In a world that is always on, in which speed of service is expected and memory capacity is taken for granted, the “cost” of outages is rapidly rising for data center providers. According to the Uptime Institute’s Abnormal Incidents Reports database, fully 70% of reported outages are directly attributable to human errors. Other sectors, notable process industries, have faced this same issue. This paper by ABB explores best practices and standards from related industry sectors that can be translated to data centers for decreasing human errors and preventing unplanned outages.
Among other things, it explains the important role of UPS in this context:
„Another aspect of special interest for data centers is energy efficiency. The UPS system is heart of the data center’s power system, a critical insurance policy against power disruptions. Now, many data centers are operating their UPS in “eco mode,” in which a bypass connects the incoming AC main power to the outgoing AC without incurring the losses associated with AC-DC conversions. Doing so raises UPS efficiencies from the 95-97 percent range up to 99 percent. That might not seem like a large improvement, but even small energy savings can become significant in dollar terms as facilities grow.
UPS architecture is also undergoing change. Modular designs now enable data centers to easily scale up their UPS as needed while eliminating single points of failure with the use of parallel feeds. Modular systems also reduce the chance for human error thanks to simplified maintenance that does not require the entire system to be taken offline. Because of these benefits, modular UPS will likely become a best practice, especially for larger data centers and those experiencing rapid growth.“