Routing Principles in SIP Cluster

[In progress]

The following sample deployment diagram illustrates the routing principles in SIP Cluster:

In the above diagram, there are two data centers (DC1 and DC2) with two SIP Cluster Nodes each. All SIP Cluster Nodes share the same switch (SIP Cluster Switch) where all Routing Point DNs and agent DNs are configured. Each SIP Cluster Node has a dedicated routing solution consisting of URS, ORS, and Stat Server. The same strategy is loaded on all SIP Cluster Nodes to enable distribution of inbound calls across the nodes and process the calls according to the strategy.

In the SIP Cluster architecture, Virtual Queue (VQ) DNs are used for reporting, monitoring, and for agent reservation. To simplify provisioning and improve performance, the SIP Cluster architecture includes a switch (VQ Switch) dedicated to VQ DNs. This enables unloading the VQ DN traffic that might be generated in complex large contact center deployments. A single HA pair of SIP Server instances running in non-cluster mode is assigned to that VQ Switch in each data center. The routing Stat Server instance is connected to all VQ SIP Severs, but each URS and ORS instance is connected only to the VQ SIP Server located in the local data center.

For configuration details, see Configuring Switch and DN Objects for SIP Cluster.

ORS Deployment

• A dedicated one-node ORS Cluster is deployed to serve each SIP Cluster node.
• A dedicated Cassandra ring is deployed for each ORS Cluster.
  • cassandra-keyspace-name: Orchestration
  • cassandra-nodes: FQDN_SIP_Cluster_Primary_VM; FQDN_SIP_Cluster_Backup_VM

To create a high performance solution, the number of ORS pairs per SIP Cluster Node can be increased.