uptime and downtime
What is uptime and downtime in computing?
In computing, uptime is a measure of how long a computer or service is on and available. Downtime is the measure of how long it is not available. Services measure uptime as a percentage of total time.
What is uptime?
Uptime tells users how long a computer or service has been available. Historically, uptime was mainly a measure of the consecutive time a single computer was powered on before it was rebooted or powered off. This could be shown with the simple uptime command. Some system administrators would take pride in having long uptimes, sometimes measured in years. They might take extreme measures such as avoiding updates or applying patches to a live system.
Modern high availability services no longer rely on a single computer to run. Clustered and balanced servers allow for a server to go down without affecting the entire service. Phased rollouts apply patches and updates to groups of servers at a time instead of all servers to allow for some servers to be available while others are restarted.
Service availability is now a better measure of service reliability than just uptime. This is often expressed as a percentage of available time versus unavailable time.
Even 99% is unacceptable for many services. This would translate to roughly three days of downtime each year. Instead, it is measured in the number of "nines" of availability. Five nines -- or 99.999% availability -- is considered the gold standard. This represents only about five minutes of downtime a year.
How to increase uptime
Strategies to increase uptime include the following:
- Minimize single points of failure.
- Use redundant systems with automatic failover.
- Use phased rollouts.
- Schedule maintenance periods.
What is downtime?
Downtime tells users how long a service is unavailable. Downtime can be planned due to maintenance or unplanned due to an outage. It is impossible to eliminate downtime, but it is important to try to minimize it. Downtime usually represents lost money due to lost revenue, unhappy customers or lost worker productivity.
Planned and unplanned downtime can be covered differently depending on the service-level agreement.
How to decrease downtime
Strategies to decrease downtime include the following:
- Have a business continuity and disaster recovery plan.
- Use monitoring.
- Use change management procedures with rollback plans.
- Test failover systems.
Be prepared for planned and unplanned downtime and consider reliability, availability and serviceability when designing, manufacturing, purchasing and using computer products and services. See why network monitoring tools must offer more insights upfront, learn about device reliability engineering and how it promotes product reliability, and explore the Uptime Institute's data center availability tier standards.