Cisco Unified Communications Manager is the call processing entity of the Unified Communications infrastructure and is responsible for call setup and teardown.  In addition to this base level of processes it is also responsible for many feature related services such as basic presence, conferencing, music on hold, and single number reach.

Call setup and teardown refers to the process of one IP phone calling another and the phone checking with the Unified Communications Manager (UCM) to ensure that the other party exists and is available.  Once these basic things have happened the next part of the process is for the first phone to figure out how to reach the other phone.  This is handled by basic routing and switching  but also now coupled with the UCM and its dial plan.  Once the two phones have determined that they are aware and reachable they begin to communicate independently of the UCM and have a point to point IP stream between the two respective endpoints.  This can be seen to be quite efficient with respect to bandwidth utilization.   Other systems that require an IP stream between the IP PBX and the respective endpoints at all times but this is not as efficient from a bandwidth perspective.

Once the conversation is complete, the two endpoints will communicate with the UCM to inform the system that they have completed the conversation and are now available.  Additionally, this is where the call detail record is written to the system. This enables users to report system usage by extension(aka - Call Detail Reporting).

In essence the Communications Manager is the "go-to guy" for all peripheral applications.  It is basically the hub for all the Unified Communication Applications and information.  It is obviously very important that this infrastructure be a reliable environment since pretty much everything is running based on the information located in the Communications Manager database.

There are several ways to ensure that this environment is reliable and resilient. For example, you can leverage servers with components that are redundant; such as disks and power supplies.    Another way to provided added resiliency is clustering the Communications Manager.  This provides the ability for the system to have increased capacity as well as redundancy within the server itself.  In this scenario the phones have multiple servers configured, in order of preference, that they "report to" such that if one server fails or goes offline the standby server becomes active.  When a failover happens it is usually seamless to the end users because  there are features which preserve calls in progress.  Only a limited number of the supplemental services, such as conference, become unavailable until the current call is terminated and the phone is registered to the standby server.  Once the primary server for the respective phone comes back online the phone re-registers to the UCM.

These varying levels of redundancy coupled with the operating system and maturity of the Communications Manager software make this solution a very redundant and resilient offering.

For more information on Cisco Unified Communications Manager go to the following website;

http://www.cisco.com/en/US/products/sw/voicesw/ps556/index.html