I've been reading up on conference bridges and the such and the folks on the support forums always provide the best information so, once again, I am need of clarification. I have so many questions regarding this so I'll try and organize my thoughts - the more I read on the topic of conferencing, the more questions crop up. I apologize if my thought process gets jumbled so here goes......
I am not intially going to be able to provide specific configs as I still do not have complete access to this network; however, a problem has cropped up that I'm trying to understand (and fix!). A DSP router has been configured to provide transcoding and conferencing (I understand this to be HW resources). The router is 2951 with 3 PVDM-256's in it. It is running 15.2 code. It has been configured for conferencing - 64 max participants. Because of the max participants (sessions, I think is what it's called on the router), it will only accept G.711 codec. I confirmed this on cisco's website and, of course, we tried to add G.729 and the command was rejected due to the max participants.
Transcoding has also been configured on this device. I don't know what the parameters are - I am trying to do this from memory of when I saw a coworker whizzing through the configs so.....
A MRGL has been configured specifically for this 64 participant conference and a phone has been assigned to this MRGL.
When someone from outside the cluster calls in, they call in on G.729 and that's where the problem arises - it appears transcoding is not doing its job and the call does not come through. If we take the phone out of that MRGL and try again, it works. I have confirmed that this conference works for everyone except one group of people in another location. My intial theory is that, for some reason, the transcoding piece is not working. I tested this by having a phone on another cluster (coming in across the WAN) call in to my conference ( I modified my phone to be in the MRGL for this big arsed conf bridge). I initiated the conference, had him call in and then I webbed to my phone to see what codec we talked. We talked G.711. I assume this other cluster has g.711 configured to go out across the WAN, hence we work. I don't know how to, if it's even possible, to make a single number negotiate a specific codec. If it is, I would like to change the number from the other cluster to force it to choose G.729 and then dial in again to see what happens. Is this possible?
One of the things I've read is that, on the call manager, there are clusterwide conference service parameters that can be adjusted - specifically the Maximum Ad Hoc Conference number. One of my questions is: is the number in this field specifically for SW conference resources or does it also apply to HW conference resources?
I've read this great document that aokanlawan posted on PVDM2 and PVDM 3 credit/allocation; it was so informative for someone like me who doesn't have much experience with this sort of thing. Again, though, reading it raised questions. He mentions that you should configure voice termination resources first then conferencing, transcoding, etc, resources second. My question here is: if we aren't using it for any kind of voice termination, I assume I don't need to configure voice termination resources?
The document alludes to the fact that the dsp resources shared - meaning, if I use up all the resources on one DSP, it automagically "rolls over" to the next DSP resource but I can't really find this anywhere else. Is this specific to the PVDM3-256 only?
Again, thank you guys sooo much for any kind of clarification you can provide. I apologize for not being able to provide more in depth information regarding the specific configurations as well.
You can see the utilization/reservation of the DSP resource using below command:
show voice dsp group all
Regarding your question,
1. I would like to change the number from the other cluster to force it to choose G.729 and then dial in again to see what happens. Is this possible?
[Ans] You can configure a new region which communicates only G729 with all other regions. Then create a device pool, assign that new region to this new device pool. then assign this device pool to that device (from which you are making call).
2. One of my questions is: is the number in this field specifically for SW conference resources or does it also apply to HW conference resources?
[Ans] This is for both HW and SW resource.
3. if I use up all the resources on one DSP, it automagically "rolls over" to the next DSP resource but I can't really find this anywhere else. Is this specific to the PVDM3-256 only?
[Ans] No, this applies to all DSPs.
During a call, you can see what dsp resources are being used by using below command:
show sccp connection
Hope this helps you in further understanding and fixing the issue.
Piyush and Chris,
Thank you very much for the information! I believe I now have some terra firma footing to stand on at this point and will proceed with checking out the good tips you guys have provided.
The reason you cannot add codecs to the conference bridges is in fact because you have max sessions and max participants defined, but having G711 only in conference bridge is actually a very good idea as most participants in most scenarios are local and if you need other participants then transcoders are used for that. Why that does not work for you is probably due to misoncifugration on CUCM side, the MRGL/MRG with the transcoders needs to be assigned to the device that needs to engage the transcoder such as the phone and/or gateway.