Francois,

Thanks for your responses, my responses are identified by <kcj>.

Kevin

-----Original Message-----
From: Francois Audet [mailto:[EMAIL PROTECTED] 
Sent: Monday, March 17, 2008 6:12 PM
To: Kevin Johns; [email protected]
Cc: Rohan Mahy; Cullen Jennings
Subject: RE: [Sip] Outbound-12 comments

 

> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of 
> Kevin Johns
> Sent: Monday, March 10, 2008 11:27
> To: [email protected]
> Subject: [Sip] Outbound-12 comments
> 
> All,
>  
> I did a review of outbound, please find my comments below. 
> This review is for sections 4-8 and the comments are primarily related

> to tightening up the procedures.
> 
> Comments are formatted as follows - <section #>, <section
> title>, <optional text location> - comment...
>  
> Regards,
> Kevin
>  
> Section 4.2.1, Initial Registration, 2nd para - Is the opening 
> sentence well understood by all? "For each outbound proxy URI in the 
> set, the UA SHOULD send a unique REGISTER in the normal way using this

> URI as the default outbound proxy."

I think it's clear. What do you propose we change?
<kcj> I was not necessarily proposing a change. I just wanted to make
sure folks were comfortable with the language 'a unique REGISTER in the
normal way'. When I read this it was unclear to me what unique was
referring to? Is the REGISTER unique in that each must have a different
reg-id or that a unique call-id is used or both? Further if the normal
way is RFC 3261 then why not state a REGISTER per RFC 3261 with the
following enhancements...?

> Section 4.2.1, Initial Registration - this section makes no reference 
> to the keep-timer. It would seem that this should be discussed in the 
> same context as detecting outbound support in the registration 
> response? It is covered in section 4.4 but seems out of place as a 
> first reference.

You mean the Flow-Timer? I'm not sure why we would want to discuss the
Flow-Timer in the section about Registration. Can you be more specific
on what you are looking for?
<kcj> Yes I meant flow-timer, sorry for the confusion. Given there is
text in 4.2.1 to examine the registration response for presence of the
outbound option-tag it seemed reasonable to identify all the parameters
the UA should look for in the response. No big deal if folks are
comfortable with the current text.

> Section 4.2.2, Subsequent REGISTER requests, The first sentence - 
> "Re-registrations and single Contact de-registrations use the same 
> instance-id and reg-id values as the corresponding initial 
> registration." Suggest making this normative

You mean saying:
"Re-registrations and single Contact
 de-registrations MUST use the same instance-id and reg-id values  as
the corresponding initial registration."
<kcj> Yes this is my suggestion

Rohan?

> Section 4.3, Sending Non-REGISTER Requests, 1st paragraph - "UAs that 
> support this specification SHOULD include the outbound option tag in a

> Supported header field in a non-Register REGISTER request." What is a 
> non-Register REGISTER request?

I think this is a typo: it should say "non-Register request.".

> Further why is this recommended? I did not see this information being 
> used by the edge proxy elsewhere in the document?

That is a good question. Rohan? Cullen?


> Section 4.4, Keep-alives and Detecting Flow Failure, 2nd paragraph - 
> "When a successful registration response contains the Flow-Timer 
> header field..." This is the first reference to the Flow-Timer header 
> and it is unclear in the context of this section what a successful 
> registration response is.
> While a proxy is only allowed to insert this in a response containing 
> the outbound option-tag in a Require header field, it would seem 
> reasonable to indicate that it is only used by the UA if outbound 
> support is also signaled. However this may cross with the current 
> thread on keep-alive usage and that a proxy may choose to use 
> keep-alives without outbound and have a desire to control their rate 
> of arrival?

Correct. It would cross with the current tread on keep-alive usage.

> Section 4.4, Keep-alives and Detecting Flow Failure, 2nd paragraph - 
> "For example, it the server suggests 120 seconds, the UA would send 
> each keepalive with a different frequency between 95 and 120 seconds."

> I think 'it' should be 'if'.

Correct.

> Section 5.4, Edge Proxy Keep alive Handling, last paragraph - should 
> recommended timer values be provided which align with the UA 
> recommended values in absence of the Flow-Timer value?

I'm not sure it would be useful, as omitting the Flow-Timer altogether
would have the same result (and would use less bytes in messages).
<kcj> OK

> Section 6, Registrar Mechanisms: Processing REGISTER Requests
> - this section has no discussion on action taken on receipt of a 430 
> response.

Correct. It describes UAS not UAC.
<kcj> I am no longer sure why I called this out, especially given my
comment below on section 7.

> Section 7 (Authoritative Proxy Mechanisms: Forwarding Requests - last
> para) does talk about invalidating bindings to invalid flows but does 
> not reference a 430 response as one possible methods for determining a

> failed flow/binding. I am not sure which section should address this 
> but some text does need to be added to cover this cause.

Yes, it may need some improvements.
_______________________________________________
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