Lab Cloud
James Brennan and Nathan Schneider
A lab cloud is a technical, metaphorical, and social infrastructure for collaborative computing within a lab. Research labs have long had to provision computing systems by technical necessity, from massive supercomputers to tiny Raspberry Pis. A lab cloud is somewhat different. Its defining affordance is enabling users to evade the regime of corporate cloud services that present themselves as necessary and ubiquitous—challenging even the metaphor from which they take their name, as the distant, ever-present, mostly benign things floating overhead. Contemporary universities have typically acquiesced to corporate clouds as the most responsible option for achieving convenience, compliance, security, and affordability. The lab cloud, in the sense we intend, arises when a lab seeks to unsettle this regime with its socio-technical practice. The motivation for a lab cloud stems not from technical requirements, strictly, so much as from a commitment to conducting research and pedagogy at some remove from the entailments of the hegemonic clouds.
A lab cloud is an exercise in “slow computing,” where slowness refers not necessarily to processing speed so much as to an intentionality about technological choices in light of social relationships and ethical convictions (Schneider 2015). This intentionality may involve introducing user-experience “friction” of a kind that corporate systems seek to avoid, for two reasons: because alternative technologies at hand may not be as user-friendly as corporate clouds and because alternative technologies may require users to reorient their imaginations around an unfamiliar paradigm. One such paradigm a lab cloud might operate by is a “governable stack”—an assemblage of technologies that participants have the power to shape through democratic processes (Schneider 2022). Labs might also think in terms of “feminist servers” (Murray 2022), “red stacks” (Terranova 2014), and the “installable base” (Roscam Abbing 2022). But a lab cloud always exists in a condition of incompleteness. The available non-corporate tools seem forever limited in their functionality, especially when user expectations are defined externally. And since the lab cloud is susceptible to participants’ modifications, it invites their imaginations to regard it as a work in progress whose progress depends on them.
The remainder of this chapter considers a specific case, the lab cloud developed at the Media Economies Design Lab (MEDLab) at the University of Colorado Boulder.
The MEDLab Cloud
MEDLab began in 2018 as an outgrowth of the Media and Public Engagement MA program in CU Boulder’s Media Studies department. It “experiments with democratic ownership and governance in the online economy,” according to its website (colorado.edu/lab/medlab). Director Nathan Schneider developed the lab as an outgrowth of collaborations with tech startups exploring cooperative ownership models, as well as communities exploring creative governance of their technologies. The MEDLab cloud is thus not only practical infrastructure but a direct manifestation of MEDLab’s mission.
The MEDLab cloud’s structure is described in a public, online handbook wiki shared with all students who become MEDLab research fellows—generally around half a dozen at a given time. The foundation of the cloud is a platform called Cloudron, produced by a small software business whose leaders Schneider has developed a relationship with. Cloudron enables relatively easy deployment of free, open-source software packages developed by diverse communities and companies. Cloudron also manages user accounts that work across most apps. The most essential components of the MEDLab cloud are an instance of Nextcloud, which provides file-sharing and other collaborative tools, and Matrix, a chat protocol. Cloudron enables the deployment of various other apps used in MEDLab projects, such as WordPress (for public websites), DokuWiki (for the handbook), and Listmonk (for newsletters).
Some tools MEDLab regularly relies on are not contained in the lab cloud. MEDLab’s main public website, for instance, is located on CU Boulder’s colorado.edu domain; there, it runs on a university-managed Drupal-based system that is optimized for meeting the university’s accessibility commitments, as well as for consistent, university-wide branding. MEDLab also uses the commercial, open-source platform GitLab to host code for its software projects; however, the lab is transitioning software projects to Gogs, a self-hosted Git platform.
