CVP Redundant Architecture with ICM

[Chad Stachowicz]

Guys,

I have a design question I want to clarify. Lets say I have 2 Call Server/VXML Server. And I have 2 total PG boxes. 1 Call Server and 1 PG on each side of the WAN. The PG boxes should be a side A and side B. They will support both CM Agent PG and TYPE 10 VRU's. Am I correct that Each PG box should have 2 PG instances? 1 PGrunning a single PIM for Call Manager / Soft ACD. and another PG instance having 2 PIM's. 1 PIM pointing to each CVP Call Server. Just want to make sure things are straight in my head! Also I want to verify that with the side A and side B PG's, only 1 PG is active at any one time. So one Call Server will always be being services over the WAN by the PG.

[Chris Deren]

You are correct in all aspects. One things to note is that Cisco recently re-certiied splitting VRUs from PGs accross WAN, the only thing you need to ensure is that you have sufficient bandwidth for this.

From IPCC SRND:

Unified IP IVR or Unified CVP PG to Unified IP IVR or Unified CVP

At this time, no tool exists that specifically addresses communication between the Unified IP IVR or Unified CVP PG and the Unified IP IVR or Unified CVP. However, the tool mentioned in the previous section produces a fairly accurate measurement of bandwidth needed for this communication. Bandwidth consumed between the Unified ICM Central Controller and Unified IP IVR or Unified CVP PG is very similar to the bandwidth consumed between the Unified IP IVR or Unified CVP PG and the Unified IP IVR or Unified CVP.

The VRU Peripheral Gateway to ICM Central Controller Bandwidth Calculator tool is available (with proper login authentication) through the Cisco Steps to Success Portal at

http://tools.cisco.com/s2slv2/viewProcessFlow.do?method=browseStepsPage&modulename=browse&stepKeyId=55|EXT-AS-107287|EXT-AS-107288|EXT-AS-107301&isPreview=null&prevTechID=null&techName=IP%20Communications

If the Unified IP IVR or Unified CVP PGs are split across the WAN, total bandwidth required would be double what the tool reports: once for Unified ICM Central Controller to Unified IP IVR or Unified CVP PG and once for Unified IP IVR or Unified CVP PG to Unified IP IVR or Unified CVP.

HTH, please rate all helpful posts!

Chris

[Chad Stachowicz]

Chris,

Is this bandwidth necessary over the visible or over the private link?

Chad

[Chris Deren]

Public network (visible).

Chris

[Chad Stachowicz]

oh no problem! Thanks for all the help!

Chad

[Chad Stachowicz]

This brought up another good question. So each network VRU has a label. This label points to a particular routing client. Now this is a type 10 VRU so in my script I do a SendtoVRU node to get it to the VXML Gateway. I have a VXML Gateway on each side of the WAN. How do I achieve redundancy in my scripts to reach both VXML Gateways? I came up with using the same incoming called-number on the VXML Gateways, and using DNS coming from the SIP Proxies into the VXML Gateways. Is this supported, am I thinking about this the right way? Also it seems in this sense that I will never be able to always use the correct local VXML Gateway if its up. Am I missing something?

Chad

[geoff]

The NVRU label (on each routing client) returned by the "Send To VRU" node (something like 8111111111) should be configured on each of the Call Servers (allowing for the correlation ID tacked on the end) through OAMP under the "Send to Originator".

That ensures that the VRU leg (and hence the VXML app) runs on the ingress gateway. A typical "branch office" deployment.

This does not go through your DNS (for SRV) or the Proxies.

Regards,

Geoff

[Chad Stachowicz]

Geoff,

I understand that and I have that working. My question is what if VXML Gateway on Side A dies completely. How does ICM know to send it to the VXML Gateway on Side B?

Chad

[geoff]

Chad, are your VXML gateways separate from your ingress gateways?

Regards,

Geoff

[Chad Stachowicz]

Geoff,

Indeed they are. I have 2 3845 Ingress gateways, and 2 5350 VXML Gateways.

Chad

[Chad Stachowicz]

This actually leads to my next questions. So I have a CVP Call Server on each side of the WAN. How do I achieve redundacy here from SIP proxy into the call servers? Also, these are going to be different routing clients in ICM, so the incoming scripts must all have to be mirror's and labels etc to function with both routing clients? Just want to make sure I'm not missings something simple..

Thanks,

Chad

[tapk5]

Exactly, if two CVP in redundancy, you have to duplicate labels against both CVP's.

Wei

[tapk5]

ICM doesn't know VXML GW, it is GK that can point side B VGW if side B dies.

Wei

[tapk5]

One PG instance is enough, you could define it as 'generic PG', so it supports two types of PIM's (CCM and VRU), so this pair of PG's have 3 PIM's (ccm, cvp1, cvp2).

It doesn't harm if using two PG instance.

Wei

[Chad Stachowicz]

Hmm in this case should I load balance out of my sip proxies to the CVP VXML Gateways regardles of where it is in the WAN? Its not possible to achieve perfect call routing from ICM because its always going to try the first label on SendTOVRU which will go into the proxy, and the best you can do at that point is load balance randomely! If there was a way to choose the order that a sendtovru node tried its labels, then it would be achievable. Maybe there is?

Thanks much!

Chad