Cache Index

The idea, in context

A cache saves work by retaining something that may be needed again. In this guide, a cache index is the lookup layer that connects a request or identity to that stored material and its reuse rules. It is a practical design term, not one universal database or industry standard. A useful record tells you more than where a value lives: it explains the context that makes the value valid. Start with that context before choosing a storage engine or optimizing the hit rate.

Identity comes first

Write down the smallest set of facts that makes two requests equivalent. A document identifier alone may be insufficient if its language, version, permissions, or output format changes the answer. Treat the key as a statement about equivalence, then test that statement against real examples.

Freshness is a decision

Choose a reason for reuse to stop: elapsed time, a new source version, an explicit invalidation event, or a change in access. Keep the clock and the policy visible in the record. An entry being present does not, by itself, mean it is safe to return.

A miss needs a clear path

Define what happens when an entry is absent, expired, or unsuitable. Recompute from the authoritative source, handle concurrent requests deliberately, and replace the index record only when the new value is complete. This makes the cache a recoverable optimization rather than an accidental system of record.