• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 2039
  • Last Modified:

CallManager Failover Design

Hi,

We have call manager cluster: One publisher, One subscriber. Also one UCCX server, one voice mail server and ISI server. They all locate at the same place. Right now we are considering a disaster recovery plan for phone system. We have multiple remote sites which have IP phones registered with the main site. Also several contact centers use the UCCX server too.

I was thinking utilize the CME in SRST mode in every gateway at the remote site. But by that way, I need configure every gateway. It seems not good idea. Right now I'm planning to buy another callmanager and UCCX server, Voicemail server at our DR site to do the failover over the WAN. My question is
1. Can I use one callmanager for both publisher and subscriber purpose? I see the normal callmanager cluster is one publisher plus several subscribers. But if the main site publisher
went down, I need another publisher at the DR site. So I need 2 publisher and 2 subscribers in my cluster. If I can use one callmanager for both function at the DR site, then I could save a lot. Any performance degradation?
2. For UCCX server, I'm not sure how UCCX server failover through the WAN. How to design it? Any suggestion?
3. I'm not sure Voicemail and ISI server can do the failover or not. Any idea?

Thanks a lot.
0
madmatrix
Asked:
madmatrix
  • 5
  • 3
  • 2
4 Solutions
 
sr75Commented:
Yes, you can have one call manager, but that depends on how many phones must use it and it must be configured independently and in duplicate to your current configuration.   Same goes for your UCCX server.  The issue I see with that is that if only one site loses connectivity to the CCM Cluster but not to the other sites, you will not be able to make calls to any other sites.   It may be better to go the CME SRST route because you can interconnect your sites if you have the gateways configured using h323 by either manually adding the dial-peers to each sites gateway or by adding a gatekeeper (at the site you were thinking about placing a single CCM at).
0
 
sr75Commented:
I never host voicemail at a singular site.  I always recommend having a CUE at each site.  That way the CUE is available in SRST mode.
0
 
madmatrixAuthor Commented:
Thanks sr75. So you mean CME SRST maybe a better solution to go if only one site lose the connection to the main HUB site. I understand that. But by this way, I need buy CME software for every site gateway, right? How can check I already have CME installed? I only see one file in the flash. I guess I don't have CME installed in every gateway.

By the way, our gateway already have dial peer pots configuration for lots of dial patterns. I think these dial-peers can be used when the gateway turns into CME SRST mode, right?
0
Free Tool: Subnet Calculator

The subnet calculator helps you design networks by taking an IP address and network mask and returning information such as network, broadcast address, and host range.

One of a set of tools we're offering as a way of saying thank you for being a part of the community.

 
madmatrixAuthor Commented:
As for CUE, that is another solution which can be built into the gateway like 3845, right? So that means I need buy CUE for every gateway.

Is there any product which can be used in SRST mode to function as UCCX?

Thanks.
0
 
sr75Commented:
Yes you need to have a licensed CME for each site and the gateway will use those dial-peers when SRST mode kicks in.

As for CUE, it is actually a service module that you purchase and install in the router (basically a Harddrive for VM storage).

I have never used UCCX, so I cannot say for sure.  I think you might be able to get away with configuring a new UCCX server at the DR site and only point the CMEs to them.  But I don't believe there is an SRST type solution for it.
0
 
madmatrixAuthor Commented:
Thanks sr75. That would be a lot of money invested in CME for each gateway. But I think that might be the safest way for failover.

As for CUE and UCCX, we have some IVR scripts configured in UCCX and also we have a cisco desktop supervisor program communicating with UCCX. The contact center supervisor can use this program to check each agent's call history and call duration blabla. The IVR script will guide customer input information and tell customer the position they are in the current call queue. IVR script can keep update that position number when customer is waiting.

I'm not sure CUE can do these functions or not. You should have idea about this. If one site lose the connection to central site, it will lose voicemail function even with CME in SRST mode since it doesn't have CUE installed, right?
0
 
madmatrixAuthor Commented:
By the way, what's the price for CUE normally?
0
 
172pilotSteveCommented:
Unfortunately, I think the answer to your original question is "it depends"...

