Re: [rtcweb] AVPF [was: Encryption mandate (and offer/answer)]

["Olle E. Johansson" <[e-mail]>]

9 sep 2011 kl. 20:35 skrev Eric Rescorla:

> On Fri, Sep 9, 2011 at 11:11 AM, Matthew Kaufman
> <[e-mail]> wrote:
>> On 9/9/11 10:47 AM, Alan Johnston wrote:
>>> 
>>>   The default will be SRTP - this can be
>>> expressed in SDP without CapNeg.  Should the RTCWEB clients choose to
>>> instead negotiate RTP, then this could be done with a second SDP
>>> Offer/Answer exchange.
>> 
>> I believe you've just designed a downgrade vulnerability.
> 
> Unless I'm missing something, if you (a) support an insecure mode and (b) allow
> negotiation of insecure vs. secure, there's not really any way to
> avoid a downgrade
> issue; the attacker can always pretend not to support security and how do you
> know better? Obviously, it helps if you can negotiate the use or non-use of
> media security over a secure-ish signaling channel, but that doesn't reduce
> the threat from the signaling service.
> 
Have we solved the issues with a "secure-ish signaling channel" anywhere? I don't think we can assume that they exist in the architecture and build solutions based on that assumption. 

SIP/TLS end2end is still very hard to handle. Single hop is definitely protected assuming that we have a CA trust path that we trust, but mutli-hop is still very hard to assure as being "secure-ish" after all these years with SIP.

HTTPS is not an answer here either, especially not with the deployment of enterprise SSL proxys that fake a site certificate in order to be able to listen-in to the traffic.

/O