Re: [xrblock] [Gen-art] Review: draft-ietf-xrblock-rtcp-xr-qoe-12

[Qin Wu <[e-mail]>]

Hi, Alan:
I think the current text of draft-ietf-xrblock-rtcp-xr-qoe-12 has already supported (i),(ii).
In addition, we follow SDP usage example defined in section 5 and section 6 of RFC5285
use of an out of range value to indicate some exceptional cases.

If we don't use out of range to deal with failure, what else can we do?

Regards!
-Qin
From: Alan Clark [mailto:[e-mail]]
Sent: Thursday, November 14, 2013 4:31 AM
To: Qin Wu; Joel Halpern; A. Jean Mahoney; [e-mail]; [e-mail]; [e-mail]
Subject: Re: [xrblock] [Gen-art] Review: draft-ietf-xrblock-rtcp-xr-qoe-12

Hi Qin

The use of an out of range value within SDP to signal that you don't want to accept something does not seem to be within the spirit of SDP.  The normal usage would be "I can do A, B, C" with an answer of "C".

In the case of QoE measurement - an RTP endpoint may incorporate one or more QoE algorithms that are used to estimate the user perceived quality of the received media stream. The offer should contain a subset of the QoE algorithms supported by the offeror and the answer should contain a subset of the QoE algorithms supported by the answerer; it is not necessary that the QoE algorithms in the Answer are a subset of those in the Offer (as the algorithms don't need to interoperate).  Unlike the case when negotiating a codec - it is not required that the two endpoints implement the same algorithm although there is some minor advantage if the two endpoints use the same algorithm if possible in order that QoE scores for the two directions are equivalently scaled.

I'd prefer to use SDP in the following way:

(i) Offer provides a list of one or more QoE algorithms that it supports, in the form of a map

(ii) Answer provides a list of one or more QoE algorithms that it supports, in the form of a map. If the Answerer supports one of the same algorithms that the Offer provides then the Answerer SHOULD answer with only that algorithm however MAY answer with more than one.

I'll work out the appropriate text changes to implement this unless there are strong technical reasons not to.

Regards

Alan


On 11/12/13, 9:14 PM, Qin Wu wrote:
Hi, Alan:
Thank for quick response.  Here are additional changes I proposed.
Regarding section 4.1, I propose the following change:
OLD TEXT:
   mapentry =  "calg:" 1*5 DIGIT ["/" direction]
                          ;Values other than 4095~4351 are valid
NEW TEXT:
  mapentry =  "calg:" 1*3 DIGIT ["/" direction]
                          ;Values 1..255 are valid
OLD TEXT:
"
If the answerer wishes to reject a mosref attribute offered by the
offerer, it sets identifiers associated with segment extensions in
the answer to the value in the range 4096-4351.
"

NEW TEXT:
"
If the answerer wishes to reject a mosref attribute offered by the
offerer, it sets identifiers associated with segment extensions in
the answer to the value in the range 512-767.
"

OLD TEXT:
"
   If a party wishes to offer mutually exclusive alternatives, then
   multiple segment extensions with the same identifier in the
   (unusable) range 4096-4351 MAY be offered;
"
NEW TEXT:
"
   If a party wishes to offer mutually exclusive alternatives, then
   multiple segment extensions with the same identifier in the
   (unusable) range 512-767 MAY be offered;
"
OLD TEXT:
"
Similarly, if more segment extensions are offered than can be fit in
the valid range, identifiers in the range 4096-4351 MAY be offered;
"
NEW TEXT:
"
Similarly, if more segment extensions are offered than can be fit in
the valid range, identifiers in the range 512-767 MAY be offered;
"

OLD TEXT:
"
Note that the range 4096-4351 for these negotiation identifiers is
deliberately restricted to allow expansion of the range of valid
identifiers in future.
"
NEW TEXT:
"
Note that the range 512-767 for these negotiation identifiers is
deliberately restricted to allow expansion of the range of valid
identifiers in future.
"

Regards!
-Qin
From: [e-mail]<mailto:[e-mail]> [mailto:[e-mail]] On Behalf Of Alan Clark
Sent: Wednesday, November 13, 2013 1:06 AM
To: Joel Halpern; A. Jean Mahoney; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Subject: Re: [xrblock] [Gen-art] Review: draft-ietf-xrblock-rtcp-xr-qoe-12

Hi Joel

Thanks for your comments.

(i) Section 1.4

Proposed change

   The MOS Metrics Report Block can be used in any application of RTP

   for which QoE measurement algorithms are defined.
to

   The MOS Metrics Report Block can be used in any application of RTP

   for which QoE (Quality of Experience) measurement algorithms are defined.
(ii) Section 3.2.2

Proposed change
"The 8-bit ID is the local identifier of this segment in the range 1-255 inclusive"
to
"The 8-bit CAID is the session specific reference to the calculation algorithm and associated qualifiers indicated in SDP (see Section 4.1) and used to compute QoE scores for this segment"

(iii) Section 3.2.1

Proposed change
"The 8-bit CAID is the local identifier of calculation algorithm associated with this segment in the range 1-255 inclusive. "
to
"The 8-bit CAID is the session specific reference to the calculation algorithm and associated qualifiers indicated in SDP (see Section 4.1) and used to compute QoE scores for this segment"

(iv) Section 4.1

Proposed change

   mapentry =  "calg:" 1*5 DIGIT ["/" direction]

                          ;Values other than 4095~4351 are valid

to

  mapentry =  "calg:" 1*3 DIGIT ["/" direction]

                          ;Values other than 1..255 are valid
and remove

 mostype = "mostype=" ("e"; Estimated MOS [P.800.1]

                           /"s";subjective MOS [P.800.1]

                           /"o";objective MOS [P.800.1]

                           /non-ws-string)
We will see if there are additional comments and then update the draft

Regards

Alan Clark


On 11/12/13, 7:12 AM, Joel Halpern wrote:
I am the assigned Gen-ART reviewer for this draft. For background on
Gen-ART, please see the FAQ at

<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq><http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.

Please resolve these comments along with any other Last Call comments
you may receive.

Document: draft-ietf-xrblock-rtcp-xr-qoe-12
    RTP Control Protocol (RTCP) Extended Report (XR) Blocks for
                       MOS Metric Reporting
Reviewer: Joel M. Halpern
Review Date: 12-November-2013
IETF LC End Date: 27-November-2013
IESG Telechat date: N/A

Summary: This document is nearly ready for publication as a Proposed Standard RFC

Major issues:

Moderate issues:
    In section 3.2.2 on Multi-Channel audio per SSRC Segment, the format description for the Calculation Algorithm ID (CAID) reads:
"The 8-bit ID is the local identifier of this segment in the range
1-255 inclusive."  I am pretty sure this is supposed to be an algorithm ID, not a segment index?

    The text in section 4.1 indicates that the number after "calg:" in the mapentry of the calgextmap is used as the ID in the CAID of the xrblock.  The packet format only allows 8 bits of value.  So why does the SDP format allow up to 5 digits?  Also, is there some reason that the special values 4095-4351 (in section 4.1) or 4096-4351 (in section 4.2) are used rather than say equally invalid 512 through some appropriate upper bound still in 3 digits?

Minor issues:
    Please ensure that all acronyms are expanded on first use.  For example, QoE is not expanded.

    The notes in B.3 indicate that mostype was to be removed from the SDP grammar.  But it is still defined.  And section 4.2 still mentions it, even though it does not get referenced by the message format. Please finish removing it.  (also "most type")

Nits/editorial comments:
_______________________________________________
xrblock mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/xrblock