Changes in Enterprise Cloud Architecture and User Behavior on Social Media in the Face of Discomfort and Difficulties

If using social media discloses terabytes of private information and users are becoming aware of the data collection risks, why do users continue using social media? If an engineering team’s current cloud architecture is expensive and requires a lot of manual labor, why do they continue using such a system? The connection between my STS research project, about users continuing to use social media despite the privacy and autonomy concerns associated with it, and my technical project, about engineers changing the cloud architecture of their applications to reduce costs and automate application management, may not seem obvious at first glance. However, both projects inspire people to avoid quick judgment about decisions and take a deeper look into the barriers to seemingly obvious alternatives. There are barriers to almost every change in human behavior or use of technology that must be acknowledged to successfully implement change.

The technical portion of my thesis is about my software engineering internship project in which my team and I transitioned existing applications to a new cloud infrastructure and the challenges in implementation along the way. My team and I migrated applications running on a Kubernetes kOps cloud infrastructure to Amazon’s Elastic Container Service (ECS) to reduce the operational costs of running applications by $150 a month and take advantage of ECS’s ability to auto-remediate software vulnerabilities that previously involved manual resolution. To migrate applications, we had to deploy the new ECS infrastructure, refactor existing application code to be compatible with the new infrastructure, and redeploy the applications onto the new infrastructure using a repeated process. My team and I faced challenges working with one application because it used outdated software packages. The outdated packages affected our decision and ability to migrate this application, forcing us to continue running it on the older infrastructure until we fixed the compatibility issues with the new infrastructure.

The deliverable of my STS research is an improved understanding of the lack of transparency by social media companies related to the wealth of knowledge they have about users and how the current solutions to improve transparency do not cause changes in user engagement with social media. Mechanisms by social media companies to increase transparency do not remove the technological determinism view users have about being unable to escape privacy risks. Technological determinism gives a name to the ill-defined lack of autonomy users have to disengage with social media or find an alternative, which can be valuable for governments to understand when attempting to make privacy-protecting laws. The barriers to users lessening interaction with social media that cause the technological determinism view include network effects, addiction, and the lack of a viable alternative. These barriers to user disengagement encourage the continued use of social media despite the data collection risks.

The fact that my internship team could not move over one application to the new infrastructure because of software incompatibilities is similar to the fact that many social media users cannot disengage with social media because of network effects and internal pressures. If I had done these two projects separately, I would not have seen how technological change for software engineers and behavioral change for social media users are both met with resistance because of pressures in sociotechnical systems to continue acting in the same way and momentum that introduces barriers to change. Actors within sociotechnical systems must face the current barriers inhibiting change to move past them and execute successful change. My thesis project demonstrates that STS perspectives can help engineers view their projects within sociotechnical systems to identify barriers to change and increase respect for the decisions all actors within the systems make.