[rtcweb] WebRTC-SIP interop: and why SDES-SRTP is a need

[Iñaki Baz Castillo <[e-mail]>]

Hi all,

I've made two "pictures" showing WebRTC and SIP interop for two cases:

1) SDES-SRTP is allowed in WebRTC:
      http://public.aliax.net/WebRTC/WebRTC_SIP_Interop_SDES-SRTP.png

2) Just DTLS-ETK-SRTP is allowed in WebRTC [*]
      http://public.aliax.net/WebRTC/WebRTC_SIP_Interop_DTLS-EKT-SRTP.png

[*] slides 30-35 in
http://tools.ietf.org/agenda/83/slides/slides-83-rtcweb-3.pdf


For those claiming to mandate *just* DTLS-EKT-SRTP in WebRTC, please
see the *cost* of such a decision, and also:

- Thanks for requiring a super Signaling+Media B2BUA/SBC in WebRTC/SIP
interop scenarios. Some vendors will be very happy and will become
very rich. Such a super device (also a DTLS to SDES conversor,
including DTLS key updates to re-INVITE) will be "a bit"... expensive.

- Thanks for disallowing *pure* SIP protocol usage (and instead
requiring SIP B2BUAs/SBCs or custom WebRTC signaling to SIP conversion
gateways). WebRTC is supposed to let the signaling protocol up to the
application, but pure SIP protocol will not be possible since a SIP
B2BUA/SBC is required, and those devices always break/limit the SIP
protocol (*always*).


So IMHO, option 2 ("just DTLS-EKT-SRTP is allowed in WebRTC") is The Barrier.


Best regards.


PS: Note that the same is true for WebRTC/XMPP-Jingle interop.


-- 
Iñaki Baz Castillo
<[e-mail]>