Unity & SRST

[d1nnsear]

UM - Unity 4.0(4)sr1 CCM 4.0.x Exchange 2003

Using the central call processing distributed messaging model to mirror the existing Exchange environment, we aim to deploy SRST at some of our smaller sites (<500 users). CCMs will be accessable across a good MPLS link.

In a worse case scenario the WAN links go down & the phones switch to SRST, I understand the Unity server cannot register with the SRST gateway.

In this case will the Unity server still answer voicemails & store them locally or will it shut down completely until the CCMs are accessible again?

Cheers,

NJ.

[mrmccann · Accepted solution]

Once the WAN link goes down, Unity will soon determine that CallManager is unavailable and will unregister all its ports with CallManager. Unity will remain running, however you will have no means of contacting the system using a phone. As soon as the WAN link comes back up, Unity will determine that CallManager has become available again and will register all of its ports. At this point you're back where you started, everything should work as normal.

[d1nnsear]

That's what I was thinking. Best get those redundant WAN links in!

Cheers for the clarification.

[eschulz]

If this is a UM deployment (or Unity Inbox), subscribers will still be able to use their PC mic/speakers to send/receive messages to/from within whatever portion of the Exchange network that remains intact.

-Eric

[d1nnsear]

Cool, I didn't realise that. Thanks Hans-Eric.

[rkeblusek · Accepted solution]

I have provided redundancy with analog ports. Install a dual integrated Unity IP and analog. Link it to FXS ports on the router and have SRST support those ports. MWI may not work but Unity will be able to run and be accessed in the absence of CallManager.

[Erick Bergquist · Accepted solution]

Getting back to the what will happen with calls when WAN goes down and if Unity will work, etc.

If unity is still registered to Call Manager (both unity and call manager servers at same location, etc) and just a WAN link to a remote site goes down, the only thing down would be that group of devices/phones at that remote site. The call manager, and unity will still be functional in this scenario. The phones would unregister with call manager since the WAN is down, but any forward no answer / busy settings they have will be in effect and if those forward settings went to VM then calls to those numbers while they were in fallback would go into their Unity VM box.

If the inbound calls to the phones at the remote site, come in a gateway at the remote site then those won't use the call manager forward settings since the inbound calls aren't hitting the CCM probably and will be routed based on SRST fallback/H323 settings on router for those numbers.

On the fallback router, it is trickier to get calls to roll to VM but is possible. You would need PSTN access so the call could forward to a VM # but you may need RDNIS support on PRI to maintain the # called so Unity directs call to right VM box. Look around cisco should provide some documents on this, as well on forums. I haven't done this myself but have seen other discussions on the forums doing this.

As the other poster said, they were using FXS ports and analog ports on unity, so there a few ways to go about this.

But it also depends on level of failure that caused the phones to go in fallback. Is the WAN just down or did the central site where call manager is take a hit? Is unity server still available? I have had one client put a home answering machine off a FXS port at a remote site just to take messages in fallback mode because leaving a message if no one answered was important to them. If site is without power this will not work and phones won't be in fallback probably anyway unless a UPS/generator is being used.

[d1nnsear]

Thanks for the posts. That's given me a good few options to think about.

NJ.