Hi,
 
>>If you, as you propose, say:
>> 
>>"If the set has more than one URI, the UAC MUST send a
>>REGISTER request to at least two of the  default outbound
>>proxies from the set."
>> 
>>..do you then really need the sentence saying:
>> 
>>"For each outbound proxy URI in the set, the UAC SHOULD send
>>a REGISTER request using this URI as the default outbound proxy."
>
>Well, yes: it says you SHOULD do them all (but MUST at least 2). The
>battery life stuff is justification for the SHOULD (i.e., why it's not
>a MUST).

There are other reasons than batterly life also, e.g. if the UA is able to 
choose between different access technologies and the outbound proxies are 
access technology dependent. Radio resources is also one issue.
 
And, in visited schenarios, the UA may be provided with a large number of 
outbound proxies in the visited network, but the home registrar still only 
wants to use a smaller number of flows, so the UA has been configured to not 
establish more than that.
 
I do NOT want to spend time on discussing all those use-case, whether they are 
valide etc, but the point is that there most likely will be a number of 
non-battery related use-cases where a UA would not want to register with all 
outbound proxies.


>>Wouldn't it be enough to say that the UAC MUST register with at least two?
>
>
>If it's clearer, I could do this instead:
>
>   For each outbound proxy URI in the set, the UAC SHOULD send a
>   REGISTER request using this URI as the default outbound proxy:
>   however, if the set has more than one URI, the UAC MUST send a
>   REGISTER request to at least two of the default outbound proxies
>   from the set. The reason for not mandating all the proxies in
>   the set is that the UA could limit the number of flows formed to
>   conserve battery power, for example. The reason for mandating at
>   least two is ensure high-availability if available. UAs that [...]
 
I think that looks good. You COULD add, after "battery power", ", access 
resources, or if the registrar only supports a certain number of flows". 
 
But, if you don't want to do that I will not insist. I do not want to delay the 
draft :)


>>I guess the question is whether UACs "with battery life
>>issues" want to be able to establish single flows or not.
>>Since the current draft does seem to allow it (at least it
>>doesn't explicitly forbid it) it could be rather unwise to
>>remove that possibility at this point, because it COULD be
>>seen as a rather big change. Maybe one solution would be to
>>strongly recommend against it, and add some words about the
>>lost functionality by doing so, but not explicitly forbid it.
>
>That's why we are asking for feedback. So far I have heard
>nobody saying that 2 as a minimum is not acceptable.
>
>But we have one stong opinion that 1 is not good.
 
I also agree that 1 is not good, because it basically removes most of the 
Outbound functionality. And, even if we would allow 1, I think it is important 
that the lost functionality is clearly indicated.

>The draft is ambibuous about 2 or 1. I for one read it as saying that 2 was 
>the minimum (because it says "However, a UA MUST support sets with at least 
>two outbound proxy URIs and SHOULD support sets with up to four URIs."
>
>So it's very unclear to to me that we would be "changing" from 1 to 2 "late in 
>the game".
 
Support a set of two, and register with two, are two different things :)

Regards,
 
Christer
 
_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [EMAIL PROTECTED] for questions on current sip
Use [EMAIL PROTECTED] for new developments on the application of sip

Reply via email to