More parts than elements: how databases multiply

Databases can be understood as the pre-eminent contemporary doing, organising, configuring and performing thing-and-people multiples in sets. Participation, belonging, inclusion and membership: many of the relations that make up the materialsocial life of people and things can be formally apprehended in terms of set-like multiples rendered as datasets. Mid-twentieth century database design derived different ways of gathering, sorting, ordering, and searching data from mathematical set theory. The dominant database architecture, the relational database management system (RDMS), can be seen as a specific technological enactment of the mathematics of set theory. More recent developments such as grids, clouds and other data-intensive architectures apprehend ever greater quantities of data. Arguably, in emerging data architectures, databases themselves are subsumed by even more intensive set-like operations. Whole databases undergo set-like transformations as they are aggregated, divided, filtered and sorted. At least at a theoretical level, the mathematics of set theory, as philosophically rendered by Alain Badiou, can also suggest some explanations of how multiples expand, ramify and split in databases. Badiou's account locates forms of instability or incoherence inherent in any set-based doing of multiples in the relation between parts and elements, between inclusion and belonging. Against the grain of Badiou's resolutely philosophical project, a set-theoretical account of databases might also point towards some vertiginous elements that elude regulation, norms and representation. In July 2007, Google Corporation switched its principal service, Google Search, over to a reindexed version of the web based on a different programming technique and database infrastructure called 'MapReduce.' That mostly undocumented switch to a slightly differently indexed web sent only mild tremors through the web (Shankland 2008). Google's MapReduce approach (Dean 2006), itself a topic of some controversy in the database design and