Cisco Cisco IP Contact Center Release 4.6.1 Leaflet

Unified CM call detail reporting (CDR) shows the call as terminated when the Unified CM failure 
occurred, although the call may have continued for several minutes after the failure because calls in 
progress stay in progress. IP phones of agents not on calls at the time of failure will quickly register with 
the backup Unified CM subscriber. The IP phone of an agent on a call at the time of failure will not 
register with the backup Unified CM subscriber until after the agent completes the current call. If MGCP, 
H.323, or SIP gateways are used, then the calls in progress survive, but further call control functions 
(hold, retrieve, transfer, conference, and so on) are not possible.

Unified CCE will also write a call record to the Termination Call Detail (TCD) table because 
Unified CM has reported the call as terminated to the Unified CCE PG. If the call continues after the PG 
has failed-over, a second TCD record will be written as a "new call" not related to the original call.

When the active Unified CM subscriber fails, the PG receives out-of-service events from Unified CM 
and logs out the agents. To continue receiving calls, the agents must wait for their phones to re-register 
with a backup Unified CM subscriber, then log back into their Unified CCE desktop application to have 
its functionality restored. Upon recovery of the primary Unified CM subscriber, the agent phones 
re-register to their original subscriber to return the cluster to the normal state, with phones and devices 
properly balanced across multiple active subscribers.

In summary, the Unified CM call processing service is separate from the CTI Manager service, which 
connects to the Unified CM PG via JTAPI. The Unified CM call processing service is responsible for 
registering the IP phones, and its failure does not affect the Unified CM PGs. From a Cisco Unified CCE 
perspective, the PG does not go off-line because the Unified CM server running CTI Manager remains 
operational. Therefore, the PG does not need to fail-over.

When a CTI Manager service fails, the Unified IP IVR JTAPI subsystem shuts down and restarts by 
trying to connect to the secondary CTI Manager service on a backup Unified CM subscriber in the 
cluster. In addition, all voice calls at this Unified IP IVR are dropped. If there is an available secondary 
CTI Manager service on a backup subscriber, the Unified IP IVR logs into this CTI Manager service on 
that subscriber and re-registers all the CTI ports associated with the Unified IP IVR JTAPI user. After 
all the Unified CM devices are successfully registered with the Unified IP IVR JTAPI user, the server 
resumes its Voice Response Unit (VRU) functions and handles new calls. This action does not impact 
the Unified CVP because it does not depend upon the Unified CM CTI Manager service for call control.

Unified IP IVR Release 3.5 provided for cold standby and Release 4.0 provides hot standby redundancy, 
but this configuration is not supported for use with Unified CCE. These designs make use of a redundant 
server that is not used unless there is a failure of the primary Unified IP IVR server. However, during 
this failover processing, all calls that are in queue or treatment are dropped on the Unified IP IVR as part 
of the failover. A more resilient design would be to deploy a second (or more) Unified IP IVR server(s) 
and have them all active, allowing the Unified CCE to load-balance calls across them automatically. As 
shown in 

, if one of the Unified IP IVR servers should fail, only the calls on that server would 

fail, but the other active servers would remain active and be able to accept new calls in the system. 

The Unified ICM is a collection of services and processes running on Unified ICM servers. The failover 
and recovery process for each of these services is unique and requires carefully examination to 
understand the impact to other parts of the Unified CCE solution, including another Unified ICM 
service.