Associate Professor of Science and Technology Studies at TIK
Hit the send button on your mobile device and you are connected. But connected how and to whom, and with what assurance about the security of the connection?
With ubiquitous use of the internet and with ransomware like WannaCry, cybersecurity has become a matter of concern as to everyday data transfers on mobile phones, smart-home devices, or cloud storage. Yet, our digital infrastructure is largely deemed invisible, perceived as simply “out there” and noticeable only upon breakdown, as Leigh Star (1999) once characterized infrastructure. What then is the materiality of the digital and where is it? Let´s take a closer look at the material politics of digital technologies – from their making and supply chains, to their disposal as e-waste.
From Closed Worlds to Hyperconnectivity
Let’s start with how data flows. The first computer-to computer network technologies took shape during the Cold War, with massive state funding of large-scale research institutions. Cybernetics as a science can be traced back to this context. Before the Internet, there was the Advanced Research Projects Agency Network (ARPANET), tied in with the nuclear race and the space race between the US and the USSR. Large-scale labs, huge state funding and big machines with growing capacities for data exchange have changed scientific practice. Especially physics but also biomedicine took place as concerted, collaborative and distributed work. Many of these literally remained within closed worlds, institutionally confined, often in military research institutes.
In contrast, much of today’s knowledge production in technosciences takes place as open science, which at the same time is also defined by rapidly changing corporate actors, new techniques and digital platforms. But software studies show that apparatuses, objects and devices are subject to continued retrofitting, building on existing infrastructure, rather than something completely novel. This is visible in many software packages that still contain structures from older data sorting machines working with punchcards. New devices align with older infrastructures in a myriad of ways. What is labelled big data might not always be that new. Big data, hyperconnectivity and machine learning not only alters but also builds on calculative devices of Cold War big science and older existing infrastructures.
The Materiality of Cloud Computing
The material trail does not only include the flow and processing of data. Big data – defined often in terms of velocity, volume, variety – demand physical server space and energy. Cloud computing is very much grounded and we find data centres in the most unusual places. Deep in the mountain, highly secured, not accessible without passing a complex several step access systems – this is where our connected lives and everyday social media usage is powered, from money transactions to government data. Information from NAV, healthcare and electricity systems, banking data are stored in these data centres. Thus, such data infrastructures are present in our everyday lives, from powering public transportation and hospitals, to running our water supply and welfare systems. While central storage may protect data better than small storages, data also become more vulnerable precisely because of the many data held by one service provider.
How do data centres relate to older technological infrastructure? Take the Norwegian company Green Mountain and its two server farms – one in Rjukan, one near Stavanger. Each of them is branded as unique precisely because of their remote location in combination with a history of older infrastructures. Interestingly, these storage systems are hardly ever built from scratch – it is older infrastructure being repurposed. One data centre, on Rennesøy near Stavanger, uses the high security infrastructure of a former NATO ammunition centre. The other one is embedded in the Hydro facilities of Rjukan, Telemark. A combination of factors, including security, protection against electromagnetic pulses, remoteness, proximity to data nodes and connection to fiberlines are listed as advantages of the site. Moreover, the data centre near Stavanger is marketed as “the world’s greenest data centre”. With the energy consumption of servers and mining of rare earths, human internet activity has a major impact on environments. Yet, its green label is due to “free cooling” from the fjord. Operating the data centre means heating up the fjord environment but paradoxically, due to no carbon emission, it is valuated as not contributing to global warming.
Hence digitalisation and the Internet – often referred to as “virtual space” do have a materiality. The example of data centres shows that new infrastructures do not come from nowhere, they are situated – and this not only applies to their regulation that might differ across countries. The very materiality of digital infrastructure, its energy use and security features are translated into assets on a global market of data services competing for customers. After all, hyperconnectivity also is accompanied by disconnections. In our accounts, often the cybersphere remains disconnected from its materiality – its making, the politics of supply chains, and disposal of devices as e-waste. The latter, for instance, implies heavy metals and persistent organic pollutants – a complex mix of legacy pollutants and emerging contaminants; their regulation and monitoring is only at the beginning.
The STS toolbox can help bring to the fore the material politics of digital infrastructures, by following the flows of data and the politics of infrastructuring. Taking material circulations, repurposing and retrofitting as point of departure in our accounts of the digital may enable us to ask new questions and participate or intervene in politics of infrastructuring.