🚀 go-pugleaf

In article <41pu48$vgt@otho.cc.flinders.edu.au>,
Rhys D. Winn <rwinn@ist.flinders.edu.au> wrote:
>>>> : It is a fact that Telstra had a software development project in
>>>> : progress in early 1995, who's aim is to monitor and log phone connections
>>>> : through their exchanges, to determine:-
>>>> :   a) Which calls are voice, and which are modem.
>>>> :   b) If modem, what baud rate and protocol.
>>>> :   c) Given (b), to determine whether the link is using encryption.
>>>> :   d) Lastly, given a positive (c), to attempt to analyse what _sort_
>>>> :      of encryption it is.
>
>>>My guess is so that they can time-charge modems, but without 
>>>time-charging voice. This means that they can get more revenue, without 
>>>risking the political firestorm that timed local calls would bring.

I worked on an ATM switch which had phone interface cards that tried
to distinguish between voice and data.  The reason had nothing to do
with cryptography or invading anyone's privacy; it was just to be able
to use less total digital transmission capacity by applying speech
compression algorithms to the voice channels.  This couldn't be done
with the data channels but the next logical step would be to
demodulate the modem tones and send the data in raw form over the
underlying digital carrier.