Climbing the Ladder of Abstraction Towards the Cloud-to-Edge Continuum • Jonas Bonér • GOTO 2023

Explore the cloud-to-edge continuum and the role of abstraction in software development as we shift towards a one-hop architecture, where data and compute move together in a hierarchical model with multiple layers.

Key takeaways
  • The concept of abstraction is crucial in software development, allowing us to focus on the essence of the system without worrying about the low-level details.
  • The cloud-to-edge continuum is becoming increasingly important, with many platforms offering solutions to bridge the two.
  • Managing infrastructure and resources is becoming too complex for developers, and a shift towards a one-hop architecture is needed.
  • The edge requires a new approach, with data and compute moving together, but there’s no one-size-fits-all solution.
  • The actor model is a good model for the edge, but it’s not the only one; other models like Erlang’s and GoF’s have their own strengths.
  • Ubiquity is key to the edge, with devices like smartphones, cars, and even tractors being potential edge devices.
  • The industry needs to adapt to a world where data is constantly on the move, with processing happening near the data.
  • The ladder of abstraction allows developers to work at a higher level, with the platform handling the underlying complexity.
  • Options can be paralyzing, so constraints can actually be liberating.
  • The industry needs to move towards a hierarchical model with multiple layers, from cloud to edge, with each layer having its own set of requirements.
  • Data sovereignty is important, with data being stored and processed in the correct location.
  • The edge means many different things to different people, but a unified cloud-to-edge continuum is the key to success.
  • The biggest problem is that developers are expected to hack the kernel, but this needs to change.
  • We need to rethink the way we approach development, with a focus on the code and the platform handling the underlying complexity.