[AVT] draft-ietf-avt-rtcp-report-extns-02.txt

[Timur Friedman <[e-mail]>]

Hello,

Please find attached the latest version of the (new title) RTP Extended 
Reports (RTP XR) draft, draft-ietf-avt-rtcp-report-extns-02.txt.  This has 
just been submitted to [e-mail]. We have updated the draft 
based upon comments from this list and from the chairs.

We hope to advance to Final Call at the March IETF in San Francisco.  Your 
comments are most welcome so that we may issue a final draft in good time.

Here are the changes since the previous version:

* Changed title from RTCP Reporting Extensions to RTP Extended Reports
  (RTP XR).  (Only very well known acronyms may appear in the title
  without expansion.  RTP appears in the titles of several RFCs, RTCP
  in none to date.)

* Reduced author list from seven authors to three editors.  Created a
  Contributors section.  Nick Duffield added as author.  (No more than five
  names allowed on the front page authors list, mentioned by Colin.)

* Added Magnus Westerlund, Dorgham Sisalem, and Adam Wolisz to the
  acknowledgments.

* Separated normative from informative references.  (Now recommended
  by the RFC Editor, mentioned by Colin.)

* Renumbered the RTP packet type (PT) value for XR packets to 207.
  This would be the next available value should the RTCP feedback
  draft, which is in last call, become an RFC.  Note that this would
  require changes to any other drafts based on the XR packet (RTCP
  Extensions for Single-Source Multicast is one).

* Renumbered the block types from 1 to 7, replacing the legacy values.
  (Requested by Colin in the absence of a clear rationale for the
  legacy values.)

* Rewrote the IANA Considerations section (as recommended by Colin).

* Changed the required behavior for the receiver when it sees a
  non-zero value in a reserved field.  Now the receiver must ignore
  the block or packet that contains this field.  Previously the
  receiver was to ignore the field if they were unaware of any defined
  semantics for the field. (Allows for future definitions of these
  fields to call for different behavior on the part of receivers.)

* All report blocks now have "Report Block" as part of their name.
  (This was inconsistent before.)

* Some of the subsections of Section 4.7 have been combined.

* Sections at the end of the document, such as the references and the
  copyright statement, have been reordered. References divided into
  Normative and Non-Normative References.

* An appendix has been created for algorithms.  The formatting for
  algorithms has been harmonized.

* The discussion of per-sender and per-packet accounting for the Loss
  RLE Report Block (Section 4.1) has been updated, with an algorithm
  in the appendix.

* Variable geometry removed from the Statistics Summary Report Block
  (Section 4.4).  Now, if a statistic is not reported, the field is
  set to zero, but the field itself is not removed.  This should
  considerably simplify parsing.

* Clarified the wording specifying the length fields.  (Thanks to
  Magnus Westerlund.)

* Clarified how late packets are accounted as losses.  (Thanks to
  Henning Schulzrinne and Magnus Westerlund.)

* Made begin_seq and end_seq 16 bits in all cases.  (Thanks to Magnus
  Westerlund.)

* Clarified what to do if there are losses or duplicates when sending
  a Timestamp Report Block.  (Thanks to Magnus Westerlund.)

* Recommendations added to the Security Considerations section.

* Gmin recommended value of 16 indicated for VoIP Metrics Report Block
  (Section 4.7)

* RERL calculation corrected in Section 4.7.  (Thanks to Magnus
  Westerlund.)

* Burst metrics algorithm modified, and discussion updated.

* Field renamed in the VoIP block: doubletalk now residual echo return loss
  (RERL). Updated discussion of this field.

* Fields reordered in VoIP block to better align on 16-bit boundaries.

* Document indenting fixed to conform to that of the RTP spec.

Regards,

Timur Friedman