We’ve already said that we believe the absolute baseline for data centre infrastructure is for it to be powered 100% by renewable energy. The other two sides of the sustainability triangle are efficiency, and innovation.
Efficiency in the way energy is used in the data centre environment, and what is done with the waste energy generated by a data centre’s operation. Innovation in how a data centre facility is powered (to address, amongst other things, reliance on diesel back-up) and how data centre sites can become net contributors to their localities’ energy requirements.
Renewable power is non-negotiable. The price of generating renewable energy is decreasing – specifically wind power, although a similar pattern must be expected in relation to tidal and solar power – and, in countries like Ireland which boasts more than 3,000 kilometres of coastline, there’s ample space for the infrastructure.
All that remains to make renewable power a preferred solution is constant demand – such as that provided by data centre infrastructure – so that that generator downtime (for lack of anywhere to put the energy) is minimised.
Again, in a country like Ireland which simply isn’t big enough to cause any meaningful latency issues, data centre infrastructure can be located in proximity to the source of renewable power, with a direct line connection (something that is being worked on under the terms of the Irish government’s Climate Action Plan 2019.)
Given the demand for power by technology, some have talked about the takeover of the renewables sector by Big Tech, citing a recent visit by Google’s CEO, Sundar Pichai, to Helsinki to unveil a $2bn (£1.6bn) funding deal for 18 power projects across three continents, including solar arrays in Chile and wind farms in South Carolina. More germane, perhaps, were the two wind farms for a soon-to-be-expanded data centre in Finland.
Efficiency of energy use (power usage effectiveness or PUE) is another fundamental whose time has come. Data centre infrastructure is a given – the modern world, reliant on internet-enabled devices and ecommerce and smart tech, cannot exist without it.
The predictions show that the dataverse (the amount of data created and copied each year) will increase fivefold by 2025. There will be as many as ten smart devices for every man, woman and child on the planet within the same timeframe.
All of this has to be squared with the needs of the planet and the very real – and very immediate – threat of climate change, and there will be ever-increasing scrutiny of who uses energy, where it comes from, where it is used and – of course – how efficiently it is used.
Increasing PUE values for data centre infrastructure is a given, whether it’s through more efficient cooling, better processing and storage tech or via solutions still in research and development.
The square on the third side is occupied by innovation – in tech and in ways of operation – that reduces the impact of data centre infrastructure on its host locality, its host economy and on global climate change and allows it to give back, mitigating against any concerns that may be voiced around scale and security.
Under this heading – and amongst many other initiatives and projects – there’s fuel cell power generation, using gas-fed cells. There’s direct line power provision, as mentioned earlier, allowing data centres to meet their power needs directly from renewable sources, without troubling the grid infrastructure of the country in which they are located.
There’s recycling waste heat for use in surrounding communities and there’s ancillary power generation to do away with diesel backup, to feed back into the grid and to make a contribution to the power needs of the host nation.
Clearly, the data centre industry is still some sizeable way away from achieving all this. Steps are being made, we’re moving in the right direction and the software, hardware and infrastructure needed to make further progress are both evolving and (in some areas) coming down in price. The elephant in the room is carbon neutrality.
Arguably, data centre infrastructure cannot be carbon neutral until every piece in its jigsaw is accounted for – and this will include the construction methods, management, logistics and materials employed in creating the infrastructure. And while we pursue the goal tenaciously, we need to be realistic and understand that, in the interim, the way forward for every industry sector, on a global basis, must feature some form of carbon taxation.
A recent article in the MIT Technology Review explains how a new model suggests that carbon taxes need to be more aggressive than has so far been proposed (in the US) – at a higher cost level and increasing every year in the short term.
According to the article, which quotes Gernot Wagner, associate professor at New York University’s Department of Environmental Studies, ‘the higher the price is at the start, the more rapidly nations and businesses will develop and deploy cleaner ways of doing things.’
It would also mean that the price on carbon can start to decline sooner, as more and more of the world’s economies – and industry sectors – shift to systems and sources that no longer pump out emissions, like those we’ve talked about here. But in addition, and importantly, the world will simply learn more about the exact impacts of climate change, and what it takes to address them.
A new model created by Wagner, Columbia University economist Kent Daniel and Robert Litterman, a former head at Goldman Sachs says that the cost of not implementing carbon tax rises at a staggering rate for every day it’s delayed.
If we wait a year to implement an effective carbon tax, the estimated cost of additional climate-change impacts will reach approximately $1 trillion. If we wait five years, that swells to $24 trillion. A 10-year delay could cost the world $100 trillion.
It’s not something we can, or should, ignore as part of our drive for a truly sustainable future for data centre infrastructure.