Maintenance Manual

Chapter 2 Fault Tolerance
Architecture
2-2
Cisco ICM Enterprise Edition Administrator Guide Release 6.0(0)
Architecture
The architecture of ICM software allows the system to continue to function if one
component fails. This ability is called fault tolerance. To ensure that ICM
software continues to operate in the case of a computer failure, all critical parts of
the system can be physically duplicated. There can be two or more physical
Network Interface Controllers (NICs), two physical Peripheral Gateways (PGs) at
each call center, and two Central Controllers. The communication paths between
critical components can also be duplicated.
The critical components of ICM software include the Central Controller
(CallRouter and Logger), PGs, and NICs. Normal Admin Workstations (AWs) are
not considered to be critical to the operation of the system since they play no
active role in routing calls or storing historical data.
When both instances of a component are available to the system, that component
is said to be duplexed; when only one of the pair is available, the component is
running simplexed. You might have some components in your ICM system that
are duplexed and others that are simplexed. For example, you might have a
duplexed Central Controller (two CallRouters and two Loggers) and simplexed
Peripheral Gateways at call center sites.
It takes more than duplicate hardware to achieve fault tolerance. The ICM system
can quickly detect that a component has failed, bypass that component, and use
its duplicate instead. ICM software can also initiate diagnostics and service so that
the failed component can be fixed or replaced and the system returned to duplexed
operation.
Approaches to Fault Tolerance
ICM software uses two approaches to fault tolerance: hot standby and
synchronized execution. In the hot standby approach, one set of processes is
called the primary, and the other is called the backup. In this model, the primary
process performs the work at hand while the backup process is idle. In the event
of a primary process failure, the backup process is activated and takes over.
Peripheral Gateways optionally use the hot standby approach to fault tolerance.
ICM software uses synchronized execution in the Central Controller. In the
synchronized execution approach, all critical processes (CallRouter, Logger, and
Database Manager) are duplicated on separate computers. There is no concept of