Security & Privacy

Simple file backup may not be enough for today’s business demands

Data Loss Prevention – not a panacea

Articles  
Conferences  
Podcasts  
White Papers  
Press Releases  

Unified Business Comm.

Virtualization

RFID

Green IT

IT Briefs

Viewpoint

Archives

Partners


 
 
 

Green IT

The four M's of data center energy efficiency


The critical usage of energy by data centers makes it increasingly important for companies to explore more green options


John Consoli, CTO, AFCO Systems

Highlights

  • For increased efficiency, the focus must extend beyond ambient room temperature.
  • Airflow management is a key factor toward enhancing data center efficiency.
  • Systems should be sized and redundancies configured to maximize efficiency.

In a recent report, the EPA sounded a clarion call to data center managers in its “EPA Report to Congress on Server and Data Center Energy Efficiency.” The report highlights the critical usage of energy by data centers and makes recommendations about how to alter the state of current and future data centers to lessen a looming energy crunch.

Data center managers who follow the proactive steps outlined below will ensure their ability to monitor and control current data center power consumption while facilitating necessary changes to increase the efficiency of their operations.

The Four M’s of Data Center Energy Efficiency

Measure: For increased efficiency, the focus must extend beyond ambient room temperature. Supply, server inlet, and return must be measured as well. Many managers who believe they are “out of cooling” may be surprised to find out how much cooling is actually being wasted. Airflow must be measured along with power consumption and air temperature. Understanding the relationships and interactions of these different components will help ensure greater control.

Manage: The EPA study clearly identifies airflow management as a key factor toward enhancing data center efficiency. A first step is to “plug the leaks.” Many data centers have floors with tiles removed or holes that no longer serve a purpose and instead, negatively impact airflow. The key is to be able to manage the flow of supply air as closely as you can to the computer inlet. Equipment such as variable speed fans at the air-handler, or preferably at the rack level, and the use of chilled water pumps will help you manage airflow at the cabinet level.

Monitor: With loads approaching 15-20 kW per cabinet, cabinet level monitoring is no longer an option. It’s a necessity! This process facilitates proactive management of the data center against “downing” events.

Maintain: Planned and systematic enforcement of best practices is a vital part of keeping the data center environment efficiently in tune with technology turnover. Trend analysis helps offset the obsolescence of legacy infrastructure. Air-handler systems should be rebalanced following a significant IT configuration. Systems should be sized and redundancies configured to maximize efficiency.

We’ve found that by taking these few simple steps, you can achieve great gains toward your data center goals for optimizing power consumption and increasing efficiency.

John Consoli is Chief Technology Officer at AFCO Systems, a global leader in the design and manufacture of advanced enclosure technology for mission-critical data center environments. AFCO Systems has the world’s largest installed base of active, air-cooled, high-density (5-8 kW) and extreme high-density (up to 20 kW) server and network cabinets. The company’s enhanced airflow management technology provides reliable, efficient rack cooling in accordance with ASHRAE standards. The company is headquartered in Farmingdale, NY, and has a global presence with operations in North America, Ireland, London, Singapore and Tokyo. For more information, visit http://www.AFCOSystems.com or call 631-249-9441.



Comments:


Add Your Comment

(We will never sell your name or e-mail address to anyone)

Remember my personal information
Notify me of follow-up comments?

Please enter the word you see in the image below: