gmane.comp.telephony.yate

T.38 Transcoding

Philipp Hoffmann — 2013-05-17T20:12:45

Hi,

I am trying to realize T.38 transcoding with Yate:

A  (SIP)--> Yate (SIP) --> Carrier (SIP) --> B (ISDN)

A supports T.38 (but no g711), the Carrier does *not* support T.38. That means, that Yate needs to transcode T.38 to g711.

How can I configure this case in Yate?

We use YATE 2.2.0-1 (default debian package).

Regards,
Philipp.

Re: German ISUP support

Diana Cionoiu — 2013-05-16T07:15:31

Hello Philipp,

DT Labs has certified SS7 implementation of Yate.

Diana

On 05/15/2013 11:32 AM, Philipp Hoffmann wrote:

German ISUP support

Philipp Hoffmann — 2013-05-15T18:32:42

Hi,

is there currently any SS7 stack with German ISUP dialect available for Yate?

Thanks,
Philipp.

Re: Perform radius authentication lookups

Paul Chitescu — 2013-05-14T09:43:45

Hi!

The Teles-UserDefined attributes are from a private dictionary and not decoded 
by Yate. Right now loading a dictionary is not supported, the RADIUS, Cisco, 
Microsoft and Quintum dictionaries are coded in tables at top of yradius.cpp

Also note that there's a level of indirection caused by passing the parameters 
through the authentication message. This matters for failed authorization.

If you want to return from a successful auth attempt:
  ret:h323-credit-amount=credit_amount
The attribute "h323-credit-amount" will be copied to parameter "credit_amount"

If you want to return from a failed auth attempt:
  ret-fail:Reply-Message=authfail_reject_text
The attribute "Reply-Message" will go to call.route parameter "reject_text"

Paul


On Tuesday 14 May 2013 12:03:42 pm Philipp Hoffmann wrote:

Re: Maximum number of extensions in Yate

Paul Chitescu — 2013-05-13T16:21:48

Hi!

Consider creating a tree of routes to limit the linear search length in each 
section.

You may do like this (based on 1st digit of number):

[default]
^[0-9]=goto start_\0
.*=echo Not starting with a digit?

[start_0]
...

[start_1]
...

...

This will reduce the complexity 10 times (if numbers are reasonably 
distributed).

Paul


On Saturday 04 May 2013 02:01:30 am ZZ Wave wrote:

Re: Perform radius authentication lookups

Paul Chitescu — 2013-05-13T16:17:01

Hi!

Generic binary parameters are currently not supported.

The User-Password was not implemented because no VoIP protocol performs 
plaintext password authentication.

You may create a proper username by generic message manipulation. A simple 
example using regexroute.conf:

[extra]
user.auth=50

[user.auth]
${called}^+\?\([0-9]\+\)$=;rad_user=\1


yradius.conf:

add:User-Name=${rad_user}


Paul


On Sunday 12 May 2013 12:06:44 am Philipp Hoffmann wrote:

Re: Problem with Callfork and RTP/DTMF in progressing state

Paul Chitescu — 2013-05-13T16:08:21

Hi!

What version of Yate do you run?

By default the DTMFs should be forked upstream - at least starting with Rev. 
4680 (2011-11-04). In Yate 4+ this should work.

Some changes in the SIP DTMF logic were made in Rev 5277 (2012-09-20)

Note that there are limitations on how a DTMF can be sent upstream in a non-
answered call. This may prevent sending them with fork or without.

The limitations are:
- SIP INFO needs the dialog to be established (so it doesn't work while 
ringing)
- RFC 2833 needs the upstream audio to be established and have some data (so 
it has clock source)
- inband needs the upstream audio to be established and a transcoder to the 
negotiated codec (and it may fail because of the codec)

You should raise the debug level of ysipchan to 8 or higher and check if you 
see messages like:

<sip/22:NOTE> Failed to send tones '6' methods=info,inband

You may see messages as the above only for some legs of a forked call. In the 
case above it was a Cisco phone than sends an 180 without SDP so no

Perform radius authentication lookups

Philipp Hoffmann — 2013-05-11T21:06:44

Hi,

I am trying to perform radius authentication requests using yradius.conf:

The radius server expect two standard attributes, which have to been set up by Yate: The "User-Name" (always the called party number, but without the first char, e. g. +49310 -> 49310) and the "User-Password" (always clear-text: "radius").