First of all, in one CCM cluster, you can ONLY EVER have one Publisher.  All writes occur to the publisher.  If your WAN is fast enough, you CAN have remote subscribers, which can operate the phones, but things like changing passwords, forward-all operations, etc won't work while there's no Publisher online.

If your hubsite goes down, your remote sites can use their local ports and dial-peers using CME or SRST (which are the same feature-set these days, it just depends on how you configure the router), but as you point out, you likely wont be able to get to the voicemail..  Assuming it's a WAN failure, and the Unity is still up, you can do some fancy dial-strings and make the voicemail work by redirecting the call back over the PSTN to a Unity access number in the hub office, but if your assumption is that the hubsite is down, then this isn't a good solution..

Each site could have it's own CUE, but then you've got a lot of cost, a configuration pain, and usability problems in that it's tough to send voicemails to multiple offices and maintain user lists, etc.. also, mobility is affected, if you care about that...  I likely can't just go log into a phone in a remote office and get my voicemail.

UCCX works in a LIMITED fashion with CME, but only if it's configured to do so from the install time, so unless you're going to have a UCCX server in each office, that's probably a non-starter too, but in reality, your inbound lines probably come into the hub office (which we're assumed is down) so you're not going to get any calls that way anyway.

I would say a better solution may be to change your fundamental plan...  Most large 7x24 organizations who had multiple sites that I've worked with used completely separate systems (Pub/Sub / UCCX) in each of two data centers.  Regionally, phones were connected to the local CCM, and they were networked either via SIP over the WAN or by PSTN T1.  The company I'm thinking of had call centers in both locations, which handled separate queues, but in the case of a failure, they would call the Telco and call-forward the phone from the downed site to a pre-setup number on the other UCCX, which was already staged with the scripts from the other site.

it worked...  not perfect by any means, but for the money, and functionality, it worked...

-Steve
0
 
madmatrixAuthor Commented:
Thanks Steve to bring up this solution. It really helps. But it could not be suitable for us now. We haven't setup connections from all remote sites to DR site. I'm sure we will get that done in the future. But without that connection at this time, two cluster will not work. So I guess CME in SRST with B-ACD should be the good way to go. Good for both scenario: WAN failure or hub down.

As for voice-mail , if WAN failure, I may use your idea for Voice mail redirect. But for Hub down, we need other solution. You said I could have problem with extension mobility by using CUE. I think EM is configured in CME, right? Why CUE can affect that?

Anyway, the future complete solution should be setting up another subscriber in the same cluster and another  UCCX to do a failover if any of remote sites lost services to hub site either wan failure or hub down. But that will involve lots of things.
0
 
172pilotSteveCommented:
By mobility, I mean that in the localized CUE situation, each mailbox will be configured on one CUE, and you'll have CUEs all over the network, so if someone from office 1 comes to office 2, their voicemail will not be on the local CUE box..  I havent' done much with the Express products, but in the "full" CM and Unity, I'm used to the situation where an employee from one office can go to another office for the day, and log into the phone, and the voicemail light will follow their mailbox, back at the central system, and they'll have FULL access to their VM box..  You'd lose that capability I think...

You may actually find that the best solution is to make the network more resilient so that you WONT have a failure of the wan, and then just have redundant EVERYTHING at the hubsite...

So much to do...  So many choices..  So many trade-offs..  All those choices and tradeoffs are going to affect the user experience, either in that the fully-available configuration will be limited to make it easier to build a fail-over scenario around it, or the user experience will change drastically in the case of a failure...

Some of the problems you'll have will be technical, and some of them will be political...  Balancing all of that against a budget will be interesting to say the least!  :-)  
-Steve
0

Featured Post

Learn to develop an Android App

Want to increase your earning potential in 2018? Pad your resume with app building experience. Learn how with this hands-on course.

  • 5
  • 3
  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now