Re: [hylafax-users] faxing over sip

[Robert Moskowitz <rgm@xxxxxxxxxxxxxxx>]

A J Stiles wrote:
> Is there a protocol where you take the raw zeros and ones that would have been 
> sent to the fax machine's modem, bundle them up into packets, send the 
> resultant packets over a packet-switched network and recreate the image "at 
> leisure"  (i.e., not relying on the carrier frequency or bit rate for timing; 
> if necessary, we can buffer it and play it back at a fixed rate)  from the 
> reassembled packets at the far end?  Because it seems to me that this would 
> be the best way to get fax working properly with packet-switching.  If SIP 
> can handle other types of packet beside audio, so much the better.
Simple answer is a resounding NO!

I worked with Hayes and company on V.32 and V.32bis, and hung around for 
V.34 (Hayes had already bet the farm on ISDN and lost to USR with the 
AT&T breakup. Dennis always said that if the breakup had been delayed 
one year we would have had real ISDN in the US, as EU has).

The biggest, single impediment is that Modeming is half duplex and 
everything is designed around the timings of the directions switching. 
The biggest 'user' of this is HDLC (ever hear of X.25? Oh were those fun 
days. And X.75 gateways. Not.) I did a lot of HDLC deployments. Then IBM 
got into the act with their 'special' 'synchronized' protocols....

IP, by nature is asynchronous, and really messes up when you try to get 
a half-duplex protocol working over it. TCP is a BIG help for this, but 
it masks all the timing problems in the network and the apps just can't 
figure out what is wrong, if timing is everything. So you fall back to 
UDP and try and build your own. This is what ITU did with T.38. Voice, 
and even Video really don't care so much. But everything done to make 
Fax work really does care.

So as Lee said, there is a lot buried inside of T.38 to make the 
endpoints think there is a working HDLC circuit there. If I remember my 
G4 days right, even that uses the timing of ISDN; that you do not find 
in IP.

To have fax work right for the world of IP would take a real revolution 
in fax. And with ITU controlling things, this will not happen. Why do I 
say this with conviction? I worked with Dave Crocker on T.37 as he was 
our (IETF's) liason to ITU. And as much as Dave (the original author of 
SMTP and that icon of email: MCImail) is a standards politician par 
excellant (he can be really a lot of fun at the IETF plenary sessions), 
the ITU people still fell flat with T.37 and went on their own for T.38.

Give it 5 years. You will start seeing the printer/fax vendors take 
things into their own hands (as they did with IPP), then we will only 
start to see systems designed well. Hey, gee, that IPP, that almost 
sounds like a great place for them to start with this! :) (oh was that 
an effort to fix up what they brought to IETF originally, "but we are 
already shipping printers working that way"....).

:)

A lot of history here, and my view of it. But I TAUGHT data 
communications back in '76. It was a teachable, one semester, SUBJECT 
back then....

I also had the privilege of using one of A/J's first accoustic modems. 
It was a plywood box about 4"x12"x14" with accoustic padding that you 
placed the headset into and closed the top over. :)



____________________ HylaFAX(tm) Users Mailing List _______________________
 To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
On UNIX: mail -s unsubscribe hylafax-users-request@xxxxxxxxxxx < /dev/null
 *To learn about commercial HylaFAX(tm) support, mail sales@xxxxxxxxx*