What can eduroam teach us about building research infrastructure
Eduroam is a fantastic piece of academic infrastructure: students/researchers from thousands of universities around the world can automatically connect to WiFi and any partner institutions using login details from their home institution. To me it's surprising that it exists, given that it has many characteristics of projects which academia is terrible at accomplishing:
- Not academically "interesting" (except possibly for some technical challenges at the very beginning)
- Output is not a research paper, instead a piece of infrastructure
- Requires perpetual upkeep/maintenance
- Needs to meet high reliability standards and be fixed quickly
How did Eduroam get built?
Summarizing wikipedia and a few other pages, it looks like eduroam started with an idea to connect different national networks in Europe, and was incrementally adopted by other countries throughout the 1990s and 2000s. Adoption was relatively easy because:
- It did not require opting into a centralized login procedure: universities could keep their existing independent authentications.
- Universities already had dedicated IT teams to provide WiFi in their buildings (so there was a team in place to handle it).
- There was already a lot of national infrastructure to connect universities within individual countries.
- Academics (presumably) did not oppose it in any way; in fact it would generally make their lives easier.
What lessons can be drawn for doing large projects in an academic context?
A lot of things which are obvious when you read them:
- Things are easier when infrastructure funding is already available outside of the grant system: that removes time limits for project completion and allows permanent employees to be hired to maintain it.
- It is easier to link up teams already building similar infrastructure rather than building teams from scratch.
- It is best when few (if any) people oppose your project.