[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: I-D ACTION:draft-wing-media-security-requirements-00.txt
"Hadriel Kaplan" <HKaplan@xxxxxxxxxxxxxx> writes:
> 1) I don't believe there was a vote on "R3: With forking, only the entity to
> which the call is finally established, MUST get hold of the media encryption
> keys" explicitly. The slide we voted on was "Forking and Retargeting
> MUST/SHOULD be secured". That was ambiguous, at least for me in Montreal.
> The way the slide was discussed it sounded like "do we still need to have a
> forked request secured", whereas what I think you really meant, now that I
> see this requirement, is "must we prevent all forked parties from seeing the
> keys, and instead constrain it to only the final chosen one". A very
> different question, with a very different answer for me.
Hmm... I certainly got the second meaning out of the discussion, but
then that's what I always thought it meant. Can the chairs weigh
in on what they thought?
> 2) R6: A solution MUST provide protection against passive attacks. I think
> you need to define what a passive attack is in the draft. It was also an
> area where I thought the voting was ambiguous. In the earlier slides they
> implied a proxy, in the SIPS/TLS hop-path of an sdesc request, would be
> considered a passive attacker; so sdesc through SIPS through proxies would
> not be sufficient to be considered a valid rtpsec model. But I don't think
> that's what the requirement voting slide implied, and not what some people
> voted on in Montreal. At least I didn't. I was voting that the keys had to
> be exchanged over a secure transport mechanism, so that a sniffer wouldn't
> see 'em. Now I know some people here feel a SIPS proxy and a sniffer are
> the same thing, but I don't. :) But I'd at least like to know whether it
> is or not in this draft's view.
I think our goal should be to design a system which to the greatest
extent possible protects against malicious proxies. I think this
implies that keys in the clear over hop-by-hop protected channels
aren't acceptable. That's what I thought I was voting on in
Montreal.
-Ekr