Here’s how Data Center Operators and Industry Best Practices can help Prepare for Tomorrow
Welcome to this installment of my series on Data Centers (DC), what they are, how they function, the opportunities for energy efficiency (EE) and the challenges found in their operation related to changing standards and growing demand.
How can utility energy efficiency programs insure they are prepared to keep up with the growth and demand of data centers?
Due to the fact there are many possible data center best practices approaches, we are only going to touch on a few in this blog. I recommend the U.S. Department of Energy’s Best Practices Guide for Energy-Efficient Data Center Design and their report from a 2016 seminar on Data Center Energy-Efficiency Best Practices. For a comprehensive list of over 35,000 articles visit energy.gov.
As a utility, there are a few important items to touch on regarding how to meet the needs of today’s data centers. Utilities must be ready to meet the demands in preparation for tomorrow’s explosion in data management. These include; airflow management, air-side and water side economizers and server virtualization. The utility must be ready to incorporate these types of energy saving technologies and many more best practices in their DC program.
The efficiency and effectiveness of a data center conditioning system is largely tied to air flow path, set point temperature, humidity and heat rejection equipment, with the goal of eliminating the mixing and recirculation of the hot air that the server racks are producing. The highest efficiencies are reached by removing waste hot air at the highest possible temperatures in one of many ways.
How do you do that?
- Hot Aisle/Cold Aisle: Arrange the IT equipment so that all heat is exhausted into hot aisles, and all air intakes draw from cool aisles. Cool air is supplied only into the cold aisle, with return air being drawn directly from the hot aisle.
- Rigid Enclosures: Build rigid enclosures to fully separate the heat rejected from the rear of IT equipment from the cool air intakes on the front.
- Flexible Strip Curtains: Use flexible strip curtains to improve the separation of hot and cold aisles by blocking the open spaces above the racks.
- Blank Unused Rack Positions: Openings created by spaces between servers in the racks form “holes” that can let hot air be pulled through the rack to the front. Block these “holes” in some manner to prevent that hot air from being pulled forward and recirculated back into the IT equipment.
- Exhaust vent hood structure: A system for removing heat in a sealed exhaust vent hood structure for redirecting air to vent outside or heat other areas of the building such as the office or warehouse.
Most standard data center cooling systems remove heat from inside by utilizing a mechanical cooling system. By using an economizer, however, when the outside air is cooler than inside air, the internal hot air is vented outside and replaced by that cooler, filtered air.
In this way, an economizer can cut data center cooling costs by more than 60% by using standard, commonly available low-cost equipment. Depending on the climate, the steady, 24-hour cooling load of a data center is well-suited to take advantage of seasonal and nighttime temperature variations to cool the space.
To work properly economizers must be engineered into the air handling system. Small data centers can be economically served by low-cost mass-produced package units. Larger data centers typically justify a more efficient chilled water system with central air handlers.
Keep in mind that in dry climates, controls should include redundant outdoor air humidity sensors to stop economization when the absolute humidity (or dew-point) is too low to prevent increasing costs of humidification on very cold days.
Water-Side Economizer (Free cooling)
Data centers present an almost constant, 24-hour, internal cooling load. Free cooling operates on the principle that during cool weather conditions, particularly at night, data center cooling loads can be served with chilled water. Dry climates can often realize excellent savings from an evaporative cooling or water-side economizer approach (Free cooling is usually best suited for climates that have wet bulb temperatures lower than 55°F for 3,000 or more hours per year)
This cooling method can be provided via a water-side economizer, which uses the evaporative cooling capacity of a cooling tower to indirectly produce chilled water to cool the data center during mild outdoor conditions (particularly at night in hot climates). A bin analysis using local weather data is required to properly assess the potential before investing in free cooling equipment.
Information Technology (IT) Systems
The fact is that active servers can use 60% or more of their full computational power on a moment-by-moment basis, leaving little extra space for additional throughput, such as now common-place live streaming and remote conferencing. For data center operators, the solution to meet these growing needs is to use server virtualization – the use of software applications to divide one physical server into multiple isolated virtual environments.
Some data center operators may dedicate each server to a specific application or task. Although a simple approach and easy to track problems when they arise, it doesn’t maximize the performance of a single server since the rest of that space is unused. One server doing one thing requires many servers, and a great deal of space to do a lot of things. That leads to overcrowding, with racks of servers consuming huge amounts of power and generating a lot of heat which will demand additional cooling.
Server virtualization addresses these issues by utilizing specially designed software on a single physical server creating multiple ‘virtual servers’. Using this approach, an operator could create enough ‘virtual servers’ to maximize the physical server’s processing power.
Although virtualization isn’t a new concept, it has been used on supercomputers for a generation. For DCs it’s a rather new approach – and one worth exploring, since it can increase both server utilization factors and server computational efficiency, allowing them to perform more work with less idle time.
The time will come when new, improved and more equipment will need to be added to the mix.
When that happens, best practices always involve purchasing equipment that meet at least the minimum efficiency recommendations of ENERGY STAR® (which average 30% more efficient than standard servers) and particularly, 80 PLUS™. 80 PLUS™ is a voluntary certification program to promote EE in computing. It certifies that products have more than 80% energy efficiency at 20%, 50% and 100% of rated load, and a power factor of 0.9 or greater at 100% load. By meeting these standards, the equipment generates 20% or less electric energy as heat at specified load levels, which reduces electricity use and bills compared to less efficient equipment.
Now that I’ve laid out some common operational and EE best practices for data center operators, what does this mean for utilities? The simple answer is: Utilities can help DC operators maximize their EE options by offering information, education, and where practical incentives for adopting most energy-efficient measures.
Most utility EE programs already offer HVAC incentives for such things as economizers, so why not add these as eligible measures for data centers?
What about offering energy audits that look at the entire data center envelope, the power chain and incentivizing airflow management strategies such as hot and cold aisle? In addition, incentivizing the purchase of new equipment that not only meets ENERGY STAR® standards, but also are 80 PLUS™ certified.
There are many best practices approaches that when implemented can provide verifiable energy savings.
As I wrote at the outset, these are only a slice of the best practices and suggestions possible for both data center operators that the utilities can incentivize. The options are limited only by the possibilities of ever improving technology and the imagination of those driving our energy industry.
DNV GL has successfully designed and implemented data center energy efficiency projects on behalf of our utility clients for more than eight years. Our team has significant experience on both new construction and retrofit data center EE projects, from small server closets to co-lo facilities. We work through the entire project life cycle of identification, justification, and evaluation, and follow this up with detailed engineering analysis and post-installation review, assuring that the promise of a data center EE project is fully realized. Visit our Data Centers Knowledge Hub for more information.
Our team is also available to work directly with enterprise-wide data center operators to assist them in company-wide energy efficiency initiatives. For more information, contact John J. Greco. John is a D.O.E. Certified Data Center Energy Practitioner, Certified Energy Manager, Certified Measurement and Verification Professional, and Certified AEE Building Sustainable Energy Technician Trainer with over 25 years of experience in the industry, and holds a Bachelor of Science degree in Electrical Engineering Technology from the State College of New York.