In an ideal world, all energy would be free, data resources would be unlimited, and every day would be Christmas, Easter and your birthday rolled into one.
But as my grandma always told me, “this ain’t a perfect world, kid.” As you can probably guess, grandma wasn’t one of those sweet, little old ladies who sat in rocking chairs all day knitting sweaters.
Enterprise executives, and the environmental lobbies that are prodding them, need to get real about two key aspects of the burgeoning “green data center” movement. The first is that no matter how often you place the word “free” in front of an eco-friendly endeavor – free heat, free cooling, free power – none of it is truly free. There is both a financial and environmental cost to everything we do.
The second is that no matter how efficient or effective your data operations become, new technologies and new expectations all but guarantee that the data center will never be green enough. A new platform or a new practice will always be waiting in the wings to build on what you’ve already accomplished.
Having said that, though, let me add yet a third truism to this equation: Doing something is better than doing nothing.
This is why we applaud companies like SAP in their drive to build computing infrastructure using 100 percent renewables, such as wind and solar – not because this supposedly puts them on a path toward eco-nirvana, but because it at least lessens the production of energy through dirtier or even dangerous methods such as coal, oil and nuclear. True, part of SAP’s plan is the swapping of energy credits/certificates and other mechanisms that allow other firms to continue using fuel-based energy methods, but this at least creates an overall decline in emissions from traditional power plants. Again, the perfect world vs. the real world.
As well, Apple seems to be out in front on the green data center issue, employing hydro and other renewable resources for projects in Oregon, North Carolina and elsewhere. Note, however, that the hydro project in Oregon requires a half-mile trench to be cut from an existing irrigation canal, which itself is closed off during the winter months. So in the end, the 3-5MW of power that the local utility expects to generate is seasonal at best.
Still, Apple is getting better reviews for its green efforts than hyperscale juggernaut Amazon. According to Greenpeace, Amazon is one of the worst performers when it comes to reliance on fossil fuels and has refused to “come clean” regarding key energy consumption metrics, including raw numbers on exactly how much energy the company requires to operate its data services. Amazon counters that it has been a leading supporter and deployer of virtualization and software-defined technologies that have reduced hardware footprints and the energy they draw. But as I mentioned, “greenness” is a relative concept. It is not enough to implement new technologies – you have to be seen as doing more than other players in your field.
But it is also true that the enterprise industry cannot conduct the green revolution on its own. It will take cooperation across multiple sectors, including utilities, government and even data users themselves. The Green Grid is taking steps in this direction with a detailed series of analyses that investigate the utility industry’s policies and incentive programs to help customers curb their energy appetites. Noting that global energy consumption increased nearly 20 percent from 2011 to 2012, the group is calling for utilities to play a leading role in education and training with regards to efficiency, plus a general simplification of existing incentive programs and the tools needed to monitor and maintain efficiency. As well, there should be a renewed effort to direct these programs at the IT industry in particular, that being the single largest energy consumer on the planet.
And ultimately, this notion of greenness will stand or fall not on the infrastructure we deploy or the way in which power is generated, but on the nature of data and data services themselves. No one can fault the need to occupy server, storage and networking components with business-related data, scientific research or perhaps even government surveillance and secrets. But when literally terabytes of storage are given over to the preservation of umpteen copies of the latest crazy cat video or hundreds of cell-phone captures of last night’s Jay-Z concert, inquiring minds will naturally want to know: Is this really the best use of our global data infrastructure and the energy needed to operate it?