Availability
This is the percentage of time that some service of infrastructure is accessible to clients and is operated upon under normal conditions.
If a system has 100% availability, then it means that service functions all the time. In other words, it is never down or unavailable.
How do you measure availability
Availability is a ratio but ratio of what though? Availability calculation is very similar to percentage change calculation.
It is the percentage of ((Total time - the duration of time the service was down) / total time) * 100.
Simple right?
You do have to remember that this measure depends on when the measurement started and whether this is expressed over a short period of time or over several years.
This metric is crucial for cloud providers as it is often part of their SLA and any breaches to the promised/advertised availability, could result in legal action.
Common availability measures
| Availability % | Commonly known as | Downtime per year |
|---|---|---|
| 99.999% | 5 nines | 5.26 minutes |
| 99.99% | 4 nines | 52.56 minutes |
| 99.9% | 3 nines | 8.6 hours |
| 99.5% | 2 nines | 1.83 days |
| 99% | 2 nines | 3.65 days |
Summary
I hope that helps you understand one of the most fundamental non functional requirements and one that is often confused with reliability.
The key to making a choice is what total time means. Different cloud providers might not report a change in availability numbers if the service was not down for all users and most providers exclude any planned downtime and sometimes unplanned downtime is also not included like the ones that arise due to malicious cyber attacks.