In the world of data centers, the terms, “availability” and “reliability” are often, erroneously, used interchangeably. While both describe a center’s performance and are related, they mean two distinctly different things.
Scientifically, reliability refers to something repeatable that, when done over and over, ends up with the same results. This repeatable reliability is measured by either what is known as meant time between failure (MTBF) or failure rate. The MTBF is the total time in service divided by the number of failures. Failure rate represents the number of failures, divided by the total time in service. Reliability also is often coupled with the phrase, “validity,” which reflects how true or accurate a measurement is to actual.
Availability measures how much something is in an operable or usable state. It is the amount of uptime divided by the total time measured. While something might be available, it might not be reliable. And even if it is reliable, it might not be valid.
A good example of this is an aging air conditioner in a computer room. It might be running consistently for years, thus it could be considered highly available. However, it might be deficient in providing good cooling conditions, so it also can be considered low on the reliability measurement. And if it’s controller isn’t properly calibrated, the tracking results could be invalid.
Overall, measuring a data center’s true reliability depends on the facility’s overall operational goals and expectations. But, in general, a reliable data center provides continuous operations as long as it is operated properly and within its design intent and limitations. The goal of most data centers is sustained operations of IT equipment — 100 percent computer room availability. To achieve this, it’s crucial to have both reliability and validity. The operating processes must be repeatable in achieving the expected outcome, and the outcome must correspond to the desired result.
Bear in mind that different factors can affect this perfect scenario, including a center’s physical infrastructure and its operating staff, among other things.
Click here to learn more about how data center availability, reliability hinge on numerous factors.