The most popular concept of sustainability revolves around energy use and while I have no issues with energy as an issue, I think in business the idea goes further — all the way to products and customers. I will leave the last couple of ideas for another time and concentrate on energy today. Actually, energy is a huge topic and the only thing I want to focus on is the data center, not whatever you have in the garage.
The data center might sound like a funny place to start but it is both germane to CRM and a great place to kick this off. As it turns out the U.S. Department of Energy (DOE) has already done the heavy lifting for data centers but for some reason few people know about it. Just for fun, see if you can guess how much of the energy that is generated for your data center goes to useful computation. While you’re thinking, let me give you some DOE data.
First off, the majority of the energy generated goes up the smoke stack. Did you know that? And we’re not talking about some close election’s 51% in this case. In the DOE’s presentation, a full 65% of the energy in the source material, let’s say coal, never leaves the building by the transmission cables, it’s just waste heat.
That leaves 35% to do some useful computing. Of that, two percent is lost on the power lines, and 33% makes it to your building. Naturally you would expect the answer to be that thirty-three percent but we’ve set the bar a bit higher by asking for useful computation. As luck would have it, the majority of the power that makes it to the data center is used for cooling and to run the lights. According to DOE 15% or less of the energy that started out as coal does something as important as running your screensaver.
DOE has developed a set of metrics that we can all use to gauge our efficiency and the metrics are so easy anybody can play. So in this case, for instance, Energy Efficiency = Useful Computation / Total Source Energy. A slightly more formalized version of this equation gives us the metric DCiE or Data Center Infrastructure Efficiency which nets out as energy for IT Equipment / Total Energy for Data Center. Armed with an electric bill and some manufacturer’s specs for your IT gear you should be able to do your own analysis very easily.
Here’s where things get interesting. DOE has identified ways to improve your data center’s performance and while it notes that a typical DCiE is less than 0.5, the best practice results can be as high as 0.85. We’re talking real money here. In a case study presented by DOE, Lucasfilms, Ltd. a movie studio that specializes in animation, was able to save $343,000 per year after implementation costs of $429,500. That means a one time investment of over $400k provides a recurring savings of over $300k for a payback time of 1.2 years.
Most importantly that means energy savings for Lucasfilms of over three million kWh. Did you know that each kilowatt generated by coal puts more than two pounds of carbon dioxide into the air?
Is this a big deal? Yes it is. The DOE Industrial Technologies Program is trying to improve the energy efficiency of U.S industry because it consumes 32 quadrillion Btu per year, almost one-third of all energy used in the nation according to the report. A trillion is 1 followed by 12 zeros, a quadrillion has 15 zeros. This is the only number I know of in general use that’s bigger than the national debt.
I don’t have any data on the efficiency of a single data center compared to a cloud data center but my instincts tell me that if we’re going to get to the best practice level of energy efficiency in the data center, one of the easier ways to do it will be to consolidate data centers into ultra-efficient clouds. Clouds can use advanced technologies to cool a data center and employ slack capacity thus leveraging the efficiency of using less gear to serve more users.
I saw some Gartner data recently showing that server deployments peaked in 2008. That could indicate the growing popularity of cloud computing but it could also simply be another reminder of the recession and the slack capacity that frequently accompanies a slowdown. If the answer tracks the former explanation it would be an example of the invisible hand of the market adjusting demand in the face of increasing costs. If it’s the latter, it could be an example of that same invisible hand adjusting supply by decreasing demand. Don’t you just love economics?
My big takeaways from this exercise are several.
- Saving energy is yet another benefit of cloud computing.
- There’s more work to be done assessing the costs and benefits, but DOE has taken a leadership position in offering some commonsense ideas.
- We can all save money in our data centers and the DOE has some practical advice.
Lastly, sustainability is about making and saving money in business. If we keep this in mind, we’ll have fewer theological debates about what’s good for the planet and whether global warming is real. It will also make it easier to see how products and customers need to be seen as sustainable resources. We’ll never agree on everything but we should be able to agree on the advisability of making money.