Assess graduate students' ability to analyze and compare consistency models (e.g., linearizability, sequential, causal, eventual), latency, and fault‑tolerance trade-offs in geo‑distributed replicated databases under network partitions. Scope includes CAP/PACELC reasoning, consensus and quorum protocols (Paxos/Raft, quorum reads/writes), replication strategies, failure models, and metrics (latency, availability, staleness). Questions should require quantitative scenario analysis, justification of protocol/configuration choices, and discussion of mitigations and limits.