Power Resiliency — Flashcards

Blackout: voltage drops to zero — complete loss of power.

Brownout: voltage drops below normal operating level — partial reduction, not a full outage; causes instability and hardware stress.

Surge (spike): voltage spikes above normal — sudden overvoltage that can permanently damage electronic components.

A battery-containing device that provides short-term backup power when utility power fails, drops, or spikes. Protects against blackouts, brownouts, and surges. Also bridges the generator startup gap (~60–90 seconds) by providing battery power while the generator warms up. Three types: offline/standby (basic), line-interactive (voltage regulation), and online/double-conversion (zero transfer time).

Simplest, cheapest UPS type. Equipment runs from utility power during normal operation; battery waits in standby. When utility fails, an internal switch transfers load to battery — with a brief millisecond delay. Provides basic blackout protection and surge suppression. Does NOT actively regulate voltage — brownouts pass through if above the switch-over threshold. Best for PCs and non-critical equipment in stable power environments.

UPS type that adds active voltage regulation via an autotransformer. Boosts voltage during brownouts and reduces it during overvoltage — without switching to battery. Battery is preserved for actual blackouts. Best suited for areas with frequent voltage sag or instability. Has a transfer delay when switching to battery but uses it far less than offline UPS. Commonly deployed for small servers, network closets, and mid-sized business environments.

Most advanced UPS type. Continuously converts AC→DC (charging battery) and DC→AC (powering equipment) — equipment always runs from the battery/inverter, never directly from utility. Zero transfer time: no switching event when utility fails. Provides fully clean, conditioned power. Most expensive; slightly less energy-efficient due to double-conversion losses. Best for data centers, healthcare, and financial systems where even milliseconds of interruption is unacceptable. Also called double-conversion UPS.

Line-interactive UPS. Its autotransformer actively compensates for voltage drops by boosting the output voltage back to normal — without switching to battery. An offline UPS would pass the reduced voltage through to equipment (causing instability). An online UPS would also handle it (but at much higher cost). Line-interactive is the cost-effective middle option that directly addresses brownout scenarios.

When battery charge drops to a configured threshold (e.g., 20%), the UPS signals connected servers to shut down cleanly — saving files, closing transactions, and powering off gracefully. This prevents the data corruption that occurs when the battery is fully depleted and power cuts abruptly. Works unattended — no human needs to be present. Critical for environments where an extended unattended outage could otherwise result in corrupted data when the battery finally runs out.

An engine-driven device that converts fuel (diesel, natural gas, propane) into electrical power for long-term backup operation during extended outages. Provides power for hours to days as long as fuel is available. Requires a 30–90 second startup delay before providing stable output — the UPS bridges this gap. Some building circuits are designated and marked as generator-backed. Requires regular testing and fuel management to remain reliable when needed.

They cover different time horizons and each has a gap the other fills:

Generator alone: equipment loses power during the 60–90 second startup delay — uncontrolled shutdown, potential data corruption.

UPS alone: battery runs out during an extended outage — graceful shutdown but no sustained power.

Together: UPS bridges the startup gap; generator provides sustained power. Seamless coverage across the full outage duration spectrum.

Organizations do not generate their own power — they depend on third-party utility providers they cannot control. A utility transformer failure, storm, grid overload, or maintenance operation can cause an outage with no fault by the organization. Power is the single foundational resource that all technology requires; without it, no redundancy, backup, or failover system can operate. Power resiliency engineering is the recognition that the external dependency must be mitigated internally.