[asterisk-users] Ast/Hyla/IAX Scalability?

[Marshall Henderson]

On Fri, Mar 13, 2009 at 11:07 AM, David Backeberg <dbackeberg at gmail.com> wrote:
>
> On Fri, Mar 13, 2009 at 11:38 AM, Marshall Henderson
> <marshallmch at gmail.com> wrote:
> > I recently read the thread entitled "Faxing Success Rate on PRI" which dealt
> > with Asterisk/HylaFax/IAXmodem. I'm successfully using this 'recipe' in a
> > few instances on systems with only a couple of analog lines all the way up
> > to a full PRI worth of Iaxmodems.
>
> Then you have probably seen that YMMD, and that some people claim
> great success with VoIP fax.
>
> Other people claim that the only way to go is a hardware fax solution,
> like the dedicated multi-modem fax cards.
>
> The only way you're going to find a solution that will work for you is
> to try it, scale it, build your own expertise with your solution, load
> test it, and watch your error rate.
>

I certainly understand the value of building the solution, testing,
patching,  and fixing problems as they arise. It was my hope however
that others would have large-scale experience with these technologies
and could share some pointers.

I'm about to perform some bulk testing between two servers to see how
the system reacts. I'm more than happy to post my findings here if
anyone has interest.

> The other consideration is your budget and your cost of dropping a
> fax. The faxmodem cards are not cheap compared to a voip solution. But
> if the faxes have a high value to the business the hardware cards are
> probably justified.
>
> Again, you'll find people arguing that their voip solution has as low
> of a failure rate as a hardware solution. I'm jealous. My voip fax
> solution does not yet have that low of a failure rate, but I'm
> hopefully getting closer to working out the last bugs.
>

Do you have any specifics to share about the problems you're finding?

> > I've also noticed that IAXmodem is compiled statically against a version of
> > spandsp included with the iaxmodem source. For a large installation, would
> > it be better to compile iaxmodem dynamically to reduce the per-instance size
> > of each iaxmodem? Or, would it be better to simply throw more RAM at it?
>
> I'm not sure what difference RAM makes. What breaks a fax on voip is
> latency and dropped packets.

Agreed. I was simply inquiring about the efficiency of IAXmodem at the
system resource level. Latency and packet drops will be minimal or
nonexistent at all in this environment.

>
> You solve both of those problems if you go the hardware solution route
> with a faxmodem card.
>

I've found hardware fax boards aren't a 100% fix either. Many of the
boards are buggy. However, I will have to say that certain
manufacturers like Mainpine are near 100%.

> The in-between solution is using a proprietary telco -> fax gateway,
> like a Cisco box that terminates a PRI and provides FXO ports that you
> plug into a single-pair faxmodem or a 'real' fax machine. That
> solution quickly becomes ridiculous when you try to scale it.
>
> > Are there any concurrency issues when receiving a large number of faxes on a
> > system using IAXmodems?
>
> File system contention, but fax files aren't very large, and I would
> call that a non-issue. Most people don't want a piece of paper; they
> want a PDF that they can turn into paper once in a while.
>

The purpose of such a system as I'm inquiring about is for digital
archival. Very little 'paper' will be in use. Buffering aside, each
fax could be written at the speed at which it is received correct? So,
if I'm receiving 50 faxes at 14.4kbps each, assuming a direct receive
frame-->block write, I'd be looking at roughly 90KBps written to disk.
Is my logic sound here?

Thank you for the response and ideas.

Marshall