Programme & speakers
Programmable infrastructures offer as many challenges as they do opportunities for addressing social issues. On 2 March 2020 the TU Delft TPM organised an afternoon of lectures and discussions on Programmable Infrastructures, supported by the new TU Delft TPM AI-Lab.
The Faculty of Technology, Policy and Management and TU Delft has a long history of studying and innovating infrastructures in both public and private contexts. The objective of this event was to lever the faculty’s core strengths to more fully understand the emergence of programmable infrastructures, identify new research goals, and develop policy solutions to address the new challenges that emerge from programmability.
The following questions were explored:
- Have we, as scholars, adequately conceptualized current developments in programmable infrastructures? Have we fully grasped the socio-technical challenges ‘Infrastructure-as-a-service’ (IaaS) poses to our work? What might we be taking for granted?
- What is the role of comprehensive engineering in transforming physical infrastructures into programmable infrastructures?
- Can legal, ethical, institutional, societal and technical mechanisms ensure that programmable infrastructures serve the public interest? Do these mechanisms exist already or will they have to be developed?
Software technologies come in a form called "infrastructure-as-a-service" (IaaS). IaaS allows organizations to scale their needs for computation and data storage. They also allow them to dynamically and iteratively develop applications in platforms with dedicated programming languages. (See, for example, the figure in this overview of Microsoft Azure.
Through IaaS, implemented across physical devices and infrastructure, digital companies are expanding their reach beyond the virtual. This means that companies seek to organize, automate and "smarten" the physical world in various public domains, including mobility-as-a-service, distributed energy services, and personalized healthcare. This wave of companies proposes implementing "programmability" to ensure that all infrastructures, not just the digital ones, are built to be adaptive, upgradable, and resilient. The term we coin here –"programmable infrastructure" – refers to an operational vision, a business proposition and an empirical topic of research. When a company proposes building a programmable infrastructure they offer to rethink and reengineer existing infrastructures, usually in the name of sustainability and equity, two of the great challenges of our times.
Researchers, advocates and public professionals should be aware that programmable infrastructures offer as many challenges as they do opportunities for addressing social issues: firstly, outsourcing the control over data and the applications and the technical infrastructure reinforces dependencies upon a narrow set of technology providers with huge capitalization; secondly, this approach leads to the loss of situated forms of knowledge as infrastructure management become automated; thirdly, redefining the social welfare function as a problem of optimization (instead of policy) shields these project from democratic processes; fourthly, the standardization of software functions challenges the ability to develop, test and validate engineering specifications in an end-to-end fashion; fifthly, the approach introduces new vulnerabilities and accumulates risk in the areas of privacy, security and safety; and lastly, interoperability across infrastructures is creating a distinct convergence towards increased levels of social, economic and political control.