API First Government
4 likes1,917 views
The document discusses the need for Estonia's government systems to shift from a traditional "box and jar" approach to a more modern API-first approach to better serve citizens. It outlines some of the challenges governments face in making this transition, including legacy systems, poor mobile usability, and the need to support third-party integration through well-defined APIs. The document proposes adopting an API-first policy, treating open data as APIs, and developing prototypes and reference architectures to help shift thinking and enable this new approach in a practical way through small, initial steps.
API First Government
- 1. API First Government Re-thinking an age-old system Andres Kütt Information System Authority, architect 8. . 6
- 2. Contents of today • The problem • Estonian approach to solving it • Things that need to be done • What are we actually doing
- 3. Our challenges are the ones of any incumbent We are measured against the standards set by the new economy organisations while dealing with a lot of legacy of all kinds
- 4. The problem
- 5. The cloud takes a different kind of information system The standard Estonian approach of a box and a jar can be deployed in cloud but gains no benefit. We’ll get all the cost and risk for nothing.
- 6. Our mobile usability is generally horrible The UI of our apps is tightly coupled to the business logic. Extending or duplicating it is typically not possible
- 7. Shift from building software to supporting an ecosystem Turns out it matters very little what your software does if people don’t use it
- 8. People expect systems that work together seamlessly From Google to IFTTT, from fitness to financial industries, companies expose meaningful APIs. Why not the government?
- 10. Adopt the API First policy Shift focus from providing things to enabling things
- 11. Estonia is a flexible and secure platform for providing services to the citizen
- 12. From bolt-on open data to integral open APIs Open data must change and become part of solution rather than being part of the problem
- 13. From box-and-jar systems to PaaS-embracing networks Cloud is not a strict prerequisite but helps drive the change in thinking
- 14. From singular UI to a flexible multitude of UX Not only mobile but third party integrations, mashups etc.
- 15. From opaque blocks of functionality to well-defined manageable services This makes it so much easier to catalogue, understand and measure
- 16. Things that need to get done
- 17. Take control of the architecture This assumes competences, a thick skin and support from higher up. Developers are pushed beyond their comfort zones and will charge more
- 18. Build an authorisation solution à la MyData Otherwise third parties would not be able to utilise APIs handling sensitive data
- 19. Invest into data protection, audit and fraud detection At which point does your data stop being private? Which consumer behaviour is fraudulent?
- 20. Practical steps we take
- 21. Talk to people All of this constitutes a massive mind-shift. It takes time and effort to bring about
- 22. Treat open data as an API Documentation, testing routines, SLAs etc.
- 23. Building open data into most new systems Requiring open data by design is a feasible start. Requiring APIs accessing sensitive data not necessarily
- 24. Develop prototypes and reference architectures To gain experience and to provide tangible support to implementers
- 25. Summary
- 26. Most of these things happen anyway We can either drive them or be driven
- 27. The shift is difficult Many strong beliefs are challenged
- 28. Small practical steps can be taken There are specific things we can do to be on the driving side