Happy Earth Day! I just wanted to take some time to acknowledge this occasion and what I think it means to data center operators.
Earth Day started around 1970 and when it gained traction, the idea was to educate and inform while celebrating our planet. Well in that vein, I did some reading today on energy efficiency and conservation. Where else, besides in heavy industry, do we have the opportunities to save millions of kWh a year? Lots of different estimates of energy consumption exist, but roughly 1.5-2% of all energy is spent in data centers in the United States (good luck finding a solid source for that number). And as IT becomes a bigger sector of the economy, so too has the consumption of data centers increased. These old centers give us so many opportunities.
I feel good about my contributions. My work at our data centers has taken more than 2 MW off the grid, year round (significantly more more in the gained economizer months), and I feel like at least 800 kW of that was stuff that if not for me, it probably never would have happened. I feel good about that, and sure someone else could have come along and pushed a lot of what I pushed, but they didn’t now did they! So that roughly 20 million kWh is something like 3,000 cars off the road, 2,000 homes worth of energy 32,000 barrels of oil, or lots of other neat ways to say “did something good for the planet“. And that was mostly in 2013, and that impact is every year for the life of those data centers. In feels good for the company terms, call it something like $2,000,000 a year that doesn’t need to be spent, with a 15 year lifespan on improvements those cash flows would have a Net Present Value of around $17,000,000 and an Internal Rate of Return around 250% (*brushes that dirt off his shoulder*).
Yet, reflecting on this Earth Day, if I am truly honest with myself I know there is a lot left to be done. Opportunities still abound, and “good ideas” for the company still exist everywhere I look. Conveying that these are in fact “good ideas” has been one of the biggest challenges. I show great NPVs, IRRs, Profitability Indexes and more. I propose projects that kick the crap out of other investment opportunities in pure financial terms, but yet still have trouble getting people to believe in my form of “magic” as it seems. A 2 year payback period for long term projects is an insane metric to pursue. At the same time, we have met and beat that metric, but only by the sheer quality of the opportunities we went after and the help of utility companies rebates. So how does the conversation evolve? How do I get better? How do we get better? How does everyone get better?
Avoiding politics as much as possible, I can still acknowledge irrefutable science here. Human activity is having significant, negative impacts on this planet. Things are changing at paces unseen before human civilization. Species are going extinct at rates 1,000 to 10,000 times faster than at any other known point, save for extinction level events. And yet opportunities to reduce human impact by reducing consumption which coincide with the profitability of businesses lay in wait. Maybe this is as much a plea as an expression of opinion. How can I do more?
I do not believe the answer lies in the next new technology, or the next new build strategy. I don’t think the answer is walking away from old centers in lieu of new ones, or abandoning technology altogether and chasing each other around on dune buggies until the apocalypse comes. No, I think the answer is right there in front of us. Finding ways to operate the locations we have logically, sensibly, safely, and using existing technologies to significantly reduce our impact while simultaneously improving the yield from those assets. While I do focus on the infrastructure layer outside of the rack for the most part, there is common sense, logical steps to be taken at all levels.
Really using the ASHRAE bands as they intend, pushing those server inlets way up, and even letting them drift out of the recommended band on the hottest days is a simple example. Heck, there are UPSs now that can seamlessly switch from bypass to double conversion, offering efficiencies of 99% or more. It is hard to get much more from a UPS than that. Improve the distribution paths efficiency, improve the cooling efficiency (reduce the need to cool at all using outside air!), and improve the IT efficiency via virtualization, server management, effective rack level practices, and intelligent equipment selection. When said out loud, these things all seem so simple, yet in practice leave so much room for improvement. Why is it one data center can operate at a PUE of 1.18 in Nevada while some in Canada can’t get below an average of 2.00? Are the systems in that center really that far of a departure from the others? (In some cases, yes they can be. DX only cooling, low supply temperatures to compensate for poor air distribution and hot aisle cold aisle containment, and very inefficient distribution can have significant impacts, but that doesn’t really make the point I’m getting at).
In the spirit of Earth Day, I hope others find my blog informational and helpful in their pursuit of energy savings. I would love to hear more about everyone else’s successes in closet environmentalism, doing things that are not only good for the company but good for the planet. Please feel free to share and comment below. If you have any questions, I will do my best to answer them, or find someone who can. And if you think you have something to help me get better, by all means please share.
Find and identify the outliers and bring them back to the mean. At the very least, I know I have some job security as these industry opportunities still prevail. Finding all those outliers won’t be easy. Hopefully by next Earth Day I will have an even more compelling story to tell.