Cloud computing is all the rage. In its simplest terms, it means outsourcing your company's information technology (IT) needs, from data and storage to software. All the servers and applications sit elsewhere in the Internet "cloud," but more literally in a data center or centers.
A recent study from Microsoft (with Accenture and WSP), "Cloud Computing and Sustainability", compared the environmental footprint of running business software internally or with an outsourced provider (in this case, Microsoft). The study showed that, compared to running their own applications, by outsourcing companies can reduce the energy use and carbon footprint of computing by up to 90 percent!
This is very good news. IT is one of the fastest growing energy hogs, accounting for at least 2 percent of global energy use. In my last book, Green Recovery, I focused on IT as one of five operational areas where green initiatives help companies save money quickly (the others were facilities, distribution, telework, and waste).
In the book, I cited statistics from IBM showing that less than 4 percent of the energy going into a data center is used to process something.
While the IT world has gotten a lot more efficient lately, there's still much room for improvement. And apparently moving your applications to the cloud can help immensely.
According to the Microsoft report (see page 6), cloud computing drives energy reductions in four related ways, which boil down to a few key leverage points:
- Reducing excess capacity
- Flattening peak loads
- Employing large-scale "virtualization" software
- Improving data center design.
Using the cloud addresses all three of the major energy-loss areas in the IBM chart: data center design tackles room and server cooling, while the other scale benefits mainly address the absurd waste, in percentage terms, from server underutilization (the far right bar).
Rob Bernard, Microsoft's Chief Environmental Strategist, likens the cloud to mass transit: "A data center essentially gets computing applications to carpool or take the bus instead of sitting in their own individual servers... but unlike mass transit vs. private vehicles, there is no tradeoff for convenience and on-demand availability."
So all of this is pretty logical. Scale is more efficient and allows for better resource planning. But I'd offer a few points worth thinking about, and one note of caution.
- The centralization of computing power should look familiar. To get some perspective on the study, I spoke with Mark Monroe, the new Executive Director of Green Grid, an organization dedicated to making IT more energy and carbon efficient. He compares the cloud to the electric grid, citing Nicholas Carr's book, The Big Switch, which Monroe says "compares utility computing development to the emergence of centralized electrical generation in the early 20th century." Like electric plants, Monroe says, central computing "utilities" benefit from scale and high utilization.
Cloud providers want to provide an hour of CPU time, a Gigabyte-month of storage, a CRM transaction, an email, or a web page for as little cost and as high a margin as possible. That just has to lead to higher efficiency than someone focused on delivering a feature internally.
In short, finding providers and partners that can take some of your energy-using operations to scale, and manage them in a shared capacity, is good for your footprint and your bottom line.
This post first appeared at Harvard Business Online.