[rtcweb] Fwd: Traffic should be encrypted. (Re: Let's define the purpose of WebRTC)

[Wolfgang Beck <[e-mail]>]

If I recall this correctly, the DTMF payload type was made for a
specific scenario:
1. an analogue phone is connected to a (SIP) device
2. the user presses a key on her phone during a call.
3. The device detects the tone generated by the phone. Relaying it out
of band with some SIP message can lead to undesired effects: a bit of
the DTMF might have escaped the SIP device, was sent as G.711, and was
detected by the remote PSTN party. Now if the SIP message (eg INFO)
arrives and gets translated by a SIP-PSTN gateway into a DTMF tone,
the remote party will see two DTMF events instead of one.

The DTMF payload solves this.

As long as we don't want to attach POTS phones to the browser, outband
signalling of events should be sufficient. If we do connect POTS lines
to the browser, we have to talk about fax as well. Thankfully, modem
ports have disappeared from today's laptops and PCs.


Wolfgang Beck