Data Centers as
Instruments of the Cloud's Territorial Expansion
global network of corporate data centers
Google Facebook Apple Microsoft Amazon
One of the most visible footprints of the extended geography of the cloud are the massive data factories and warehouses that dot the globe. Data centers have emerged as the quintessential building type of the information economy and have become the subject of fascination and study for artists, architects, social anthropologists and media scholars, among others. If the factory was the architectural embodiment of industrialization, and the downtown tower stood in for finance economy, the data center materializes the contemporary data economy. However, the significance of data centers goes well beyond the formal and the symbolic. Data centers can also be conceptualized as instruments of territorial expansion critical to the extension of the operations of the cloud
Our increasing dependence on mobile devices and the expansion of sensors and extractive data devices in everyday spaces have translated in massive amounts of data generated on a daily basis, and in turn in a continually growing global network of data centers. In 2018 Cisco’s Global Cloud Index forecasted that the amount of annual global datacenter traffic will grow three-fold from 2016 to 2021, reaching 20.6 Zettabytes per year by the end of 2021. There is a reciprocal relationship between the growth of data centers as storage and processing epicenters of the cloud and the massive rise of personal mobile computing and the extension of computing capacities to everyday spaces and infrastructures of urbanization. Hence, the importance that data centers hold as instruments of expansion within the extended geographies of the cloud should be obvious.
Within this expansionist context, reliability is an important factor. Building up redundancies in the network through creation of multiple data regions and zones is an essential part of making a cloud platform reliable. This is also a spatial project, as regions and zones of the cloud are materialized through data centers which are in turn grounded in the specificities of geography and location. North Carolina is among states in the US that have lobbied heavily for data center investments. The presence of major data centers by the like of Apple, Google, and Facebook has been followed by other tech companies who have contributed to the build up of a data center geography in the state.
Locating North Carolina's data center geography in space and time
Locating the physical infrastructure of the cloud, like data centers, entails a spatial logic based on a geopolitical cocktail of economic, political, and environmental incentives. Specifically, the locative strategy of data centers is based on a set of gold standards. Whether a region is earthquake-prone, bound to political instability, or climatically extreme, will be initial decisive factors. Energy reliability, good fiber optic connectivity, proximity to major target markets, low energy cost, access to plenty of water, plenty of inexpensive land, climate, and tax incentives each play an important role in where a cloud provider decides to plant its data centers. In locating its nodes, the cloud has followed in the footprints of previous rounds of capitalist spatial production. Cloud firms often locate themselves on top of the ghosts of the manufacturing economy or within the spatial vestiges of the military industrial complex of the Cold War. This is not solely a reuse project however. The infrastructural and security surplus built into these leftover spaces means significant savings for datacenter companies.
North Carolina's data centers have grown rapidly since 2007 when Google (A) erected one of its first purpose built data centers in Lenoir. The move was followed by Facebook (B) and Apple (C) within a couple of years.
As much as the spatial projects of the cloud are often touted as kick-starters for economic regeneration, these territorial projects tend to be more opportunistic, making decisions based on the historical build-up of infrastructural capacities as well as the political willingness of local governments. The situation creates a power relationship tipped heavily in favor of cloud firms. Desperate states and local governments competing to lure data center investments offer cloud companies enticing incentive packages designed specifically for them. The resultant geography is one in which the cloud is grounded in the economically-weathered small communities that are willing to play with its rules.
Given the common spatial logic of their locational strategies, data centers tend to cluster together. Above: 1) Cluster of data centers outside of Charlotte, NC; 2) Google's data center in Lenoir, NC; 3) Facebook's data center in Forest City, NC;
4) Apple's data center in Maiden, NC.
In addition to political and economic willingness, the locational strategies of data centers mean that the local geography (vegetation, topography, etc.), infrastructural capacities (electricity, water, communication, etc.), and the local built environment, each play a part in deciding where and how a data center is to be located.
The physical layers of landscape, infrastructure, and the built environment directly contribute to the locational strategies of data centers
Beyond its locative strategies, a typical data center's defensive spatial strategies use topography, vegetation, and hard barriers like fences and walls to cut it off from its immediate context while maintaining heavy dependence on regional infrastructures of power, water, and data. Space and its construction play a critical role in grounding the cloud.
topographic separation and land forming
vegetation buffers and defensive landscape design
fences, walls, and hard barriers