Is there any modem connection sniffer hardware/software?

I mean a device (tap) that could be connected to a phone line,
and to record network traffic made over dial-up connection over
that telephone line (for V.90 protocol for example). I am not
reffering to those "com port sniffer" or "modem sniffers" that
have to be installed on the "target" PC, but something that would
be used outside the house, by tapping the telephone lines.


Any hint and/or link is more than welcome!

Zorpetus wrote:
Since you would be recording an analog (voice-like) signal, I
beleive the device you are looking for is called a "tape
recorder".

-reed

Reed <reedh@rmi.net> wrote:
That doesn't provide data though, just noise.

I don't know if anyone does in fact make such a device. Years ago
HP made a piece of test equipment that would lock up on various
private line modem protocols, but did not deal with v.32 (and v.34
didn't exist yet).

Hence such a device can, and probably does exist. It cost about
$10,000 a pop too.

--
FloydL. Davidson <http://web.newsguy.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) floyd@barrow.com

OK, but what about "echo cancelation" on V.90 modem communication?
I don't think that this communication couldn't just be recorded to
some tape recorder (or maybe you can, but some special kind of it is
required).

Anyway, you have it recorded (or you are collecting those signals in
real time) - but what to do with them - how to convert them into some
digital data (frames or packets)?

zorpetus@yahoo.com (Zorpetus) wrote:
Because of the way both v.90 and v.34 modems both synchronize
and use the same frequency spectrum, tapping a connection is
necessarily a "4-wire" operation rather than a "2-wire"
operation. A typical analog POTS voice conversation can be
tapped by merely bridging the 2-wire drop line, but to decode
the modem tones in each direction requires that the line be
broken and two hybrids be inserted to provide a separate path
for send and receive (a 4-wire interface).

<-- Subscriber ISP -->
+-------+ v.90 Receive +-------+
Tip | /----|-<----------------<-|----\ | Tip
& >--|-X--B | | B--X-|--< &
Ring | \----|->---------------->-|----/ | Ring
+-------+ v.34 Transmit +-------+

This device would be inserted into the line, and then each of
the lines in the center would be "tapped" with something that
could decode the data.

Note that this is an extreme simplification! The "B" is a
balance network, which must approximately match the 2-wire Tip
and Ring line, and which gets approximately 1/2 of the power of
any signal. That means a 3dB loss in each hybrid, minimum
(actual figures would be more like 3.5 to 4.5 dB). Hence there
would almost certainly have to be an amplifier added to each
side of the 4-wire interface (making this a "repeater"!).

And the echo cancellation issue would have to be dealt with for
the monitoring devices. That could be done either by adding
echo cancellation to the amplifiers (probably a good idea, given
that this is going to require at least 7 dB of gain, and might
cause a very noticable degradation of the voice quality on the
phone line), or it could be only applied to the bridged
connection going to the monitoring device.

Obviously this is not an extremely simple device, but then again
it isn't rocket science nor does it actually employ anything
unusual. All of these parts discussed so far, with the single
exception of the amplifiers, are included in virtually every
v.90 modem!

Hence this could be done with something that amounts to stuffing
about 4 times as much circuitry as a regular modem has into one
box.

That is correct. As you note, there has to be "echo
cancellation", or actually a separation of the transmit and
receive signals. Once past that, the only thing is that it
would require a fairly high quality tape recorder, since most of
them have far more phase noise (which our ears are almost immune
to) than a modem signal can tolerate.

That is just about the only part of this which is not "off the
shelf". The "modem" portion used to monitor the data needs to
be able to work without being able to communicate to the distant
end. It just has to deal with whatever is being sent and cannot
request changes. Obviously normal modems don't work that way,
but building one that does is not difficult, it's just
expensive... ;-)

--
FloydL. Davidson <http://web.newsguy.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) floyd@barrow.com