(4-Dec-09) No one disputes that UPS systems are an essential part of any data center’s business continuity planning. However, selecting the right UPS involves weighing everything from runtime needs to long-term scaling.
UPS Type & Availability
Reaching any backup power decision starts with determining how much power is needed across all necessary storage, server, and network resources. Traditional UPSes were centralized—one big, fully loaded, high-capacity box with octopus cabling routed to an array of servers—whereas newer solutions might involve putting one UPS in each rack; each of these UPSes, in turn, network back into a central management system that admins can use to track battery levels and remotely control tasks. Between these two extremes sit centralized but modular products, which are essentially high-capacity shells that managers can fill up with batteries over time as scaling is needed.
Although a more modular approach might seem smarter because it allows organizations to simply pay capital expenditures as needed rather than all up-front, this benefit must be weighed against availability needs. The more pieces a solution has, the more possible failure points there are.