Cache = a copy of data in a fast place. Instead of hitting the database or API every time, you read from the cache. If the data is there (hit), you get a fast answer. If not (miss), you fetch, then store in the cache for next time.

Why cache? Speed and cost. Less load on the source (DB, API). Lower latency for users. Trade-off: data in the cache can be stale until it expires or is invalidated.

Cache = faster responses and less work for the backend. Trade-off: data can be stale until TTL or invalidation.

In-app = memory inside your process (e.g. a map or Redis next to the app). Fastest, but not shared across instances.

Distributed cache = Redis, Memcached. Shared across many app instances. Good for session data, API responses, or computed results.

Edge / CDN = cache at the edge (CloudFront, Fastly). Caches static assets (images, JS, CSS) and sometimes API responses. Serves from a location close to the user.

Cache when: read-heavy, same data asked often, and a bit of staleness is OK (e.g. product catalog, leaderboard).

Avoid or be careful when: data changes often, must be exact (e.g. balance after payment), or cache invalidation is hard. Set TTL (time-to-live) so old data expires; invalidate on write when you need freshness.