For reference:

--
[nas]
add:User-Name=${caller} ;;;;; question: how can we cut the first char (+)? ;;;;;
add:User-Password=radius
--

The query goes through the radius server, but Yate do not provide the "User-Password" attribute.

Trace:

--
<yradius:NOTE> Using sections [nas] and [radius common] for authentication
<yradius:GOON> Ignoring unknown attribute of type 1
--

In the source code, the "User-Password" attribute is definied as binary, while "User-Name" is defined as string. Is it possible, to set a binary (string) attribute in the yradius.conf file?

Thanks,
Philipp.

Re: Re: Yate as MGCP Media Gateway

Diana Cionoiu — 2013-05-10T17:57:29

Hello,

The MGCP gateway in Yate is not full because it wasn't ment to be used. 
It was ment just for testing.
However we will be interested to develop more on that. Please drop me an 
e-mail at diana at null dot ro if you are interested in buying such a 
gateway.

Regards,
Diana

On 04/26/2013 03:01 AM, Eugene Prokopiev wrote:

SS7 Localy blocked ciruit

Herman Bigos — 2013-05-09T17:16:12

Hello.
My config is YATE 4.3.1+ AS5400.

I've got problem with localy blocked circuits:
Circuits are blocked by Yate itselfs from unknown reson.


status sig isup1/97
%%+status:sig isup1/97
module=sig,trunk=isup1,type=ss7-isup;circuit=97,span=mg70+1,status=Disabled,lockedlocal=true,lockedremote=false,changing=false,flags=0x1
%%-status

control isup1/ISUP unblock circuit=97
Could not control isup1/ISUP unblock circuit=97

control isup1/ISUP unblock circuit=97 force=yes
Control 'isup1/ISUP' OK

--------------------------------------------- but circuit is still
blocked----------------------------------

status sig isup1/97
%%+status:sig isup1/97
module=sig,trunk=isup1,type=ss7-isup;circuit=97,span=mg70+1,status=Disabled,lockedlocal=true,lockedremote=false,changing=false,flags=0x1
%%-status



---------------------------------------------------------------------------------------------------------------------

So, anybody knows whats up ?

Regards
Hermozol

dtmf debug

2013-05-06T11:45:00

Hello

I have Cisco 5400 with SLT as PSTN to SIP GW. I have problems with  
DTMF that is received from PSTN site. When someone call my NGN number  
and press and hold dtmf tone for longer time, I receive it double.  
There is no problem for GN.
Only difference I see in IAM is that when someone call NGN ther is:

PropagationDelayCounter='126'
HopCounter='30'

when someone call GN number i see:

PropagationDelayCounter='0'
ParameterCompatInformation.PropagationDelayCounter='transit,nopass-param'
ParameterCompatInformation='31 c0'

Is there a way to check what goes as DTMF from PSTN site ? Or can  
PropagationDelayCounter cause this behaviour ?

Greetings
Andrzej

Re: IVR with javascript

Damian Wolgast — 2013-05-04T17:53:32

Hi Mahdo,

to get dtmf use a javascript handler and register it to receive dtmf.messages.
in javascript.conf add line
<dtmf name>=script
like
mydtmf=/tmp/dtmf.js

dtmf.js is like

function doSomething(message)
{
 // your logic
Engine.debug(message.text);
}
Message.install(doSomething, "chan.dtmf", 20);


this outputs the pressed number in yate debug output.


play file: Channel.callTo("dbwave/play//path/to/file") or with callJust
you cannot use that in a dtmf handler, you need to use that in your main js routing script

your ivr logic of course is if/else etc...

Regards,
Damian

Am 04.05.2013 um 06:05 schrieb Mehdi Shirazi:

Re: IVR with javascript

Mehdi Shirazi — 2013-05-04T04:05:09

Hi Domian
Thanks for reply.

