Cisco Cisco Agent Desktop 8.5 Referências técnicas

CAD 7.0 uses the “warm standby” approach to fault tolerance and autorecovery. No 
manual intervention is required to recover a failed service.

Data and features may be lost at the time of the failure. For instance:

■

Active monitoring and recording is stopped. They can be restarted manually 
after the failover.

■

Enterprise data for the call in progress is lost at the time of the failure.

All CAD features are fault-tolerant to a single point of failure with several exceptions. 
They are:

■

Playback. Recordings are tied to a specific service, and thus are not 
replicated.

■

SPAN-based monitoring and recording. Desktop monitoring can be used for 
CAD agents if fault tolerance is required. SPAN-based monitoring can be 
used as a backup to desktop monitoring.

CAD uses LDAP replication to provide fault tolerance for configuration information, 
such as work flows, agent hot seat settings, and so on. It uses MSDE merge 
replication to provide fault tolerance for Recording & Statistics service-related data, 
such as call logs, agent state logs, recording logs, and so on.

A subset of the base services fail over together. These services will either all be active 
or all be inactive on the same box:

■

Chat service

■

Enterprise service

■

LRM service

■

Recording & Statistics service

■

IP Phone Agent service

The LRM service controls the failover logic for this subset. Two failures of the same 
service within five minutes causes failover of the subset. One LRM failure causes 
failover of the subset.

The service autorecovery feature enables Agent Desktop and Supervisor Desktop to 
automatically recover their connections to the Cisco Desktop services in the case of a 
service restart or a network outage.