Sophie Lovell and Sebastian Schumacher on the rise and rise of the data centre for uncube magazine. “The internet forgets nothing. Everything is saved – nothing is thrown away and it all needs storing somewhere. IBM estimates that we generate some 2.5 quintillion bytes per day, that 90 per cent of the world’s stock of data is less than two years old and that this volume is currently doubling every 18 months. This exponential growth is creating an exponential storage problem as well as an exponential energy problem: where are we going to put it all and where will all the energy come from to keep our memories “alive”?
So where does all our data get kept? It may be digital but it still takes up space, and it takes up a great deal of energy too. As long ago as 2007, according to The Economist, we began experiencing a data housing crisis. The same source states that the amount of data generated surpassed the available storage space. In 2008 American households alone generated 1,200 Exabytes (1018 bytes) of data. Add industry and science output to that figure (for example: when the new Square Kilometre Array (SKA) telescope in South Africa and Australia goes online, in the next decade, astronomers expect to process ten petabytes of data from it every hour or 1015 bytes) and you start to get an idea of the size of the issue involved.
Global tech giants such as Google, Facebook, Microsoft and Amazon are all building data centres at an incredible rate to cope with the demand for digital space. Facebook alone expects to have five new centres in operation by 2017. Apple has recently announced plans for new facilities in Denmark and Ireland. And don’t forget all the other providers: Deutsche Telekom, for example, completed Germany’s biggest cloud backup centre in Saxony-Anhalt in 2014, containing 30,000 servers and which is already expanding.
So there is a quiet building boom going on of windowless, high-security homes for our collective memory – housing everything from kitten gifs to government secrets. A new architectural typology is growing with it, where the well-being of bits and bytes takes priority over living occupants. This also means a new vernacular of citizen-less urban sprawl developing as these server farm hosting facilities are placed near power supplies or where there is space to generate their own power.
This is because another major consideration when it comes to data centre location is cheap and clean energy. Our brains consume more energy and generate more heat than most of the rest of our bodies.
So it goes with servers. Data centres are responsible for an estimated two per cent of US energy consumption. According to The Guardian, each of Facebook’s data centres uses enough energy to power 30,000 US homes. Greening data storage is, therefore, a big issue. Thankfully there are signs of change in this respect. Apple’s new server farm in North Carolina, for example, uses on-site biogas fuel cells and acres of solar panels to power it and they claim that all of their server farms worldwide are now powered by renewable energy. Facebook too is building a 17,000-acre wind park to power its Fort Worth plant in Texas.
In terms of location, countries like Norway, Sweden, Iceland and Canada are also finding themselves highly desirable as low environmental temperatures can help save a fortune on cooling and local water sources mean cheap hydroelectric power. Facebook already has their first server farm in Europe in the small town of Luleå, Sweden to take advantage of this, claiming it to be “the most energy-efficient computing facility ever built” and they are currently building a second. Iceland too has two huge data centres: Verne Global and Advania Data and are investing in attracting more big companies to join them.
Finding detailed‚ reliable information about data centres is difficult since just about all the big companies are rather less than fully transparent in this area and many facilities are shrouded in secrecy. This leads to another interesting potential driver for data centre location: data sovereignty. Amazon Web Services, Microsoft and IBM have all recently announced new centres in India to serve the big local consumers there – but the question still remains: under whose jurisdiction is the data held?
This is an issue that has gained considerable traction in Europe where the European Union has raised the issue of the privacy and security of their citizen’s data on foreign servers, something that is particularly relevant in the wake of Edward Snowden’s NSA revelations.
So unless the internet can learn to forget, or we can find more efficient storage solutions, this is an exponentially messy memory storage problem that’s not going away. And it will most likely come closer to home when we decide that, despite all the promise of the world wide web creating a global community, we may actually prefer to have these giant “clouds” – anything but fluffy ones – on the ground nearby so we at least have the feeling they are under local jurisdiction.
Get used to server farms, this is an architectural typology that is going to become as common as the agricultural kind in the not too distant future.”
Read the article online at uncubemagazine.com