Please guide me How we can have DTMF interaction in javascript IVR? 
for announcement or multi_level IVR menu we still need chan.masquerade 
( example ?) or functions like "callTo","playFile"... do the job for us? ( I remember your question
in forum http://forum.yate.ro/index.php?topic=117.0 ) still not clear
 for me, in javascript we should forget and don't worry about this 
warnings(in php samples)? 
-/* This is extremely important.We MUST let messages return, handled 
or not */
-/* If Yate disconnected us then exit cleanly */
...


Regards
M.shirazi


--- On Fri, 5/3/13, Damian Wolgast <damian.wolgast-Z55j7XUYj0LZaZF71Vwp9w< at >public.gmane.org> wrote:

From: Damian Wolgast <damian.wolgast-Z55j7XUYj0LZaZF71Vwp9w< at >public.gmane.org>
Subject: Re: [yate] IVR with javascript
To: "Mehdi Shirazi" <mahdi_shirazi-/E1597aS9LQAvxtiuMwx3w< at >public.gmane.org>
Cc: yate-uHKunLg9Q/3XMkR9fcqaOA< at >public.gmane.org
Date: Friday, May 3, 2013, 12:22 PM

Hi,
this clearly describes what to do:
http://docs.ya

Re: Maximum number of extensions in Yate

Ionut Muntean — 2013-05-03T23:31:57

Are we talking about the maximum number of extensions (routes) or the 
maximum number of concurrent calls? The extensions (routes) are defined 
in regexroute or anything else like a database, the maximum concurrent 
calls is something else.


On 5/4/13 2:01 AM, ZZ Wave wrote:

Re: Maximum number of extensions in Yate

ZZ Wave — 2013-05-03T23:01:30

SIP only, G.711A, G.729

Yate works in "normal" mode, with full signaling and RTP proxy

regexroute is about 4k lines


2013/5/2 Diana Cionoiu <diana-liste-uHKunLg9Q/3XMkR9fcqaOA< at >public.gmane.org>

Re: IVR with javascript

Damian Wolgast — 2013-05-03T16:22:36

Hi,

this clearly describes what to do:

http://docs.yate.ro/wiki/How_to_do_routing_using_javascript


in short:
- enable javascript module
- create your routing script (example is provided)
- save script where yate can access it
- open javascript.conf and point to the script in section 'general' on routing=<path to script>

Regards,
Damian

Am 03.05.2013 um 14:12 schrieb Mehdi Shirazi:

IVR with javascript

Mehdi Shirazi — 2013-05-03T12:12:42

Hi
Is it possible that someone kindly provides 
 sample IVR in javascript?

Thank you

RTP Packet Size in yate?

Bipin Patel — 2013-05-02T14:49:36

Re: Maximum number of extensions in Yate

Ionut Muntean — 2013-05-01T23:19:25

Yate, goes, with RTP and codec transcoding as it is needed, up to more 
than 670 calls per machine (max load for the CPU's). The machine has to 
be at least 8 core (threaded or not), at least 3.4 Ghz. SO flavor, disc 
and memory (4Gb mem is trivial) do not matter. No H323 -> SIP ... only SIP.

A custom build "cluster" with a cheap machine acting as a load balancer 
(Yate) that receives the SIP calls, balance them (signaling only, no 
RTP) to a pack of 6 other cheap machines (also Yate) that do the routing 
from Postgres, CDR build and handle the RTP. The total load for this 
"cluster" is 4k+ concurrent calls. All done with extmodule&php scripts. 
Of course, the Postgres is on another server that can handle the 
requests in real time.

It's all on the processor speed and thread capabilyties of the machine 
Yate runs on (more threads possible, more calls). This is done for a 
wholesale company (no users, no register, only calls routed from the DB).

Hope this helps.

On 5/2/13 1:45 AM, Diana Cionoiu wrote:

Re: Maximum number of extensions in Yate

Diana Cionoiu — 2013-05-01T22:45:58

Hello ZZ,

Which protocol? What about routing?

Diana

On 04/27/2013 06:16 AM, ZZ Wave wrote:

Re: Send MAP messages

Chris Hölzel — 2013-04-30T10:36:18

Hi there,

unfortunately this doesn't belong to yate.

It is an own application.
It is definitly not opensource and you have to buy a aculab board to use
this software.
We would not sell our product or provide any kind of source code. This
system is only for internal usage.

Sorry.

Chris

On 03.04.2013 10:34, Akib Sayyed wrote: