Worldwide, data centers use about 30 billion watts of electricity annually. This is roughly equivalent to the output of 30 nuclear power plants.
This data alone is quite interesting.
However, when one understands that data centers can waste 90% or more of the electricity they pull off the grid by running their facilities at maximum capacity around the clock, whatever the demand, this data moves from interesting to sobering. Even more sobering is the fact that there are affordable, no risk ways, to drastically reduce energy waste but companies choose not to implement them.
Adaptive Computing is in a unique position to see the source and direction of green computing because we span both the enterprise data center through our private cloud solutions and non-traditional data centers through our high performance computing solutions. From this unique vantage point it is clear that HPC, as it has done in so many other areas, is leading the efforts behind the adoption of green computing within universities and government agencies.
Furthermore, green discussions appear to be occurring much more in EMEA than in North America. This is not surprising given energy costs in EMEA are often double those of the U.S. and it’s also not surprising that the U.S. consumes about 1/3 of all energy attributed to data centers. I’m not a hardcore environmentalist but I’m certainly someone who is against wanton waste that could be easily reduced with minimal risk.
What do I mean when I say “easily reduced with minimal risk?”
I mean implementing policy-based solutions that automatically and intelligently consolidate compute resources when resources are underutilized. The solutions put the resources freed up from consolidation in a state of off, hibernate, sleep, or standby, and then intelligently bring resources back online when certain thresholds are met or service level agreements (SLA’s) are at risk.
The key points here are “automatically,” “intelligently,” and “policy based.” Those keywords apply to both enterprise data centers and traditional HPC environments. Coincidently, Adaptive Computing is a leader in green computing and provides all three of these key ingredients.
How can I say this? Because we’ve been doing this for years in HPC and Adaptive Computing has 40+ patents issued or pending, many of which cover the base technologies needed to implement green computing.
I often participate in sales meetings with prospective customers in which we spend a large portion of our time discussing optimization through intelligent policy management. All too often, when discussing consolidation type policies, we hear comments such as “I already purchased the hardware,” “I don’t want to risk application performance,” or even worse, “We are not concerned with saving money at the moment.” Granted someone who is responsible for performance with little budgetary accountability, usually makes these types of comments and always when the CIO isn’t in the room.
There are ways to drastically reduce or eliminate risk and there will come a day when saving money will be a priority. Plus it’s simply the right thing to do and there is a need for more of that in the corporate world today.
- Set green data center goals, track them and then publicize your accomplishments.
- You will be one of, if not the coolest person, in the company by proactively doing a good thing for society while saving money for the company. You will lose the coolness if you have to be told to do it.
- Be proactive!
A recent survey by the consulting firm McKinsey & Company analyzed energy use by data centers and found that, on average, they were using only 6-12% of the electricity powering their servers to perform computations. The rest was essentially used to keep servers idling and ready in case of a surge in activity that could slow or crash their operations. So even if a data center improved their utilization to 30% or 40% we could reduce power consumption by 3-4 times and running ones servers all at 30-40% will not add significant risk.
In summary, there is considerable energy waste in today’s data centers. There are easy and affordable solutions that eliminate perceived risks associated with better utilization of compute resources. Policy based solutions that incorporate automation and intelligence are key to reducing risk and increasing manageability.
Green computing is the right thing to do, saves money, and you’ll be a hero. Speaking of hero’s, it turns out that one of my hero’s was wrong when he said, “It’s not easy being green.”