Why was DTLS-SRTP chosen in WebRTC?

Question

I would like to know the reasons that led DTLS-SRTP to be the method chosen for protecting the media in WebRTC.
I suppose it was considered that it is better to exchange the SRTP key material outside the signaling plane, but why not allowing other methods like SDES ? To me, it seems that it would be faster than going through a DTLS handshake, and as secure as DTLS-SRTP if you can guaranty that the signaling channel is secured...

Answer 1

Because the author of the SDES specification said "please don't use SDES". http://www.tschofenig.priv.at/wp/?p=993 has a good summary of the IETF meeting in Berlin where that decision happened along with links to the slides