Agent Availability for Routing

Handling how DNs become available for routing is unique to SIP Cluster.

In the SIP Cluster architecture, URS is connected to the default port of the SIP Server and is only aware of a call on that particular SIP Server. The routing Stat Server is connected to the T-Controller port of the SIP Server through which all calls in the SIP Cluster are passed and, as a result, the Stat Server is aware of all calls on different SIP Servers in the SIP Cluster.

This section discusses DN states and DN ownership based on the type of agent phones: SIP phones or PSTN phones.

DN State for SIP Phone Agents

When the phones are not registered, the initial state of their DNs is out of service (OOS).

DNs are not registered

The following sample diagram illustrates the flow when one of the SIP phones registers:

1. SIP phone 1 registers by sending a SIP REGISTER request to data center 1 (DC 1).
2. SIP Server 1 becomes an owner of DN1.
3. DN1 is set to in-service state (InSrv) on SIP Server 1.
4. DN state is passed to the local routing Stat Server through the T-Controller port (TCport).
5. DN state is passed to the remote SIP Server 2 through the T-Controller layer.
6. SIP Server 2 distributes the DN state to the Stat Server 2 through the TCport.

DN State for PSTN Agents

DN ownership for PSTN agents is defined by the T-Library registration (TRegisterAddress) sent from the Workspace Web Edition (WWE) application in to which an agent logs in. The SIP Cluster Node received the registration becomes the owner of that DN.

If WWE is not connected, Genesys Web Services (GWS) does not register agent DNs and their states are out of service (OOS).

When WWE1 connects to GWS:

1. Agent 1 logs in to WWE and GWS registers on DN1.
2. SIP Server 1 becomes an owner of DN1.
3. DN1 is set to in-service state (InSrv) on SIP Server 1.
4. The DN state is distributed the same way as for the SIP phone agents.

Other Scenarios

The following sample diagram illustrates how an inbound call is routed to an agent when the cluster nodes are connected and both the DNs are registered:

• Both DNs are in service.
• Global Routing: Call can be routed from any node to any agent.

The following sample diagram illustrates the flow when the connection between the two SIP Cluster nodes is lost:

• The SIP Cluster nodes are not connected through the T-Controller layer.
• As soon as connection is lost, both SIP servers understand that the owner of the DN is not connected any more and events for that particular DN cannot be received.
• SIP Cluster node moves DN to OOS state, if connection to DN owner is lost.
• Local Routing: each SIP Cluster node can continue routing to the locally owned DNs.