Cisco Cisco Customer Voice Portal 8.0(1) Design Guide

The configuration discussed in this section protects against all of these situations. However, the 
following two situations cannot be protected against:

Someone stops the process with calls in progress. This situation occurs when a system administrator 
forgets to put the Call Server out-of-service first to allow calls in progress to finish before stopping 
the process.

The Call Server exceeds the recommended call rate. Although there is a throttle for the absolute 
number of calls allowed in the Call Server, there is no throttle for call rate. In general, exceeding the 
recommended calls per second (cps) for an extended period of time can cause erratic and 
unpredictable call behavior on certain components of the CVP solution if one of the components is 
not sized correctly or if the call load is not balanced according to the weight and sizing of each call 
processing component. (Refer to 

 for call server call rate details.)

For call survivability, configure the originating gateways as described in the latest version of the 
Configuration and Administration Guide for Cisco Unified Customer Voice Portal (CVP), available at 

The survivability.tcl script itself also contains some directions and useful information.

In the event of most downstream failures (including a Call Server failure), the call is default-routed by 
the originating gateway. Note that survivability is not applicable in the Unified CVP Standalone and 
NIC-routing models because there is no Unified CVP H.323 or SIP Service involved anywhere in those 
models.

There is also a mechanism for detection of calls that have been cleared without Unified CVP’s 
knowledge:

Unified CVP checks every 2 minutes for inbound calls that have a duration older than a configured 
time (the default is 120 minutes).

For those calls, Unified CVP sends an UPDATE message. If the message receives a rejection or is 
undeliverable, then the call is cleared and the license released.

The CVP SIP Service can also add the Session Expires header on calls so that endpoints such as the 
originating gateway may perform session refreshing on their own. RFC 4028 (Session Timers in the 
Session Initiation Protocol) has more details on the usage of Session Expires with SIP calls.

Calls are handled as indicated for the following scenarios:

Calls in progress

If the Unified CVP SIP Service fails after the caller has been transferred (transfers include transfer 
to an IP phone, VoiceXML gateway, or other egress gateway), then the call continues normally until 
a subsequent transfer activity (if applicable) is required from the Unified CVP SIP Service. If the 
caller has not hung up and is awaiting further activity, there is a period of 9 to 18 seconds of silence 
before the caller is default-routed by survivability to an alternate location.

If the call has not yet been transferred, the caller hears 9 to 18 seconds of silence before being 
default-routed by survivability to an alternate location. (Survivability does not apply in NIC-routing 
models.)