Most significantly, MEDLab’s entire Cloudron system resides on the servers of the commercial hosting company DigitalOcean; the lab is currently exploring alternative options, including hosting through a worker-owned cooperative or managing a physical server locally, so that MEDLab participants can hear the fans and whirs of their own cloud. Students could then experience aspects of online platforms normally hidden from view, such as energy consumption, heat, and physical maintenance. A locally, physically managed system further challenges the corporate cloud metaphor, presenting a reminder that every cloud is really someone’s server.
A student experience with the MEDLab cloud
By James Brennan
I joined MEDLab as a research fellow on arrival at CU for the Media and Public Engagement Program. My first introduction to the other student lab members was at our weekly meetings hosted in the Media Archeology Lab. The room presented not just new faces but also a museum of working old computers. Cluttered tech adorning the shelves evoked the legacy of hackers which we—as aspiring researchers and hackers ourselves—stood to inherit. We then introduced ourselves and the projects we were working on in the lab. The activities we shared were as varied as our backgrounds, from producing the lab’s monthly radio show to my own work assisting communities in adopting decentralized tools.
Both undergraduate and graduate students participate in the lab’s endeavors. Each offers their own expertise. We keep in touch through the Element messaging service on the lab cloud. It operates much like Slack with an important exception: we host it. On joining, I encountered some minor but persistent difficulties in setting up my devices against the verification and privacy measures of the service. In spite of this initial effort, it has since been easy to forget the security present in the background of our normal discussions. It is unsurprising that controversial or confidential content is not sent over the lab channels, yet this is precisely how security ought to be: ubiquitous and requiring little attention, regardless of need. That MEDLab practices what it preaches represents not only adherence to an ethical commitment, but also effective pedagogical practice.
This became more evident the following term, while I worked as a teaching assistant for the department’s Introduction to Social Media course. Leaning on MEDLab’s established cloud system, we were able to host a Mastodon instance exclusively for the class. I was responsible for ensuring the 250 students all succeeded in logging into our very own, albeit clunky, network. Rather than standard tech support emails sent into a corporate void, the face-to-face assistance not only better met students’ needs but established relationships of guidance from the outset. The alignment in a self-hosted social media network for a class on the same topic allowed the natural incorporation of practices of social media into assignments about social media. Students who chose to lean in to this experiment in self-hosting demonstrated their own unique understanding and style of network use. They decorated profiles, shared memes, and added a personal flair to what might have otherwise been a dryly academic experience.
This is one example of a lab cloud growing beyond a research purpose. Locally possessing these tools provides capacity for building richer community in countless more ways still.
James Brennan is a masters student at the University of Colorado Boulder in the Media and Public Engagement program.
Nathan Schneider is an assistant professor of media studies at the University of Colorado Boulder, where he leads the Media Economies Design Lab. His most recent book is Governable Spaces: Democratic Design for Online Life.
References
- Murray, Padmini Ray. 2022. “A ’Feminist’ Server to Help People Own Their Own Data.” The Bastion, August 12, 2022. https://thebastion.co.in/politics-and/tech/a-feminist-server-to-help-people-own-their-own-data/.
- Roscam Abbing, Roel. 2022. “On Cultivating the Installable Base.” In Proceedings of the Participatory Design Conference 2022—Volume 2, 203–7. PDC ’22. New York, NY, USA: Association for Computing Machinery. https://doi.org/10.1145/3537797.3537875.
- Schneider, Nathan. 2015. “The Joy of Slow Computing.” The New Republic, May 20, 2015. https://newrepublic.com/article/121832/pleasure-do-it-yourself-slow-computing.
- Schneider, Nathan. 2022. “Governable Stacks Against Digital Colonialism.” tripleC: Communication, Capitalism & Critique. Open Access Journal for a Global Sustainable Information Society 20 (1): 19–36. https://doi.org/10.31269/triplec.v20i1.1281.
- Terranova, Tiziana. 2014. “Red Stack Attack! Algorithms, Capital and the Automation of the Common.” Effimera, March 8, 2014. http://effimera.org/red-stack-attack-algorithms-capital-and-the-automation-of-the-common-di-tiziana-terranova/.