Posted January 3, 2012 by Spyros in General Tips & Tricks

The Expectations of Uptime in the Real World


Server uptime is one of those “real-world” expectations that is often misunderstood by webmasters and web designers when selecting the right host for their professional web creations. The term uptime is interchangeable with reliability, and it is an important statistic that web hosting providers carefully monitor. Hosting providers that have servers with record uptimes often instruct their marketing and sales personnel to tout those long-running statistics to prospective clients. For techies involved in managing server farms, a lengthy server uptime record is like a badge of honor.

What Uptime Really Means

Uptime can be defined as the uninterrupted time elapsed since any machine or device began operating. Any interruptions in a machine’s operation, including reboots, qualify as downtime. Information technology (IT) professionals may obsess over uptime, but several devices outside of IT have posted impressive uptime records. The ATS-3 satellite, launched by NASA in 1967, was continuously operational until 2001. That’s 34 uptime years. Home refrigerators are known to run for years without being turned off, save for power outages, and nuclear plants have similarly amazing uptimes.

Server computers are regularly subject to downtime caused by issues such as operating system (OS) upgrades, patches, power outages, disaster prevention measures, etc. For these and other reasons, servers that post long-running uptime statistics never fail to impress IT professionals.

Record Uptime Statistics

In 2006, IT industry giant asked its clients to send the company proof of their Netware server uptimes. One client submitted a screen capture of a server running on a 486 that was last rebooted in 2000. The client apologized for not having implemented a patch that fixed a memory leak which required an annual reboot, otherwise the server uptime would have been closer to 9 years.

Uptime statistics are reported by the OS, but this measurement has proven unreliable at least once. In 2009, a Linux server user detected a bug which returned uptime results of nearly 40 years for some Linux OS versions. Still, many Linux servers running on different hardware platforms have reported uptimes of 2000 days.

Apple’s Mac OS X may not be an industry favorite when it comes to server machines, but a website that compiles Mac OS X uptime stats has been tracking a server in Sweden with an uptime longer than 8 years.

Uptime versus Availability

Sever uptime statistics can give webmasters looking for a new place to park their websites a pretty good idea about the host’s level of commitment to its clients. Uptime records can also reveal patterns of administration employed by hosting providers to maintain their servers.

When evaluating different hosting providers, webmasters will invariably hear about 99 percent or higher uptimes. This figure sounds reasonable until actual time is taken into account. A downtime measurement of 9 percent means that a server was not operational for about 87 hours in a year. Assuming that a webmaster needs to see his or her sites up and running around the clock, 365 days a year, those 87 hours could means a lot of lost business.

Availability, the capacity of a server to handle all requests, is a more accurate representation of the dependability webmasters should look for. Besides availability and uptime. what other factors are vital when looking for the ideal web hosting provider?

This is a guest post. Mila Johnson is an internet entrepreneur who uses GoDaddy discount codes to save money on her domains.