Are you sure about that? Next time you finish an outgoing call, don't
hang up and see what happens (I'm assuming that your outgoing call is
going out the analog line, and not a VoIP circuit or anything). I'm
pretty sure you'll find that it takes about 11 seconds before your
system hangs up.
Jim
Nabeel Jafferali wrote:
Thanks for the info Jim.
Can you answer one basic question that's confusing me though - why do I get
far-end disconnect supervision fine on an outgoing call, but not on an incoming
call?
--
Nabeel Jafferali
X2 Networks
-----Original Message-----
From: Jim Van Meggelen [mailto:[email protected]]
Sent: February-09-09 11:49 PM
To: Nabeel Jafferali
Cc: [email protected]
Subject: Re: [on-asterisk] Audiocodes - Disconnect Supervision
Waaaay back in the early 90s when voicemail was not yet ubiquitous,
far-end disconnect supervision for loop start analog lines was something
that didn't really exist. They were lines, not trunks (the CRTC acually
tariffed PBXs and Key Systems differently, as well as the kind of
circuits you could connect to them). It was never a problem before
voicemail, because as long as your phone system hung up the line (which
is what would happen when you hung up your phone), the circuit went
on-hook at your end, and the call was terminated correctly.
If you wanted far-end disconnect supervision, you had to get a
ground-start trunk. These were analog circuits that didn't use a loop to
request dial tone, but instead used a ground (we had an extra wire in
our butt sets so we could ground the line to the frame to get dial tone
when we were testing.
As voicemail systems became more common, far-end disconnect supervision
on small key systems became a problem. Lines would appear to be hung,
but in reality what was happening was that the voicemail simply didn't
have the smarts to realize the far end was disconnected, and so it
wouldn't hang up the call.
The solution was a hack. Carriers began to implement disconnect schemes
on analogue lines. They would send a signal of some sort to indicate to
the system that the far end had disconnected. This was never perfect,
and still isn't (although it's pretty darn good).
So the problem is not only related to Asterisk. Any voicemail system on
a loop-start analog system has to take into account the facts that a)
the disconnect will not come immediately, and b) sometimes it will not
com at all.
I don't think that there will ever be a way to solve this problem 100%
of the time, but you can get close.
The first thing to set correctly is the silence threshold. Three seconds
is too short, ten seconds too long. What seems to work is about 5 seconds.
maxsilence=5
The next thing to figure out is if we end a recording after 5 seconds of
silence, we will set our minimum message length to 6, which is one
second higher than the silence. This will ensure that any message that
is a result of the caller hanging up and not leaving a message, will
always be 5 seconds (because that's the silence threshold), and thus
will not be a valid message (because it is not the minimum of 6 seconds).
minmessage=6
You will have to tell your clients that any messages less than 6 seconds
may be treated as a hangup by the system. They might not like it, but
I've never seen anybody unable to get over it.
If you really want to play it safe, you should put something in your
dialplan to set the absolute maximum length of a call. That could even
be a few hours if your clients talk on the phone for a while, but at
least that way the call will always disconnect eventually.
The simple fact is, if you want reliable, fast, far-end disconnect
supervision, you cannot use analogue circuits.
Jim
Nabeel Jafferali wrote:
Thanks for the info.
Yes, I do get that, meaning the call does hang up 10-11 seconds after the
remote party hangs up. This also means Asterisk voicemail gives me 10 second
blank messages if the call had ended up in voicemail.
Any way around that?
--
Nabeel Jafferali
X2 Networks
-----Original Message-----
From: Jim Van Meggelen [mailto:[email protected]]
Sent: February-09-09 5:53 PM
To: Nabeel Jafferali
Cc: [email protected]
Subject: Re: [on-asterisk] Audiocodes - Disconnect Supervision
You have to wait for the far-end disconnect from the carrier. If you
have a butt set you can monitor the line and listen for it. I have found
it usually takes about 11 seconds to get the signal.
Nabeel Jafferali wrote:
I have an Audiocodes MP-118FXO in production. When a call is made and the
remote party hangs up, the Audiocodes hangs up the call immediately. But if a
call is received and the remote party hangs up, the Audiocodes does not hang up
immediately.
Anyone experienced this issue with Audiocodes or any other gateway in general?
Any tips?
--
Nabeel Jafferali
X2 Networks
--
--
Jim Van Meggelen
[email protected]
http://www.oreillynet.com/pub/au/2177
"A child is the ultimate startup, and I have three.
This makes me rich."
Guy Kawasaki
--
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
