Cisco Cisco IP Contact Center Release 4.6.2 Leaflet

In most cases, Setup and the PG Explorer will allow you to configure machines that do not meet these 
restrictions. However, when you start up the PGs, only two will actually start.

Unified Expert Advisor implements an active/standby high availability model for its runtime servers and 
a buffering high availability model for its reporting servers.

Unified Expert Advisor runtime servers operate in an active/standby fashion. Typically the primary 
server will be active and the backup server (also called the high availability server) will be in standby 
mode. Under normal conditions, all calls are handled by the active server. The only thing the standby 
server does is keep track of expert advisor login and presence states, but it does that only in order to 
avoid having to gather the data all at once if a failover occurs.

The rules for setting up duplexed PGs and Unified Expert Advisor runtime machines, as depicted in 

, are as follows:

PG A should connect to the primary runtime server.

PG B should connect to the backup runtime server.

There may be no cross-connections between PG A and the backup runtime server, or vice versa.

This is because the primary and backup servers are envisioned to be located at physically different sites, 
across a WAN from each other. The WAN link between PG sides is guaranteed (via Unified ICM design 
guidelines) to be high bandwidth, low latency in nature, but there is no such guarantee for the connection 
between a PG and its peripheral. Also, transmitting data across a WAN link could lead to security 
concerns that are not a problem with LAN links.

PG A

PG B

Unified Expert
Advisor
Primary

Unified CCE

187481

Unified Expert
Advisor
Backup