Conference question (CUCM 9.1) ...and one other

voip7372 · ‎10-31-2013

This is new to me, so pardon my ignorance ;-) We're in the early stages of deploying a CUCM (latest version of release 9) and I had a training class that raised a few questions and concerns.

Question 1:

Regarding conferencing, if we only have the CUCM servers (two C240 M3 UCS servers) installed and nothing else (no other hardware - nothing for transcoding), you can use ad-hoc conferencing for quite a few users, as long as ALL of them are using G.711 or G.722? You can NOT add in anyone from the WAN at a branch site to the conference if that call from the branch is coming in as G.729 codec, correct? Even if it’s only a 3 party conference call, you can’t add that 3^rdperson if they’re coming in as G.729? Do I understand that correctly? I understood from the class that if we want to do any type of basic conferencing that have mixed codecs (G.729 and G7.11 for example) we'd have to install a box that does the transcoding, like a router setup just for that.

Question 2:

While I've got your attention, let me know if this makes sense. We purchased two Cisco C240 M3 servers and these will run the virtual machines for CUCM and so on. Assuming I want to split the country in half with a single cluster, would it make sense to have one physical server in the eastern (midwest actually) US with a Publisher virtual machine (VM) and a Subscriber VM...and further west, install the second C240 M3 server and only have a Subscriber VM on that one with Call Manager service running? For the eastern server, the Publisher will be running TFTP service and Subscriber running Call Manager service (so phones will register to that server)...and that Subscriber would be in the 'eastern CM group', for example.

The physical server further west with the Subscriber VM running the Call Manager service would be for registering/serving the phones in the western half of the US. Maybe we call the CM group for the western phones, 'western CM group'.

Total phones for all servers will be less than 10,000.

Does that make sense? Forgive me if I'm thiinking about it wrong, but like I said, this is new to me and I'm used to working on a multi-site Avaya setup.

Chris Deren · ‎10-31-2013

1. not necessarily, if the conference bridge is in g711 region via its device pool configuration, then you can bridge in remote users as the phones will then negotiate G711, keep in mind the calls traversing the WAN will use G711 bandwidth

2. Sure as long as you meet all latency and bandwidth requirements documented in SRND and other application design documents.

With ~10K phones and WAN cluster you should have at least 2 call processing servers at each data center to provide 1:1 redundancy model. With clusters with more than 1250 phones you need dedicated TFTP as well as documented in SRND. If you are going to rely on CUCM software media resources you should run that against sizing tool as well as those put some heavy stress on the servers.

HTH, please rate all useful posts!

Chris

voip7372 · ‎10-31-2013

Thanks.

Ok....well, that's the thing...we want to keep the WAN calls as G.729 to reduce the bandwidth needed. So, I guess it makes more sense to install a box for transcoding. I guess we can assume this transcoding device (router as I understand it) can be thought of to do the same thing as the Avaya 'MedPro' card (media processor card - provides DSP for converting TDM calls from ISDN to VoIP and vice versa....and I guess it also takes care of any of the codec conversion)?

Thanks for the insight. We still have to work with our vendor to get the exact design settled, but I just wanted to have some ideas before the meetings. As I understand it, the C240 M3 servers are pretty robust and I believe we should be able to install all the CUCM VMs we need to ensure no one VM/server is overloaded. So....possibly 2 Subscriber VMs at the western site and 2 Subscribers at the Eastern site, plus the Publisher and a dedicated CM server running TFTP (at the eastern site)? Sound reasonable?

Chris Deren · ‎10-31-2013

Correct, transcoders would be the way to go, or you can simply use the Cisco router with PVDMs (same that you would use for transcoding) as your conference bridge as those can negotiate all codecs.

The servers are robust and will allow for few VMs depending on which OVA files are used as over-subscription is not supported, so it really comes down to how many CPUs are needed (simple math).

HTH,

Chris

voip7372 · ‎10-31-2013

Thanks again.

Chris Deren · ‎10-31-2013

No problem, please mark the thread as answered if you have no further questions on this topic.

Chris

Sent from Cisco Technical Support iPhone App

voip7372 · ‎10-31-2013

Thanks. I didn't mark the post as an official question, so the option to select it as answered isn't there.

Chris Deren · ‎10-31-2013

Interesting, I would imagine since you posted the question the option should be there. In any case thanks and good luck.

HTH, please rate all useful posts!

Chris

Rob Huffman · ‎10-31-2013

Hey Chris,

You will also notice that you can't VIP endorse anything in a thread

where the OP did not mark their original query as an actual question.

I think once anyone responds it's too late to go back and change it

Cheers!

Rob

"Why do the best things always disappear "

- The Band

Chris Deren · ‎10-31-2013

Thanks Rob, I did not realize you can post un-official question, I suppose I don't post too many questions :-)

Chris

Rob Huffman · ‎10-31-2013

You are always welcome my friend

"it's the old question is not a question, question"

Cheers!

Rob

"Why do the best things always disappear "

- The Band