Consistency specifies what data reads may observe after writes, making it a central trade-off in distributed systems and a common interview topic. Under CAP, consistency means every read receives the latest write or an error; partition tolerance forces a choice between consistency and availability during network failures. section 6.2
Use these distinctions when analyzing a distributed architecture:
- Strong consistency provides the latest value or an error, while eventual consistency allows replicas to converge over time. section 6.2
- CP systems such as HBase prioritize consistency during partitions; AP systems such as DynamoDB prioritize availability and typically provide eventual consistency. section 6.2
Interview scenario
When a network partition separates replicas, ask whether the system should reject or delay reads and writes to preserve consistency, or continue serving potentially stale data to preserve availability. Justify the choice against the application’s requirements. section 6.2



























