:IP:port tuple) so, wouldn't be 404/604 the appropriate
> choice?
Depends on a particular implementation. Some may consider this
to well be a 408.
Thanks,
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / [email protected]
Web: http://ect.bell-labs.com/who/vkg/
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
