I have been digesting the latest vicious assault that Greenpeace has made on the data-centre industry - with their naming and shaming of the social-network operators and their, so called, ‘dirty data’.  Greenpeace seem to have accepted that they have been unsuccessful at influencing the US government (and every other one) to accelerate the construction of renewable energy sources and so have decided to go after the publicity generated by attacking the social habits of the younger generation, i.e. Facebook, Google et al.  All of this attack is flying under the banner of ‘cloud’ computing – although it’s not entirely clear that Greenpeace actually understands what that means (a bit like most of the rest of us!) – and ‘dirty’ electricity, mainly that generated in the USA with a high coal content in the fuel mix. 

The attack doesn’t do the rest of the industry much good as their targets, like Facebook, are in no way representative of mainstream data-centres and it’s probably worth noting that many other parts of the world don’t burn brown-coal to generate power.  The fact the Greenpeace also doesn’t like ‘nuclear’ (do they want most of the world to continue in fuel poverty?) is also a feature of their energy policy, despite it being low-carbon.  I’ve said it before but it is worth saying again that Facebook’s 1.07 PUE is pretty remarkable (and unobtainable for 99.9% of the rest of the market) so picking on them is rather perverse.

But are Greenpeace in danger of alienating their future support base by telling the world’s social media addicts (the younger generations) that their beloved services are unethical?  It doesn’t seem to make good long-term PR sense.

So what are the dynamics of the growth in power consumption of ICT?  What Greenpeace seem not to understand is that in the context of the growth rate in digital services (fuelled no doubt by Jevons Paradox* and only partially mitigated by Moore’s Law) the growth in renewable energy capacity is out of step.  The growth in data is >60% CAGR and Moore’s Law (so successful for 30 years in predicting the doubling of capacity every two years) is struggling to maintain 40% CAGR with the steam running out in about 5 years time?  So, fairly simplistically, data-centre power is growing >20% CAGR whilst the capacity of renewables is growing at what, 5%, 10% per year?  We need a paradigm shift to break that cycle and no one knows where or when it will come.  Even if it came ‘today’ it would take at least 5 years to gain a major share of the installed base.

Then we could consider the lower-carbon enablement that data-centres bring to our society – better communications, more education, less travel, smart-grid enablement, smart-buildings … even more effective democracy as we saw in the Arab Spring … Even Greenpeace must understand that; they actually use the very networks that they hold in so much disdain to publicise themselves and peddle their messages.

The bottom line is that even through this economic recession cycle data-centre power growth has many-fold outstripped the growth in renewable power capacity in the utility.  So in many markets in the data-centre world it’s a ‘zero-sum game’; for every ‘clean’ Watt that the giant data-centres of the social network guys burn someone else (us) has to burn a ‘dirty’ one.  Of course we have the low-carbon power countries, e.g. Scandinavian hydro and French nuclear (:o), but a lot of data-centres still need to be built close to their clients and markets.  Maybe that’s Greenpeace’s plan – to drive all the large data-centres out of the States?

*Jevons Paradox:  William Stanley Jevons was one of the great English Victorians.  An economist, he wrote a book in 1865 entitled The Coal Question.  This book postulated that if you improve the energy efficiency of a process the overall fuel consumption tends to rise rather than fall due to increased demand for the service/product as the cost and availability improves and this became known as the Jevons Paradox.  In ICT terms this has manifested itself as simply if you make the speed faster and the cost less the result is a huge acceleration in usage.  Sometimes now referred to as the ‘rebound effect’, Jevons Paradox has proven itself true for over 150 years.