The Internet runs on a simple equation: bandwidth and computer power are growing exponentially as their price keeps falling. Which is fine, since the demand is also surging the same way. There is one glitch though. More Internet traffic means more data centers across the globe and more electricity to power and to cool millions of overheated processors. Problem is : energy supply can’t follow the same trend. The cost of electricity consumed by data centers doubled between 2000 and 2006 and could double again by 2011, as mentioned in this Business Week article. This trend is about to become the main problem of the Internet sector. Especially with oil at $110 a barrel (one analyst even sees spikes at $150 ).
To address the question, two simultaneous races have begun. One is to use new technologies to lower the energy consumption of data centers. The other is to find the best locations, regardless of where the traffic demand resides, since the expense of transmitting data is small compared to the processing cost. A worldwide scouting operation is underway: Iceland or Siberia are on the radar-screen of data-hungry companies such as Google or Microsoft. Over there, energy is abundant, and cooling is natural. Political problems can be natural as well… No one wants a Putin-like dictator cutting the supply of bits coming from Northern data centers the way he does with Ukrainian pipeline.
It will be years before research for energy efficiency or Siberian computer farm will yield any result. In the meantime, content could take the heat. YouTube is said to account for 10% of the entire Internet traffic. This raises questions about YouTube economics : a free service, consuming so much bandwidth/energy, yielding very little ad revenue ? Can’t go for ever. In the media business, we were used to think that digital costs of distribution were negligible. That might be true compared to a newspaper, but certainly not in absolute, future terms.