Skip to main content

Shepherd writeup
draft-ietf-rtcweb-video

1. Summary

The abstract does a wonderful job of explaining what this draft is about:

  This specification provides the requirements and considerations for
  WebRTC applications to send and receive video across a network.  It
  specifies the video processing that is required, as well as video
  codecs and their parameters.

As far as where you should point your fingers:
- Sean Turner is the document shepherd.
- Alissa Cooper is the responsible Area Director.

2. Review and Consensus

I’d characterize discussions about this draft as a lengthy, lively, and
spirited and that’s putting it mildly, but I also think if Chuck were
representative of this then this meme would sum it
up:http://cdn.meme.am/instances/500x/58865172.jpg.  There has easily been in
excess of 1,000 email messages about this draft and this draft has been
discussed at many f2f meetings ​(IETF 81, 82, 85, 86, and 88) before the WG
consciously decided to take an entire year off from discussing it before
re-engaging around the IETF 91 timeframe.

From the start, VP8 and H.264 were the front runners though H.261, H.263,
Theora, and Motion JPEG were also considered.  From the 10,000 foot level, the
draw to H.264 was its widespread support and the draw to VP8 was its BSD-like
license and irrevocable free patent license on its bitstream format.  Of
course, VP8 was shown to be widely supported and the IPR “freeness” of VP8 was
questioned (discussed in response to question #3 below). And, H.264’s
widespread support was attacked based on the number of profiles and H.264’s
licensing cost was described as “low”.

Of course, the topic that dominated the discussion was IPR.  But, from the
start everybody that this was going to be the case so there’s text to address
it in the charter, i.e., there’s the possibly that an IPR-encumbered solution
might be selected.  See Section 3 for more on IPR issues.

Lest you think that the entire debate was about IPR, early on technical topics
included video quality and performance as well as status of VP8
standardization, which BTW is moving its way through ISO.  None of these topics
however were interesting enough to maintain the WG's attention while the IPR
debated raged.

As far as the consensus process goes, there were actually three attempts at
reaching consensus.  The first attempt did not result in WG consensus.  After
the 1st attempt failed, the WG explored an alternative decision making process
(the 2nd attempt), which also was fodder for the IETF discussion list, based on
this msghttp://www.ietf.org/mail-archive/web/rtcweb/current/msg09909.html. 
This too didn’t result in WG consensus but a couple of important topics did
fall out: 1) it helped enumerate more options (i.e., it wasn’t just H.264 vs
VP8) and 2) confirmed that the WG did in fact want to decide on an MTI codec.

Leading up to IETF 91, the chairs proposed a series of physical and vocal
calisthenics, which can be found
here:http://www.ietf.org/mail-archive/web/rtcweb/current/msg13358.html, that we
thought would help the WG reach consensus (i.e., third time being the charm and
all).  But, Adam took the wind out of our sails with his so called “novel
proposal", which can be found here:
http://www.ietf.org/mail-archive/web/rtcweb/current/msg13432.html. Based on
list support for the proposal we modified our questions to refer to Adam’s text
prior to the meeting.  So at the 2nd RTCWeb session @ IETF 91 we had
presentations from the draft author on the compromise text and from the VP8 and
H.264 “camps” in support of the compromise.  Then, we had an open mic session
followed by some hums.  The rough consensus in the room was to adopt the
comprise text but there were some objections raised and I pointed these out in
the message sent to the mailing list confirming the room’s
consensus:http://www.ietf.org/mail-archive/web/rtcweb/current/msg13696.html. 
After the allotted time elapsed (and a few hundred more emails), I determined
that the consensus in the room was confirmed.

I’ll note that during the second WG consensus attempt, I purposely requested
that the other two chairs step down from the podium and I, along with some help
from my ADs, made the consensus call and it was later confirmed on the list.

I do not believe there is need for special attention from any directorate
beyond the normal directorate reviews this draft will receive during the IETF
LC.

3. Intellectual Property

Note: Because IPR concerns were discussed so frequently, Scott Brander
presented the IETF’s IPR rules
(http://www.ietf.org/proceedings/85/slides/slides-85-rtcweb-3.pdf).

The following is the answer I got when I queried Adam about the IPR:

<--

I have disclosed all IPR that I am required to under section 6.6 of BCP 79
(i.e., none).

In the interest of full disclosure: although they do not cross the threshold of
requirements under section 6.6; and, in fact, are so nebulous as to be nearly
meaningless to report under section 6.1.3, there are three matters that I will
point out for the sake of completeness.

Due to statements made at the microphone in Toronto, I am aware that there is
possibly a patent or patent application covering the CVO technique cited in the
video draft. I have not taken steps to verify these statements, and no
disclosure has been filed against the draft on the topic.

Although no disclosures have been filed against the video draft on the topic,
the fact that H.264 has patent encumbrances is widely accepted within the
industry. Although most claimants guard the purported patent numbers as trade
secrets, I do note that the following declarations are asserted against the
H.264 RTP format required by the video document:

https://datatracker.ietf.org/ipr/485/
https://datatracker.ietf.org/ipr/502/

Although no disclosures have been filed against the video draft on the topic,
the fact that VP8 has patent encumbrances is widely accepted within the
industry, although the number and identity of companies holding relevant IPR is
a matter of heated debate. I do note that the following declaration is asserted
against the VP8 RTP format required by the video document:

https://datatracker.ietf.org/ipr/1622/

And the following declaration is asserted against the corresponding bitstream:

https://datatracker.ietf.org/ipr/1571/

-->

Cullen pointed out that the MPEG LA patent list is available.  You can search
for "MPEG LA patent list" and get a link.

4. Other Points

***DOWNREF ALERT***: RFC 6386 is a DOWNREF.  Please make sure to explicitly
note this in the IETF LC and add it to the DOWNREF registry.

There are no requests of IANA in this document.

There are normative references to two IETF drafts, I-D.ietf-payload-vp8 and
I-D.ietf-rtcweb-overview; the VP8 draft has been submitted the IESG for
publication (i.e., it’s on Ben’s plate); the RTCWeb overview draft will be
progressing immediately before this draft or at the same time (i.e., there’s no
crazy dependancies here).  Note that unlike many overview drafts this draft is
standards track so there’ll be no downref issue.

There are normative references to specifications from ITU, IEC, and 3GPP but
all are standards so there’s no concern about downrefs.

There may have been more messages sent about this 9-page draft than there are
alphanumeric characters in it.

Back