Highly available
High availability is the ability of a server system to
provide a service to a user with as little perceived downtime as possible.
Though not guaranteed, different ranges of high availability exist,
from five nines
with 99.999% uptime, to 95% uptime.
If a system component fails for any reason, the high availability solution ensures that another component takes over for the failed component or that another server with the same component is available to handle the client request.
This documentation addresses only the hardware and software issues. Other issues that might impact high availability at a customer's system include: human error, process problems, and external issues.
Restriction: High availability does not mean that a service
is guaranteed to always be available. Analysts describe a range of
high availability targets, from the so-called five nines availability,
with 99.999% uptime, at the high end, to basic availability at 95%.
Five nines availability translates to five minutes or less downtime
in a full year of 24 by 7 operations. By contrast, 99% availability
allows up to 87 hours of downtime per year, and 95% allows up to 436
hours, or 18 days, of downtime, with the cost of providing high availability increasing
exponentially as the target moves from 95% to 99% to 99.999%.