Cisco Cisco Customer Voice Portal Downloads

Page of 108
High Availability
Method
Scalability Method
Path
first, and fails over to the
backup if necessary,
Server and one backup, or
if CSS is used, an array of
remaining there for the
Media Servers can be
used.
life of the call.
Subsequent calls try the
primary first again. Note
that most files fetched
from the Media Server
remain cached on the
VXML Gateway. This
algorithm is not in effect
if CSS is used.
If request fails, routing
script may select a
Unified ICME routing
script identifies a
Request from VXML Gateway to VoiceXML Server
different server and try
VoiceXML Server by
again. There is no
name or IP address. If
automatic failover to a
CSS is used, a farm of
backup VoiceXML Server
without a CSS.
VoiceXML Servers can
be used.
N/A
Pre-established
relationship.
HTTP Call Result from VXML Gateway to IVR Service
N/A
Pre-established
relationship.
Micro-application Result from IVR Service to VRU PG and
Unified ICME
N/A
Pre-established
relationship.
Agent label from Unified ICME and VRU PG to IVR Service
N/A
Pre-established
relationship.
Agent label from IVR Service to H.323 Service
If the Gatekeeper goes
down, its HSRP partner
Request goes to the
Gatekeeper, which returns
Directory query from H.323 Service to Gatekeeper
begins handling requests
the IP addresses of one or
at the same IP address. If
more endpoints
the entire pair is down,
represented by the agent
label.
the H.323 Service takes
itself out of service and
will not receive new calls.
Alternatively, H.323
Service now supports
Alternate Gatekeepers.
You can configure
multiple gatekeepers; if
the H.323 Service’s
primary one stops
responding, the H.323
Service will make the
next one primary and use
Planning Guide for Cisco Unified Customer Voice Portal 4.0(1)
66
Chapter 6: - Creating a Failover Strategy (Designing for High Availability)
High Availability H.323-Based Call Flow