CAP & consistency

You have already met a version of this: with replication, a follower can be briefly behind the leader, so a read might be slightly stale. That was consistency showing up in practice. This unit gives you the vocabulary and the tradeoff behind it, which interviewers ask about often.

The core question is: when data is stored in more than one place, how fresh must a read be? Strong consistency means every read sees the latest write, always. Eventual consistency means reads might be briefly out of date, but all copies converge to the same value soon. Neither is "better", they are a tradeoff.

Heads up: CAP is a topic people love to over-formalize and argue about. We will keep it at the practical level that actually helps you in an interview: strong vs eventual, why you cannot always have strong for free, and how to choose per piece of data.