The Weakness in Data Center Power-Backup Legacy Design
Data centers are ravenous, intolerant beasts consuming, according to some reports, up to 2% of the world’s electricity production. Power – clean and constant power – is an imperative. Customers dependent upon monster data centers for business-critical applications expect zero downtime. Consequently, the focus of data center facility managers is on reliability.
Data centers deploy large and expensive battery-based Uninterruptible Power Supply (UPS) systems to ensure continuous service to a demanding customer base. To meet expectations for 100% uptime, facility managers trade electrical efficiency for the assurance of knowing that their backup systems will start the millisecond they are needed. Despite calls by investors and the general public to lower utility costs and reduce their carbon footprint, facility managers have largely rejected industry efforts to introduce ‘green’ or ‘eco-mode’ solutions, fearful that these complex systems will fail when needed.
However, there is a significant weakness in the legacy design of power-backup systems at data centers. Developed before the advent of modern electronic voltage regulators (EVR), data center UPS systems are designed to do double duty – supply backup power in the event of a blackout, and continuously monitor and condition power.
The implied promise of a UPS is uptime. It is ironic, then, that a report sponsored by Emerson Network Power [2] found that one of the leading causes of unplanned power outages was failure of the UPS system itself.