Fault-Tolerant Distributed Minesweeper

   page       attach   
sommario

We built a cooperative, distributed Minesweeper game. But instead of relying on an external database to keep things in sync, we designed an in-memory Replicated State Machine (SMR) from scratch. Using pure Python TCP sockets, the architecture enforces Strong Consistency across a three-node cluster through a two-phase PREPARE/COMMIT chain replication protocol. If the leader dies, the cluster recovers automatically. We fully containerized the entire stack—even the graphical Tkinter client—proving that complex fault tolerance and deterministic state can work without heavy-weight dependencies.

prodotti