Cisco Cisco Unified Customer Voice Portal 10.5(1)

Creating a Failover Strategy (Designing for High
Availability)

Unified CVP is designed with high availability in mind. Most Unified CVP call flow
components—those components that carry call traffic—are designed for n+1 redundancy. If
any single component fails, there is no caller-observable impact to the handling of new incoming
calls, except possibly a prorated reduction in total call capacity. However, calls that are in
progress using the impacted component may be affected, though they will usually be caught by
the Survivability script on the ingress gateway and given default treatment or transferred to a
default target rather than dropped. The Unified CVP solution components that support n+1
redundancy are: Call Server, VoiceXML Server, Ingress Gateway, VXML Gateway, Media
Server, and ASR/TTS Servers. However, Media Server and ASR/TTS Servers can only be
duplexed if a Content Services Switch is not used.

A second set of Unified CVP solution components—SIP Proxy Server, Gatekeeper, Content
Services Switch—allow for redundant pair (duplex) configuration.

Connections to the Reporting Server are resilient: information is buffered at the source in case
of a connection failure, and then transmitted to the Server when it later becomes available.

The Operations Console does not support redundancy because the ability to access and update
configurations is not usually real-time critical. Similarly, VoiceXML Studio is an offline tool
for application script development, which does not require high availability.

Note: The material included in this chapter is intended to compliment, not to replace, the high
availability information provided in the Cisco Unified Customer Voice Portal (CVP) Release
4.0 Solution Reference Network Design (SRND) document. You are urged to consult that
document as well.

This section contains the following topics:

•

•

•

Planning Guide for Cisco Unified Customer Voice Portal 4.0(1)