Databases allow to store and access data, and to keep data independent from the software applications (i.e. the same set of data ideally can be used with different software applications). Think of DBMS as an operating system for data. DBMS structure data. For example, in relational database systems the structure involves tables, which again consist of table columns, which describe attributes, and table rows, which comprise of a single data set of a certain entity. DBMS also provide the means to restrict access through a login.


Participants of a database dérive, initiated by the author to discuss database infrastructure.

While DBMS may appear to the user like a black box, because one can only perceive its input and output, an infrastructural perspective may give us a more complex idea of what’s happening. Media scholar Lisa Parks, following Susan Star and Geoffrey Bowker  (1999) proposed to use infrastructural inversion as a strategy to explore those parts of an infrastructure that are visible to us. Parks argues, that infrastructure is often of high complexity, that even with full access to all resources it would be difficult to describe an infrastructural setting in its entirety and to make it comprehensible.

Rather, Infrastructural inversion tries to deduct its underlying classification systems and standards from a small portion of the infrastructural complex. One example that comes into my mind for databases is the project Other People also Bought by artist Sebastian Schmieg and Jonas Lund (2013). They wrote a script that would pick the automated shopping recommendations that Amazon pulls from its database based on user behavior. The artists’ script then adds one of these recommended products to the shopping cart, thus altering not only the user data but also triggering another round of recommendations. With this playful project not only do Schmieg and Lund explore the recommendation algorithms, but in a deeper sense expose the particularities of data storage and the structuring of data. They address the very existence of a new kind of data – the meta-data that was recorded from user interactions. These recordings create a new form of data based subjectivity, a particular data body of the particular user, who in most cases is forced into producing his/her own transactional data body.

Further, I remember more artists works, worth being discussed from a perspective of database infrastructure – classics such as Lisa Jevbratts 1:1 (1999), Heath Buntings The Status Project(since 2004), Paolo Cirio’s Amazon Noir  (2007/08), and Natalie Bookchins Testament (2009), all of whom either created databases for their own use or explored existing ones. Although it was not their direct aim to address the database as infrastructure, and although they have been interpreted mostly in the context of, a re-reading from a database perspective could reveal more about the underlying politics of databases.


Lisa Jevbratt 1:1. Each pixel represents one IP Address as of 2001. A mouse click on the pixel leads the browser to the respective IP address

Furthermore, non-artistic examples for infrastructural inversion of databases exist as well: For instance one could explore the popular WordPress content management system that runs on a MySQL database backbone using the plugin Query Monitor. It makes visible each database call ‘inside’ the WordPress system and gives us an idea what might be stored in database tables such as wp_posts, wp_comments or wp_users.

Other, more limited means to explore DBMS are APIs, which are sets of commands to access the – so to say – public part of large DBMS of companies such as Google , Facebook, Twitter, and so on. However, this seemingly open access is restricted in its own way. Exploration is only open to a certain extent that fosters, but does not jeopardize, corporate exploitation of data (cf. Taina Bucher).

While accessing databases through their interfaces forms one area of contesting DBMS, another form is infrastructural tourism (Shannon Mattern), which means actually visiting the material parts of a certain infrastructure. Boaz Levin and Ryan Jefferey have done so, when discussing the ideology behind the “Cloud” in their video essay All That is Solid Melts Into Data (2015) and Timo Arnall visited and filmed the Telefonica Data Center in Spain for his film Internet Machine (2014). Both films trace the physical layers of network infrastructure and – in my reading – also refer to infrastructural inversion.

We can only speculate whether analogous to “food porn” a genre called “infrastructure porn” will develop in the near future. Actually, there is a whole film genre that could be called “infrastructure porn”, but for reasons of deference it is labeled disaster movies. Popular culture thus offers a reading of infrastructure in the moment of breakdown, which is the moment where infrastructure gets most visible. Disaster movies allow a society to confronts its inner, unspoken fears and phantasms of losing its infrastructural base. Manhatten for instance is flooded and frozen in The Day After Tomorrow (2004), a large part of infrastructure gets destroyed, while just the The New York Public Library remains a survivors’ hideaway.

It is in the moment of breakdown, that we are confronted with the black-boxed database management systems. But can we really look inside these systems? I doubt it. Therefore let’s return to infrastructural inversion and to ‘superficially’ investigating the visible ends of DBMS. I have identified visual signs that may indicate the existence of a database and tested them through an ongoing series of database dérives. Visible indication of a DBMS may be given trough:   1.) the form field for data input, search queries and updating data;   2.) the list view or the presence of structured data;   3.) the use of abstract organizing principles, i.e. ID-numbers;  4.)  a login that allows for gradual access levels and individualization;  5.) the recording of meta data of user interactions such as change logs, i.e. in Wikipedia it records different article versions and the editors’ user names;  6.) external data access through a standardized interface, the API.

This list is by no means complete, but it may give a first idea of what I mean with infrastructural inversion. And although this list provides us with clues about the structural aspects of database infrastructure, it does not include infrastructure as capital investment, infrastructure as something that is learned through membership, or that is stretched in time. Processes of establishing and standardization, of conflict resolution, maintenance practices, governance and control, as well as the everyday use from a user’s perspective can only vaguely or not at all be addressed with this approach.

Another infrastructural area of DBMS that could be considered and has not been addressed yet, comprises of the production process and its producers (often male engineers). If we look for any access points, where the un-black-boxing could take place, we may end up with the producers and managers who heavily influence the structure of data storage and who have yet to be recognized as addressees for political claims. We may also end up with legislation, regulation and standardization, yet another field that is highly unattractive for artists. There may, however, be hope: The Political Spectrum (2008) was a collaborative mapping effort initiated by artist Julian Priest to discuss the changing regulation of airwaves. It is a beautiful project that demonstrates the potentials and restrictions of artistic work in the infrastructural field.


Julian Priests participatory work The Political Spectrum asked visitors to add their own comments to a board where he described regulation for the available air wave frequency spectrum per country.