Re: [rtcweb] 答复: 答复: Fwd: I-D Action: draft-westerlund-rtcweb-codec-control-00.txt

[Martin Thomson <[e-mail]>]

On 21 May 2012 18:25, 邓灵莉/denglingli <[e-mail]> wrote:
> [...] I would suggest two options about the security threat statement:
> 1, remove the item c from Section 8; or
> 2, add a few words to notify what is left out and may be of concern in the field and people would know what to expect in reality.
> Would you agree?

I'm not sure about the full implications of these sorts of mechanisms
yet.  I was speaking more generally about rtcweb.  In general however,
it is true that a site (or any entity on the signaling path) can
degrade the experience.

Adding a mechanism like what Magnus proposes lets any participant in a
call perform the same sorts of degradations.  As a site operator, I
want to ensure that this doesn't happen, so that my users always get
the best experience from my site.  So that's not necessarily an
exposure that I like.  Of course, I haven't really analyzed the upside
of this particular trade-off yet...

--Martin

p.s. it might be possible for a mechanism like this to "fix"
downgrades chosen by the site, though that wouldn't be advisable.  The
proposal indicates otherwise: that the constraints set during
negotiation are firm.