CallManager Clustering
CallManager clustering is a method of seamlessly distributing call processing
throughout a converged IP network. By using clustering, several CallManager
servers can share the burden of call processing, which becomes particularly
important in larger or widely distributed IP Telephony implementations.
A cluster is defined as a set of Cisco CallManager servers sharing the same
database and resources.The roles described in Table 4.1 can be assigned to members
of a CallManager cluster.
The AVVID solution provides options that allow scaling and load-distribution for
both IP telephony and voice/video conferencing features. Clustering is a technique
used to enhance both the capabilities of the network as well as the redundancy
within the network.With the networking capabilities available today, it is
possible to cluster both voice networks and video networks. Clustering techniques
allow you to scale your networks and, when the need arises, to add more
users or services.
Cisco CallManager clusters can be configured to distribute call processing
and device registration between multiple servers on the same segment. Up to
eight servers can be part of a cluster, with common database information and
real-time device registration data being replicated throughout using intra-cluster
communications.With common information shared throughout a number of
servers, redundancy is achieved. If a single server fails, another server can transparently
take over call processing for a group of devices. For very large or multisite
installations, several clusters may be used, with information being shared via intercluster
communication. However, you should use this solution with care, as many
Cisco CallManager features will not function between clusters.
Voice and video conferencing is facilitated by terminals producing
voice/video data streams, and by Multipoint Control Units, which control the
conference. For voice or video conferences larger than those supported by a
single MCU, Cisco offers a feature known as cascading. Cascading allows you to
cascade two or more MCUs, in order to provide a single larger conference. In
addition to providing highly scalable conferences, this also provides load distribution
between multiple MCUs, and allows voice and video streams to be localized
by the use of MCUs on different segments.