Degrees of Downtime

What exactly does high availability mean?

Every federal agency defines the term differently, depending on the level of risk it can handle. Networks striving for 99.9 percent — or “three nines” — availability will suffer roughly nine hours of downtime annually. The higher the uptime percentage, the shorter the window of downtime: 99.99 percent availability, for example, correlates to about 53 minutes of downtime a year, and 99.999 percent means just five minutes of downtime a year.

What exactly does high availability mean?

Every federal agency defines the term differently, depending on the level of risk it can handle. Networks striving for 99.9 percent — or “three nines” — availability will suffer roughly nine hours of downtime annually. The higher the uptime percentage, the shorter the window of downtime: 99.99 percent availability, for example, correlates to about 53 minutes of downtime a year, and 99.999 percent means just five minutes of downtime a year.

So why doesn’t everyone strive for “five nines” availability? As it turns out, “the difference between those nines is extremely expensive,” says Info-Tech Research Group lead research analyst Darin Stahl.

Although many federal networks require 99.999 percent uptime, Defense Research and Engineering Network (DREN), the Defense Department’s research and engineering network, targets an uptime of 99.9 percent.

According to a DOD spokeswoman, the expectation of availability is lower because the agency uses the network to test new strategies, devices and software, and because a lower uptime percentage is more cost-effective. When it comes to DREN spending, she adds, DOD prefers to dedicate its DREN dollars to substantially faster throughput, rather than higher availability.

May 10 2011