DISTRIBUTED CONSENSUS
Byzantine fault tolerance in asynchronous networks requires a minimum of 3f+1 nodes to withstand f simultaneous failures. The impossibility result of Fischer, Lynch, and Paterson demonstrates that no deterministic algorithm can guarantee consensus in an asynchronous system with even one faulty process.
Recent advances in randomized protocols offer probabilistic termination guarantees that circumvent the FLP impossibility, trading determinism for liveness under partial synchrony assumptions.