Network Uptime Calculator

The calculator models series availability as A^n, since every component in a non-redundant chain must be up simultaneously for the path to work — if any single link fails, the whole path is down. Even a 99.9%-available component compounds downward as you chain more of them: five in series at 99.9% each works out to roughly 99.5% for the full path, not 99.9%.

Once you specify more than one redundant path, the calculator switches from the series model to a parallel model: the whole path is only down when every parallel path is down at the same time, so the probability of total failure is (1 − A_series) raised to the number of paths, and availability is one minus that. This is why even a mediocre series availability can jump close to 100% with two or three genuinely independent redundant paths.

It's a shorthand for how many consecutive 9s appear in your availability percentage — 99% is 'two nines,' 99.99% is 'four nines' — and the calculator derives it as −log10(1 − availability), so it scales continuously rather than jumping only at round percentages. A higher nines value corresponds to exponentially less downtime per year.

The math assumes every component's failure is statistically independent, but the explainer flags that redundant paths sharing a power feed, a single fiber conduit, or the same firmware bug aren't actually independent — a single event can take out what looks like redundancy on paper. Treat the calculator's output as a best-case theoretical ceiling, not the availability you'll actually observe.