Ignore the data-center power crisis? Don’t press your luck

analysis
Apr 3, 20072 mins

Even organizations that are ineligible for PG&E's Virtualization/Server Consolidation Projects Incentive Program (or other utilities' variants) should give virtualization, and other greener, more energy-efficient technologies, some serious consideration. In a conversation last week, Bogomil Balkansky, director of product marketing.at VMware, outlined the "triple whammy" that's driving up power and cooling costs

Even organizations that are ineligible for PG&E’s Virtualization/Server Consolidation Projects Incentive Program (or other utilities’ variants) should give virtualization, and other greener, more energy-efficient technologies, some serious consideration.

In a conversation last week, Bogomil Balkansky, director of product marketing.at VMware, outlined the “triple whammy” that’s driving up power and cooling costs in the data center. To simply ignore the harsh reality of what some industry leaders have described as a power crisis would be to, well, press your luck.

Ignore the data center power crisis? Don't press your luck
And Whammy No. 3: Energy costs have been going nowhere but up. Balkansky said that on average, they’ve increased 20% since 1999.

“Under the weight of these three factors– quadrupling power consumption, double the number of servers, and the increasing cost of electricity — they’ve all compounded and lead to a situation where expenditures for power and cooling has become a significant force to reckon with,” he said.

He also noted that some utilities are struggling to keep pace with companies’ swelling thirst for energy, resulting in a datacenter migration from metropolitan regions like Silicon Valley and New York to destinations in Idaho and Washington State. Building a new data center is, of course, a monumental task that can take a couple of years and many millions of dollars.

During our conversation, Balkansky also noted that VMware is working on adding power management features to its product line, which strikes me as an invaluable addition. The way it would work is, it would dynamically transfer underutilized VMs to as few servers as possible during downtimes and power down the servers not in use.

That’s certainly something to watch for down the road, though Balkansky couldn’t say when we might see it.