Interruption of a site's electrical supply, where how the switch to backup power is handled decides between no visible impact and a full shutdown.
Power supply failure refers to the interruption of electricity to a data center, whether it originates from the public grid, a human intervention or an internal fault at the site. A properly designed data center has several layers of backup, UPS units then generators, meant to take over within seconds with no perceptible service interruption. The real peril therefore does not lie in the outage itself, which is frequent and normally handled transparently, but in the failure of the switch to these backup systems, whether through a power surge when supply is restored, human error during a maintenance intervention, or a synchronization fault between sources. This peril illustrates a feature specific to electrical risk in data centers: the triggering event is often minor and handled thousands of times without issue, while the actual loss arises from a procedural failure during an otherwise anticipated event. For insurers, the analysis focuses less on the probability of a grid outage than on the demonstrated reliability of switchover procedures and how often these backup systems are actually tested, data rarely available before underwriting.
On May 27, 2017, work on the power supply at a British Airways data center near London caused a power surge when electricity was restored, taking down the airline's IT systems, canceling hundreds of flights and costing roughly £80 million.
power supply failure, panne électrique data center, electrical outage data center