Assess students' ability to design a Byzantine fault‑tolerant consensus protocol for geo‑distributed data centers under partial synchrony, focusing on specifying system assumptions (n, f, network model), choosing quorum and replica placement strategies, selecting leader-based or leaderless architectures, and applying optimizations for latency and bandwidth (batching, pipelining, erasure coding, optimistic fast paths). Expect concise protocol sketches, correctness arguments for safety and liveness under partial synchrony, analysis of message and bandwidth complexity, and an evaluation plan for geo‑distributed deployments and failure scenarios.