Data centers typically waste energy in a number of ways, in large part because historically they have been designed with reliability and capacity as top priorities, leaving efficiency to take a back seat.
But an enterprise can save a lot of money by running its data centers in a way that reduces power consumption and, consequently, energy expenses.
The first step in devising a data center’s energy-management strategy and processes is measurement: Determining how to allocate energy costs and carbon to IT users.
What you choose to measure, however, depends on your enterprise’s energy management goals.
An APC by Schneider Electric white paper identifies three different energy-assessment goals and their implications:
One-time or periodic benchmarking of performance
This is where you start if you’re not even sure that your data center needs an energy-efficiency program. Your benchmarking may well prove your data center is more energy-efficient than comparable facilities. Conversely, it may demonstrate a need for an ongoing energy-management strategy.
One note of caution: Achieving this baseline energy-assessment goal will not provide the information an enterprise would need to devise an actionable plan that would result in lower energy costs or carbon output. To get that data, must focus on Goal Nos. 2 and 3:
Pass-through allocation of energy of carbon to others
Data centers that offer outsourced storage services to other enterprises can save money by allocating, or even directly billing, data center customers for energy or carbon. Making customers aware of their own data center expenditures will motivate them (well, the smart ones) to cut costs through virtualizing servers, turning off dormant servers, implementing power management functions and managing excess storage.
We at APC by Scheider Electric believe that, based on the maturity and virtualization level of a data center, cost reductions of 10% to 80% are possible.
Use of information to reduce infrastructure energy or carbon impacts
Without supporting equipment and systems, a data center is just a bunch of servers that don’t work. The physical infrastructure that supports a data center – the power and cooling systems, lighting, controls – is a heavy energy consumer and generator of carbon, and therefore a prime target for data center cost reduction.
There’s a metric for benchmarking data energy infrastructure energy use: Power Usage Effectiveness (PUE). In data centers where it’s suspected that physical infrastructure power consumption exceeds that of the IT load, there may be many opportunities to reduce power costs.
Since this goal is to reduce energy or carbon impacts across a complex physical infrastructure, it requires data wherever power is being consumed – lighting costs, air-conditioning, heating, etc.
APC/Schneider Electric estimates that managing the data center’s physical infrastructure for energy efficiency can result in savings of 10% to 40%.
For more information on data center energy efficiency, read the APC by Schneider Electric white paper, Allocating Data Center Energy Costs and Carbon to IT Users.