[asterisk-users] Faxing through Zap cards

Karl Fife asterisk-users at kfife.mailworks.org
Wed Sep 3 14:11:06 CDT 2008


On Tue, 2 Sep 2008 11:38:17 -0500, "James Sneeringer"
<jsneerin at gmail.com> said:
> On Sun, Aug 31, 2008 at 1:45 AM, C F <shmaltz at gmail.com> wrote:
> > No, in the beginning you asked because you don't have the experience
> > so folks like myself that do have the experience answered. It might
> > work for you, no one knows and you THINK it will work, it's a hit and
> > miss, stability is huge issue, thats where experience comes in. If you
> > want something that I or the other people here just think works, then
> > just get an ATA. If you want something we have experienced and know
> > that it works, then get a channel bank.

I'd like to draw on your experience.  At one point you mentioned that
the
fax stability goes from perfect to "anybody's guess" when the call
leaves the PRI card.  I think I understand the underlying architecture
well enough to know why this is the case.  Here's the question:  

In an installation where there are only Analog Telco drops, can
pri/channel bank reliability be achieved on analog cards by keeping fax
traffic *within* a single Digium TDM card *because* of the fact that
card would not be subject to the limitations of the PCI/PCX interface
bus and/or underlying OS?  For example 4 analog fax lines into (and out
of) a single TDM800--4 telco lines to 4 FXO, 4 fax machines from 4 FXS). 

Do you have any practical or theoretical knowledge as to whether similar
reliability to the PRI/Channel-bank setup can be achieved PROVIDED that
traffic is never allowed to leave the internals of the card.  Depending
on how ZAP services the card, there may be exactly ZERO difference
between the aforementioned setup and one involving multiple SEPARATE
cards.  If traffic stays within the card, where (if anywhere) does the
process becomes compromised?

Certainly it would be trivial to design a card that could handle fax
pass-through, so the logical conclusion seems to be that NOT having done
so was done to achieve a GREATER good in a mutually exclusive design
trade-off.  I'm sure that I (and others) would be very interested to
gain a better understanding of this if you (or anyone) can speak
intelligently to it. 

Thanks

-Karl 




More information about the asterisk-users mailing list