In an earlier post, we discussed data center energy saving steps that were primarily about changing behavior, such as powering off unused equipment or tuning systems. These ideas and others can be found in an article in the March issue of Schneider Electric’s Intelligent Energy publication, “Checklist: the Top 12 Data Center Energy Saving Best Practices (pages 12-13).
While all 12 ideas can have rapid impact on savings, it’s worthwhile noting that some do require investments in new equipment or expense in reconfiguring data center layouts. But even those steps with a price tag can bring relatively rapid energy savings, so they should be considered as part of the investment plan for the data center. It’s possible some of the following ideas have been overlooked in the data center’s investment plan, or that some of these lower-cost steps could be front loaded in the plan to gain some return before more costly steps are implemented.
Among the ideas to consider:
- Idea 2: Deploying hot-aisle/cold aisle arrangements. This step involves orienting the rows of racks so that the fronts of servers face each other. There are infrastructure costs to consider, of course, such as containment, and recabling (either subfloor or above the rack) as well as raised flooring and perforated tile commonly needed for cold aisle arrangements. Better or different monitoring may also be needed. For more details on room options, check out this post last fall by Paul Desmond, “Examining Your Cooling Options: Room, Row and Rack,” as well as a podcast Desmond hosts on, “Boost Cooling System Performance with Containment Systems.”
- Idea 7: Row-based cooling. There also is a trend towards cooling entire data centers with row-based cooling, rather than using a hybrid strategy of row-cooling for racks and supplemnetal cooling for ancillary systems. Use of row-based cooling reduces the mixing of hot and cold air streams and reduces the length of air flow paths. While it’s likely that row-based cooling should be part of a larger plan, it is possible row-based cooling is being underutilized. For more on this topic, check out the post by Mr. Desmond, “Why Row-based Cooling May Be All the Cooling Your Data Center Needs.”
- Idea 5, Instrumentation to Monitor Energy Consumption. The data center should be instrumented to identify and warn about suboptimal conditions. Assuming a data center operation has data center infrastructure management software (DCIM) or will soon deploy DCIM, additional instrumentation may be in order to fully reap the benefits of monitoring and other functions in DCIM. The old axiom, “garbage in, garbage out” hold true for most software management systems, so a bit more instrumentation could be exactly what’s needed to ensure that DCIM gets fed a stream of high-quality, relevant data on conditions.
- Idea 9 (High Efficiency UPS) and Idea 10 (Variable Frequency Drives). These energy-saving best practices are very similar in that they require new gear, but carry energy savings that can be quantified, leading to relatively short payback periods. For example, at 30 percent load, the newest UPS units pick up more than 10 percent in efficiency when compared with the average of currently installed low utilization systems.
Crunch the numbers on these areas of investment. It could be that at a relatively modest cost (versus bigger ticket investments such as entirely new data centers or large chillers), the ideas above could bring significant savings.