High availability network built to carry an emergency notification system

The problem

Three campuses depended on an emergency notification system as the primary way of alerting people, in an environment where audible alarms alone cannot be relied on to reach everyone. That makes the network underneath it a life-safety system in practice, not an IT service. The notification system itself runs on a single-wire architecture, so the network had to be designed around keeping that path up.

What I did

• Designed, built, and tested the network from the ground up across all three campuses rather than adapting what was there.
• Built the headend on dual virtually stacked Cisco cores, with a stacked access layer beneath it.
• Ran dual-path redundancy throughout, so no single link or device failure takes the notification path down.
• Deployed high availability firewalls and an SD-WAN architecture across the campuses.
• Tested the design against failure rather than assuming redundancy works because it is configured.

Where it landed

Three campuses on a network engineered to the standard the system deserves: if the notification system needs to alert, the infrastructure carrying it is not the thing that fails.