This is a waiting room implementation as part of my Bachelor's thesis in Computing Science. It is currently far from done, and I would not recommend using it in its current state as changes are still being made fast and often.
This contains some things that should happen soon™. These are all things I intend to do before the end of my thesis, and they're roughly in order of priority. I will split them up into smaller tasks as I go along.
- Finish writing the distributed waiting room simulation:
- Make the letting out of users trigger when it hasn't happened yet and a new root is selected.
- Implement Round-Robin Probe Target Selection.
- Better document the recently added code
- Detect node failures automatically (using simplified SWIM)
- Add recovery from node failures
- Write automated deterministic testing system (TODO: Split)
- Write benchmarking skeleton and run benchmarks (TODO: Split)
The following are things that I will likely not do before finishing my thesis, as I am focussing on a simulation only for now. They are here so I don't forget about them. I do intend to do them at some point, but only after my thesis is done.
- Document metrics and move them out of to
waitingroom-metricscrate - Move settings parsing with foundation out of
waitingroom-coreso the waiting room can be used without foundation - Set up docker container images to make running prometheus and grafana for the dashboard easier
- Re-make parts (most) of
waitingroom-httpto make the code more self-documenting and overall better - Add cross-node message passing to
waitingroom-httpto make distributed implementation work - Make
waitingroom-httpwork with the distributed waiting room - Write a proper readme with usage instructions etc.