Balanced Call Processing
You can balance calls between the servers in the CallManager cluster through the
use of CallManager groups and device pools, as defined in the previous section.
By allocating different devices to different pools, and then assigning these pools
to different CallManager groups, you have a very flexible method of achieving
the load balancing you require.
It is recommended there be some common characteristic shared by each
device in a pool, such as physical location, IP subnet, or device type.
In Figure 4.4, we can see two device groups, and two server groups.Within
our server group we have three servers; Server A and B are dedicated call processing
servers, and Server C is configured as a combined database publisher and
TFTP server, but also has call processing capabilities. Server group 1 defines
Server A as the primary server, Server B as the secondary server, and Server C as
the tertiary server. Server group 2 defines Server B as the primary server, Server
A as the secondary server, and Server C as the tertiary server. Device group 1
uses Server group 2, and Device group 2 uses Server group 1.This means that
under normal circumstances call processing would be evenly balanced between
Server A and B. Server C would only be used for call processing functions if both
Server A and B became unavailable.