Eventual consistency allows replicas to return stale data temporarily while guaranteeing convergence to the latest state after writes stop. It trades strict correctness for availability, partition tolerance, resilience, and performance. section 6.2
Use eventual consistency when temporary inconsistency is acceptable:
- Choose it for highly available distributed systems such as DynamoDB, Cassandra, or DNS. section 6.2
- Contrast it with strong consistency, which returns the latest write or an error and suits financial data. section 6.2
- Explain the choice through CAP trade-offs: partition tolerance remains necessary, leaving consistency and availability to balance. section 6.2
Interview answer
When asked about consistency models, explain that replicas may temporarily disagree under eventual consistency but converge without new writes. Contrast this with strong consistency, then justify eventual consistency for systems where uptime matters more than immediate freshness. section 6.2









