[bfcpbis] Comments on Appendix B of RFC 4582bis

[Gonzalo Camarillo <[e-mail]>]

Hi,

during our last face-to-face meeting, I agreed to have a look at
Appendix B of the bis draft:

http://tools.ietf.org/html/draft-ietf-bfcpbis-rfc4582bis-02#appendix-B

I think it is a good idea to have that type of appendix because when
people (e.g., the IESG) review the draft, the first question we will get
is why have we defined an application-layer UDP-based mechanism instead
of reusing an off-the-self transport mechanism... and that is indeed a
good and very natural question.

The Motivation section is good. It could stress a bit more that BFCP
over UDP was already out there in real deployments and that specifying a
common way to exchange BFCP messages where TCP was not appropriate was
needed in order to avoid the existence of many different ways to do that
in the market. In that way, the text will flow well into the next
section (alternatives considered).

I would move the description of the UDP-based approach in the draft from
the introduction (B.1) to the Alternative Considered Section as *the*
alternative that was finally chosen. Also, that section needs to explain
that we asked the transport area for ways to tunnel TCP over UDP but
that, at that point, there was no consensus on how to do that.

Another alternative that should be described in the Appendix is the SCTP
over UDP approach ala RTCWeb. The text should say that such an approach
was not mature enough (not even fully specified) at that point,
unfortunately.

Cheers,

Gonzalo