Testing a P2P Network in a Bottle: the Docker Cluster
Roadmap note — open work items #108, #66, #131, #132, #133
In 2026 the project delivered something its predecessors only simulated: real multi-node LibreSocial clusters in containers (Docker/Podman, documented in the archive's "LibreSocial Container Setup" manual). The next step is to make that setup a testing instrument rather than a demo.
Open work:
- Distributed testing: master/slave testing setup in Docker containers (#108, high priority): run the framework's Tester against a live cluster — one driver node, N peers — and collect results the way Liu (2013) and Dlikman (2018) did with virtual networks, but on real JVMs and real TCP. This closes the loop between simulation and reality.
- Execution environment for 6 LibreSocial instances (#66): a reproducible lab recipe (compose file + resource notes) so any student can stand up a six-peer world in minutes.
- Dynamically handle wait time for display startup (#131), audio error message on noVNC startup (#132), "Potential Security Risk Ahead" warning (#133): polish the self-service experience — each of these is a papercut that currently makes first-timers hesitate.
Why care: LibreSocial's claims (decentralization, self-organization, monitoring) can only be shown, not argued. A one-command cluster plus scripted tests turns every thesis question — "what happens when 30% of storage nodes churn?" — into an afternoon experiment.