Redundancy in Network Design: Avoiding a Single Point of Failure

A single core switch that every other switch connects through, a single internet connection with no backup, or a single gateway that every sensor depends on are all common in networks that were built incrementally rather than designed. Each one works fine daily, right up until it fails and takes down everything depending on it at once.

A backup internet connection (often a 4G/5G failover) keeps a site's dashboards and alerts reachable even if the primary line goes down. Redundant switching at critical points means one switch failing doesn't isolate an entire section of a building's devices. These additions have a real cost, which is exactly why redundancy is worth deciding deliberately — for the parts of a system where downtime is genuinely costly, not applied uniformly everywhere.

Not every network needs full redundancy at every point — a small home network can usually accept an occasional outage without real consequence. A hospital, a factory production line, or a facility with life-safety systems is a different calculation entirely, where a network outage has real operational or safety consequences, and redundancy budget should be prioritised there first.