Begin a discussion of the best ways to optimize a data center and your conversation is likely to focus first on reducing capital and operating costs while boosting energy efficiency. A direct current (DC) power distribution system, for example, can be five to 15 percent more efficient than alternating current (AC). Moreover, coupling the DC data center with a micro-grid powered by biofuel, solar or wind sources can create zero-net-energy (ZNE) buildings that sell excess power back to the utility, resulting in net-zero energy costs.
Beyond efficiency and choice of power systems, however, other considerations impact optimization. Reliability and availability are “givens” in today’s data centers, but they still require choices to be made. For instance, the tier-level target must be appropriate to the reliability required. A self-healing function in the power supply network can increase reliability, but may add to the cost.
Of course protection and safety must be rigorously implemented, and plans to scale up the IT component of a data center in the future must also incorporate plans for more backup-power generation and decisions on configuring uninterruptible power supply (UPS) systems. Alternatively, if a smaller footprint is desired for the center as a way to minimize real estate costs, designers must build in server racks and other equipment with higher power density, which means higher cooling bills.
The owner, architect and engineer must collaborate on the best way to meet each of their goals, understanding that an optimal design is always possible if the objectives are clear and discussions are held on fundamental assumptions, such as the operating life of the design, geography and site constraints.
Current trends offer some guidance for these considerations as designs favor data centers that are greener, through zero-net-energy facilities; modular in their UPS and generator systems; “cloudier”, benefiting from colocation; and hotter, with higher-power density servers.