Power Redundancy

Power is the foundation everything else depends on. Power redundancy means no single failure takes systems offline. It's implemented at multiple levels:

Dual power supplies — enterprise servers, switches, and storage systems typically have two power supply units (PSUs). Each PSU connects to a different circuit or PDU. If one supply or its circuit fails, the other carries the full load. This is the most important single level of redundancy for any critical device.

Dual PDUs — power distribution units in a rack are typically deployed in pairs (PDU A and PDU B), fed from separate circuits or separate UPS units. PSU A of each server connects to PDU A; PSU B connects to PDU B.

UPS (Uninterruptible Power Supply) — provides battery backup to bridge the gap during a power event, and in some cases provides power conditioning. A UPS protects against brief outages and allows for controlled shutdown if the outage extends. UPS runtime is typically minutes to tens of minutes — enough to ride through a brief utility event or execute a clean shutdown, not hours of operation.

Generator — for extended outages, a generator provides sustained power after the UPS battery is exhausted. Typical transfer time from utility to generator is 10–30 seconds, which the UPS bridges. Generator fuel capacity and testing frequency are critical operational concerns.

Diverse utility feeds — for the highest-availability facilities, two physically separate utility feeds from different substations enter the building on different paths. Expensive and uncommon outside of colocation facilities and critical infrastructure.

For most organizations, dual PSUs on critical servers, dual PDUs per rack, and a properly sized UPS covering the network closet and server room is the practical target.