The deployment consists of one Cisco UCM Cluster and one Cisco CMS Cluster distributed accross three sites, US, UK and GE. In this type of deployment, there are some recommendations that should be taken into considerations
The first recommendation is to reduces the number of distribution links required to connect the participants in the conference, and therefore reduces the load across the CMS node and distributed calls accross the WAN. To achieve this requirement, we use the CallBridge Group feature. We create Three CallBridge Groups for each site, and we associate the local to each CallBridge Group.
The second recommendation is to optimize the outgoing calls from the Cisco UCM Cluster.
A SIP Trunk must be configured to each CallBridge.
Three Route Groups for each site must be created to group the SIP Trunk to the remote CallBridges located in each site. For example the Route Group RG-UK will group the CallBridges of the Site United Kingdom UK (UK-CMS1 and UK-CMS2).
Three Device Pools and three UCM Groups must be created for each site.
Each UCM Group must be associated to the corresponding Device Pool, for example, the UCM Group RG-UK will be associated to the Device Pool DP-UK.
Each endpoint will use the Device Pool to register to the local Call Controls.
All CallBridges share the same space and a single Route Pattern must be created with a single Route List. The Calls received by the Call Control must be routed to the local CallBridges. To achieve this, we need to use the Local Route Group with a single Route Pattern and Route List. The decision to route the calls to the local CallBridges is done through the Device Pool where the Local Route Group for each site is associated.
For the CallBridge Group to work, There are two requirements:
The first requirement: We need to enable the "Accept Replaces Header" on all SIP Trunk so that the call control is able to move SIP calls between Call Bridge Groups as instructed by the CallBridges when sending the SIP INVITE with Replaces Header.
The second requirement: we need to assign the appropriate Rerouting CSS on all SIP Trunk, this Rerouting CSS must include the Line Partition of the endpoint so that the Call Control is able to send a SIP UPDATE to the endpoint when moving the call to another CallBridge hosting the meeting.
Finally to optimize the outgoing calls from the Call Controls's perspective, we need to enable the "Run On All Active Unified CM Nodes" option on the Route List and the SIP Trunks, this option ensures that the Call Control where the endpoint is registered to is always used to route the egress calls.
