Explaining Co-Develop’s Understanding of Digital Public Infrastructure

by David Eaves and Jordan Sandman

Co-Develop recently released a framework for how we understand digital public infrastructure (DPI). If you haven’t had a chance to read it, please take a look. As with all of our work, we welcome your comments and feedback as we co-create this work alongside a dynamic ecosystem of partners. 

Why did we write a framework for digital public infrastructure?

Definitions of digital public infrastructure have been offered by several organizations in the DPI space. As Co-Develop begins to engage with potential partners and make grants, we wanted to transparently state our understanding of DPI so partners can easily and confidently understand how to intersect with our strategy. We also hope our definition clarifies what DPI is for newcomers to the space and helps them understand its intellectual underpinnings. 

To what degree did we incorporate or differ from other organization’s work?

Existing definitions that we benchmarked include those from GovStack, the World Bank, and EkStep Foundation. We would not suggest that our definition should be used by everyone in the space. Instead, we released it to contribute to an ongoing discussion that we hope will continue in the future. We submit that there does not need to be immediate convergence on a single definition of DPI. When convergence does happen, it will likely result from country implementers who use a common definition for policy and planning coalescing on a version they understand. 

How is our framework different from others’ work? 

Our definition is influenced by the work of Brett Frischmann, who emphasizes the functional role of infrastructure – meaning the real-world functions it offers rather than its specific technical components. Consequently, our piece embeds DPI in its socio-political context – that is, we attempt to describe the conditions that have to be present to make a digital capability “public infrastructure” acknowledging that those conditions are perpetually changing as technologies, economies, and political philosophies evolve – rather than a static definition of technology components that are considered to be DPI at this current moment. 

This is also a reaction to some challenges around two distinct definitions of “infrastructure” that converge in this context. Specifically, the difference between infrastructure as it is typically defined in the tech field (any foundational software that underpins other applications) and infrastructure as it is defined in public policy (essential social functions that should be guaranteed to be inclusive, safe, and equitable by governments). Microservices architecture is powerful (and there is much that should be done to commoditize it - more on this another time) but defining any foundational software as DPI, as opposed to a specific, essential set of services around which a general consensus exists, risks creating a definition of infrastructure that is almost infinitely broad in its scope and thus diluting its value.

How do digital public goods (DPGs) fit into our DPI framework? How are DPGs and DPI different?

We recommend looking at the GovStack and Digital Public Goods Alliance’s work to learn more about how digital public goods can play a role in the creation of a country’s DPI (click here for the definition). One type of digital public good -  open-source software - can be adapted by a country to build DPI capabilities. Not all DPI must be built with open source code, but there are many benefits for countries of doing so, including reducing vendor lock-in and increasing transparency and adaptability.

Is this framework going to change?

Physical infrastructure – roads, bridges, etc. – has existed for thousands of years and yet the set of systems that are characterized by its definition are politically contestable, and its definition continues to evolve. We expect the definition of DPI to similarly evolve. Here the conversation is as important as the ultimate answer.

How does sectoral digital infrastructure fit into DPI?

Our definition focuses on society-wide capabilities that transect all sectors and are so essential to well functioning societies that social values should be baked into them. 

As we discussed in the definitions piece, there are some sectoral solutions like health care records, eCitizen platforms, GIS, and mobile wallets that, while may not be society wide, are both important and meet many of our criteria for DPI, but still fall outside of our immediate focus. In time, these examples may be included in our definition. For now, many of these sectoral solutions likely benefit from society-wide infrastructure like digital identity, payments, and data exchange. For example, health care records that can leverage a digital identity and data exchange system will be able to much more easily authenticate users and share sensitive information with consent.