Resilience Metrics & Definitions

Clear definitions - availability, reliability, endurance, and ride-through - keep the resilience claims precise and honest.

Say what you mean

Resilience is a loose word until it is defined by measurable quantities. This page fixes the terms used across the section so that every claim can be checked against a number, and so that no property is oversold by vagueness.

The core metrics

• Availability: fraction of time producing power (single Aegis unit 0.86–0.995)
• Reliability: probability of no failure over a period (MTBF-driven)
• System availability: the redundant set's uptime, higher than a single unit
• Endurance: how long the plant runs without external resupply (years, fuel-limited)
• Ride-through: the duration of disturbance the plant survives without dropping load

How they relate

Availability and reliability describe a single unit; system availability is what redundancy builds from them; endurance is set by on-site fuel and maintenance; ride-through is set by storage. A resilient installation reports all four, because a high value in one cannot hide a low value in another — high endurance is meaningless if system availability is poor, and vice versa.

The reference points

The section's honest anchor is the availability gate: a single unit at 0.86–0.995 against a hyperscale Tier III target of 0.99982. Every resilience metric is reported relative to that gap, and the redundancy architecture is what closes it at the system level.

All the metrics here are physical and reliability quantities, and the discipline is to report them together, since a single strong number quoted alone can hide a weakness that only appears when availability, endurance, and ride-through are read side by side. Reporting all four together also prevents a common failure of resilience marketing — quoting one strong number in isolation — because a claim is only meaningful when availability, reliability, endurance, and ride-through are stated side by side.