Arsenal / Resilience planning

Availability & DR Calculator

“Five nines” sounds great until you see how little downtime it allows.

Convert uptime targets into real minutes, turn MTBF and MTTR into availability, see how backup frequency sets your RPO, and find out what redundancy actually buys you.

Network+ N10-009
Security+ SY0-701 / SY0-801
Runs in your browser

01 / The nines

Uptime % → allowed downtime

Availability target (%)

Period	Allowed downtime

Basis: 365-day year (8,760 hours); month = one twelfth of a year (about 30.4 days); week = 7 days.

Reverse it: downtime budget → availability

Downtime

Unit

Per

02 / Reliability metrics

MTBF + MTTR → availability

MTBF (hours)

MTTR (hours)

Availability = MTBF ÷ (MTBF + MTTR). The system is up for MTBF hours on average, then down for MTTR hours while it is repaired.

Know the terms

MTBF	Mean time between failures. Average run time between failures of a repairable system. Higher is better.
MTTF	Mean time to failure. Average life of something you replace rather than repair (a fan, a drive). Higher is better.
MTTR	Mean time to repair (or recover). Average time from failure back to service. Lower is better.
MTD	Maximum tolerable downtime. The longest the business can survive an outage; RTO must fit inside it.

03 / Disaster recovery

RTO vs RPO scenario picker

RPO · looks backward

Recovery point objective: how much data (measured in time) you can afford to lose. It is set by how often you back up or replicate.

RTO · looks forward

Recovery time objective: how long you can take to get the service running again. It is set by your recovery method and site.

Scenario Backup / replication schedule Failure strikes this long after the last good backup: Recovery method

Times are example planning values for practice, not vendor guarantees.

Business RPO target (hours)

Business RTO target (hours)

04 / Redundancy math

Serial vs parallel (redundant) components

Serial (in a chain): every part must work. Multiply the availabilities: A = A₁ × A₂ × … The chain is always weaker than its weakest link.

Parallel (redundant): only one copy must work. Multiply the chances of failure, then subtract from 1: A = 1 − (1 − A₁)(1 − A₂)…

Each row is a stage in the path. Stages connect in series; set “copies” above 1 to make that stage redundant (parallel).

Stage	Availability %	Copies	Stage result	Remove

Parallel math assumes the copies fail independently and failover is instant. Shared power, a shared switch, or a failover that never gets tested breaks that assumption, which is why real designs remove single points of failure everywhere in the path.