If brief inconsistency would be catastrophic, which CAP property should you prioritize?

If brief inconsistency would be catastrophic, which CAP property should you prioritize?

• A. Consistency
• B. Availability
• C. Partition Tolerance
• D. Latency

Answer: (A)

Strong consistency is the property to prioritize when even a brief moment of inconsistent data could be catastrophic. In a distributed system, you want every read to reflect the most recent write, so all replicas agree on the same value at the same time. That eliminates scenarios where different nodes show different balances, counters, or states, which could lead to serious errors or fraud. To maintain this level of correctness, you may need to withhold or delay responses during network partitions or otherwise sacrifice some latency or availability. The trade-off is worth it when correctness is non-negotiable. The other concepts focus on responsiveness or system resilience during faults, but they don’t guarantee that all observers see the latest, identical state across all nodes, which is precisely why they’re not the best fit when inconsistency would be catastrophic.

Strong consistency is the property to prioritize when even a brief moment of inconsistent data could be catastrophic. In a distributed system, you want every read to reflect the most recent write, so all replicas agree on the same value at the same time. That eliminates scenarios where different nodes show different balances, counters, or states, which could lead to serious errors or fraud.

To maintain this level of correctness, you may need to withhold or delay responses during network partitions or otherwise sacrifice some latency or availability. The trade-off is worth it when correctness is non-negotiable.

The other concepts focus on responsiveness or system resilience during faults, but they don’t guarantee that all observers see the latest, identical state across all nodes, which is precisely why they’re not the best fit when inconsistency would be catastrophic.