The Materiality of Big Data Technologies

The next speaker in this ICA 2018 session is Zane Cooper, whose interest is in the material constitution of big data. Big data make use of earth and labour that do not easily track with its digital manifestations: they generate a long supply chain of physical hardware that supports the big data cloud. There is therefore a need to distinguish between what big data infrastructures are (their constitutional logic) and what they do (their operational logic).

The metaphor of contamination might be valuable here. Contamination is collaboration in the service of ‘precarious survival’; it is a building and a working across difference. Contaminated diversity is a tangled, morally ambiguous interdependence.

Cloud data storage is controlled by a handful of major companies. But these companies themselves depend on hard drive manufacturing, which is itself a highly concentrated industry. These themselves depend on permanent magnet manufacturers, which rely on rare earths mainly mined and processed in China. The foundational, functional unit of the cloud data centre is the 3.5” Winchester hard drive, a technology that has existed since 1986 and persists even in spite of the considerable growth in solid state storage.

Such hardware form factors are persistent because of the specific magnet technology that is used in these devices. In the late 70s, conflict in Zaire rendered previous magnet technologies inaccessible, and a new magnet technology entered the market a few years later; it has been the dominant technology since the late 1990s. Such magnets rely on rare earths, and such magnet manufacturing – part of it was previously done in the U.S. – subsequently moved almost entirely to China: its marketshare grew from 46% in 1994 to 97% in 2013.

These developments took place before the rise of big data to public attention, even though there were already concerns at the time. Now, it has become even more crucial to untangle these complex connections between materiality and big data.