Expanded server capacity for call reliability and redundancy

Call delivery now operates across two geographically separate sites.

Inbound traffic is distributed at the carrier, not by a load balancer in front of a single location, so routing does not depend on any component of the primary site being reachable.

Architecture

• Carrier-level distribution. Inbound routing is held in the carrier's hunt configuration, which selects an available site independently of our infrastructure.
• Automatic endpoint re-registration. Handsets are provisioned with both sites and re-register to whichever is serving traffic, without reprovisioning or manual intervention.
• Independent media paths. Each site terminates its own RTP, so a call in progress is not dependent on a shared media relay.
• Capacity sized ahead of demand across both sites rather than at a single point.

Failover is automatic and requires no action from an administrator or end user.