This audio was created using Microsoft Azure Speech Services
There’s an old saying: “Don’t throw the champagne out with the cork.” What’s the champagne of the data center? Efficiency. Yet poorly designed and implemented data center power and cooling infrastructure needlessly wastes more than 60,000,000 megawatt-hours per year of electricity worldwide every year. For context, consider that 1 MW of electricity can power about 1,000 homes. That’s a lot of inefficient cork-popping.
As a data center manager, it’s unlikely you’re wasting energy or creating inefficiencies on purpose, but you may not even realize which practices are the catalysts and how easily the loss can snowball. If you’re not careful, efficiencies can slip through your fingers. Here are seven ways to avoid killing your data center efficiency.
- Don’t oversize.
As our White Paper 37 explains, “the utilization of the physical and power infrastructure in a data center or network room is typically around 50-60%. The unused capacity of data centers and network rooms is an avoidable capital cost, and it also represents avoidable operating costs, including maintenance and energy.” Want to make some simple savings? Scope your physical infrastructure carefully, and expand later as necessary.
- Don’t place cooling too far from IT load.
This can have a significant negative effect on performance due to air mixing . Techniques such as in-row cooling put heat rejection in much closer physical proximity, increasing efficiency.
- Don’t block cool airflow.
When cool air can’t get where it needs to go, hot spots and inefficiencies can occur. Consider the placement of your CRAC’s (computing room air conditioners), maximizing airflow to let them do their work without wasting power. When cool air flows below a raised floor, make sure it is not blocked or diverted by cables and other obstructions.
- Don’t over spread the IT load.
If your IT equipment is scattered across too many racks (which may have been done during a virtualization project), you may be wasting resources. Consolidating equipment to make the most of individual rack capacity should yield improvements. In general, efficiency increases dramatically as the load is consolidated to minimum amount of racks and close coupled cooing is deployed. Get full details in White Paper #46: “Cooling Strategies for Ultra-High Density Racks and Blade Servers.”
- Perform proper maintenance.
Preventative maintenance both reduces risk and ensures equipment is operating at peak efficiency. Letting the cycle slip opens up the door to downtime.
- Utilize containment.
Air from the hot aisle and cold aisle should not mix. Doors and roofs should be added to the hot or cold aisle and blanking panels put in the racks to prevent escape. Read White Paper 135: “Impact of Hot and Cold Aisle Containment on Data Center Temperature and Efficiency” for more information.
- Implement DCIM.
To see any meaningful improvement in efficiency you’ll need to benchmark before, and measure after any change. DCIM software and monitoring can help tell you where electricity is being used, or wasted.
Even small changes can make a big difference, but beware: when multiple issues occur together, efficiency is thrown out. Solve them all, and you might have cause to pop some real, rather than metaphorical champagne.
Want more tips? Check out White Paper 49: “Avoidable Mistakes that Compromise Cooling Performance in Data Centers and Network.”
For a thorough assessment, use our Data Center Efficiency Calculator.
