Various degrees of high availability can be achieved with Session Initiation Protocol (SIP) trunking. As with any failover solution, a higher the level of resiliency will come at an increased cost. Following are some different scenarios along with their corresponding high availability mitigation:

• Mediation Server failure.To protect against a Mediation Server failure, first install a second Mediation Server alongside the existing Mediation Server. Next, add this Mediation Server to any voice routes containing the existing Mediation Server. For outbound calls, Office Communications Server automatically load balances between the two Mediation Servers and route calls failover to just one of the Mediation Servers in case the other server goes down. Therefore, the service provider must be configured to accept SIP trunk calls originating from two source Internet Protocol (IP) addresses.
  
  Handling failover for incoming calls is done in an analogous fashion. The service provider should be configured with both gateway Internet Protocol (IP) addresses corresponding to the two Mediation Servers. The service provider should load balance between these servers by sending some calls to one server and some to the other. More importantly, if one server is no longer responding, the service provider must take that Mediation Server out of the pool and only send incoming calls to the active Mediation Servers.
  
  
• Connection failure.To protect against a connection failure, use a second connection for the redundant Mediation Server instance. This second connection ensures that if one of the connections goes down, only one Mediation Server will be affected. The service provider must be able to detect this connection failure for incoming calls and send calls to the other connection.