From: "Perry E. Metzger" <pmetzger@lehman.com>
Date: Tue, 8 Feb 94 17:02:13 PST
To: rcain@netcom.com (Robert Cain)
Subject: Re: Crypto Regulation Reform
In-Reply-To: <199402090010.QAA22469@mail.netcom.com>
Message-ID: <199402090042.TAA09799@snark>
MIME-Version: 1.0
Content-Type: text/plain



Robert Cain says:
> Perry E. Metzger sez:
> > 
> > > Uhh, could you tell us?  Sounds like quite a breakthrough.  Credit
> > > card sized?  Much cheaper than a modem, like $50 maybe?  And it
> > > digititizes and securely encrypts speech (full duplex?) on the fly?
> > 
> > By definition anything that does this in the digital domain needs a
> > modem, so it can't be cheaper than a modem. None of the analogue
> > methods are going to be terribly secure.
> 
> Remember that a "modem" such as we are used to is a much more complex
> device (at least the firmware, and you do pay for that :-) than what is
> required for simply modulating and demodulating a fixed rate, framed
> bit stream.

This is embarassingly wrong, Robert.

> Today's modem chip sets invariably have a general purpose microprocessor
> to do all the Hayes type stuff and a DSP to do the actual bit stream
> modulation/demodulation (and digital filtering and echo cancelation,
> etc.) where my device can be the DSP alone and requires no RS232 ports
> or the like.  This will result in a saving.

Have you actually looked at one of the Rockwell chipsets in real use,
Robert? They have "all in one" solutions these days. Getting cheaper
than what they sell is almost impossible -- you cannot achieve savings
by "leaving things out" because there is nothing available to leave
out. With the cost of a codec to do something like QCELP and the chip
to do the encryption, you are going to be at least as expensive as a
normal modem anyway just for the parts to manage that component of the
work.

> In short, what is required for a voice-only device such as I am
> initially thinking about is a subset of what is required for a
> computer modem.

I'd be very suprised to see your price predictions come true. I'd be
less suprised to see a secure voice product becaue the mechanisms to
build such things are well understood and hardly revolutionary.

.pm