Summary:

Subodh Bapat, who heads up computing giant Sun Microsystem’s energy efficiency charge, advocates the notion of “energy-proportional computing.” While most data centers today use basically a constant amount of power regardless of workload, engineers and data center designers are working on making servers use the equivalent […]

Subodh Bapat, who heads up computing giant Sun Microsystem’s energy efficiency charge, advocates the notion of “energy-proportional computing.” While most data centers today use basically a constant amount of power regardless of workload, engineers and data center designers are working on making servers use the equivalent amount of energy for how much work they do.

With energy prices, and computing demands, on the rise, the trend makes sense. Companies across the board, from Sun to Google to Advanced Data Centers to International Data Security, are innovating on a variety of ways to make data centers more efficient. But sometimes it’s the simplest steps that can do the most good. While we’re all waiting for energy-scaling data centers, Sun’s Bapat offers three simple suggestions for companies looking to get the biggest bang for their buck when it comes to cutting data center energy:

1) Refresh: Identify which servers are being used efficiently and which servers aren’t. Older servers tend to be energy hogs, while newer hardware like Sun’s CoolThreads server line can conserve energy and lower electricity bills.

2) Virtualize: Invest in technology like virtualization to consolidate work loads onto less machines. Less servers mean less rack space and less costs for cooling.

3) Assess: This is the age of transparency, so why not use the power of software and energy auditors to know for sure how efficient your data centers are? Assessments can help prompt changes to the basic architecture of the data center, among them placement of heated and cooled hardware (don’t mix ‘em).

By Katie Fehrenbacher

You're subscribed! If you like, you can update your settings

Comments have been disabled for this post