Re: [AVT] Comments on draft-ietf-avt-tfrc-profile-08.txt

[Ladan Gharai <[e-mail]>]

On Jul 2, 2007, at 11:25 AM, Gorry Fairhurst wrote:

>
> Ladan
>
> I enclose some comments after reading your latest rev of
> draft-ietf-avt-tfrc-profile.  These comments concern questions on RFC
> keywords, and some other queries. I will send editorial comments  
> separately.

    Thank you for raising the point of the RFC keyword usage in the  
draft,  as it was somewhat lax, particularly in Section 7.  I have  
addressed them below and welcome more feedback.
>
> Please do keep me in the loop on further revisions, I'd be happy to  
> read.

    Thank you, I will be submitting a revised draft before the monday  
deadline.
>
>  Best wishes,
>
>  Gorry Fairhurst
>
> ----
>
> Questions:
> Section 7., paragraph 3:
> OLD:
>     Based on initial assumptions on round trip time if more than the
>     recommended 5% is needed for RTCP bandwidth, the applications  
> should
>     use the SDP bandwidth modifiers RS and RR [3556] to signal the  
> amount
>     of RTCP bandwidth needed. Should this assumptions change once  
> the RTP
>     flows are running, the application can recalculate the amount  
> of RTCP
>     bandwidth needed and resignal this new value using its signaling
>     protocol of choice.
> = The phrase "should take into account" needs some clarification
> (a) Is this a RFC2119 "SHOULD"?
> (b) Please clarify "take into account", what does this imply, is it  
> just a
> matter of increasing the advertised bandwidth value?

     This paragraph seems to be a mine field of RFC keywords! below  
is a slightly rewritten one, which I hope is less ambiguous.
>
> - The phrase "should use the SDP bandwidth modifiers RS and RR"
> Is this a RFC2119 "SHOULD"?
>
> Section 7.
>
> - The phrase "the application can recalculate the amount of RTCP  
> bandwidth"
> - Is this a RFC2119 "MAY" or "SHOULD"


     Based on initial assumptions on round trip time if more than the
     recommended 5% is needed for RTCP bandwidth, the applications  
SHOULD
     use the SDP bandwidth modifiers RS and RR [3556] to signal the  
amount
     of RTCP bandwidth needed. If the round trip time assumptions  
change after the RTP
     flows start running, the application MAY recalculate the amount  
of RTCP
     bandwidth needed and resignal this new value using its signaling
     protocol of choice.

>
> Section 8., paragraph 0:
> - The phrase: "the AVPF T_rr_interval variable must not be set"
> - Is this a RFC2119 "MUST NOT"?

      This should be a MUST NOT - as setting T_rr_interval to a value  
higher than the current RTT will disrupt the TFRC timing requirements
>
>
> Section 8., paragraph 2:
> - The phrase: "must signal their use of the AVPF profile"
> - Is this a RFC2119 "MUST"?

       Yes.
>
> - The phrase: "and optionally an initial"
> -  Consider changing to "and optionally MAY sends an initial"

      This should probably be consistent with para 3 of section 7 -  
which says SHOULD


_______________________________________________
Audio/Video Transport Working Group
[e-mail]
https://www1.ietf.org/mailman/listinfo/avt