[Asterisk-Dev] Very Weird problem with MeetMe, SIP, Zap and the combo of the three

[Nir Simionovich - CTO]

Hi All,

  I have a really funky problem, which I can't seem to pin point.I have a 
setup that looks something like this:

SS7 Networks --SS7--> Veraz IGate4000 --SIP--> Asterisk 

Now, Asterisk has a second connection, that looks like this:

Asterisk --PRI--> Avaya CTI

  Now, I'll describe several sceanrios that I'm testing, with some really
Weird results:

Scenario 1:

A. User calls in from the SS7 network, via SIP to Asterisk.
B. Asterisk originates a call to Avaya CTI via PRI
C. Both users are now put into a MeetMe room
D. Avaya user talks, SS7 user hears. SS7 user talks, Avaya heards nothing.

Scenario 2:

A. User calls in from the SS7 network, via SIP to Asterisk
B. Asterisk dials out via the PRI to the Avaya to a specified location.
C. Avaya picks up the call, then generates an outgoing call.
D. The call is picked up.
E. User at the avaya end speaks, SS7 user hears. User at SS7 speaks, avaya
   user doesn't hear.

Scenario 3: 
A. User calls in from the SS7 network, via SIP to Asterisk
B. Asterisk dials out via the SIP connection to a specified location.
C. The call is being picked at the remote end and we have 2 way audio.

Scenario 4:
A. User 1 calls in from the SS7 network, and is put into MeetMe room 1000.
B. User 2 calls in from the SS7 network, and is put into MeetMe room 1000.
C. Users 1 and 2 talk, but can't hear a thing.

Ok, now, I'll go through the conclusions I've reached:

1. There is no network blockage, which is proved by Scenario 3.
2. A test system, working with the same Asterisk version (1.2.0-stable) is 
   working fine in a different location, with a slightly different setup.
3. Scenario 3 negates issues of SIP interoperability between Veraz and 
   Asterisk, on the following tested codecs: g711, IPP-g723.1, IPP-g729 and 
   Digium-g729
4. All MeetMe application entries ares started with: MqAx options.
5. Scenario 4 and the options defined in section 4 indicate the MOH should 
   have been heared, but it wasn't.
6. Now, if I wouldn't have tested scenario 4, I would have said that the PRI 
   is causing issues, however, scenario 4 indicates that something else is 
   wrong. 

So, anyone has an idea of what's going on here? Or better yet, a proposed 
course of Action?

Regards,
  Nir Simionovich