Cache Types and Placement

Cache types and placement explain one of the most important practical truths about caching: not all caches solve the same problem. A cache inside one application process behaves very differently from a shared distributed cache, an HTTP edge cache, or a query cache near the database. They reduce different kinds of work, fail in different ways, and require different invalidation and observability strategies.

The four lessons move from closest to farthest and from simplest to more shared forms of behavior. The first covers in-process and host-local caches. The second covers networked shared caches used by several instances or services. The third covers HTTP-facing caches such as CDNs and reverse proxies. The fourth compares application-object caches with lower-layer database and query-level caches so the reader can choose the right level of reuse.

Use this chapter when the question is not “should we cache?” but “where should the cache live?” The goal is to leave the child lessons with a clearer sense of how proximity, sharing, protocol semantics, and authority boundaries change the real cost and value of each cache type.

In this section

• In-Process and Local Memory Caches
  The fastest cache layer in most systems and the coherence risks that come from keeping separate local copies in each instance.
• Shared and Distributed Caches
  Networked caches shared by many instances and the trade-offs between coherence, latency, and operational overhead.
• CDN, Edge, and Reverse Proxy Caches
  HTTP-facing caches that sit close to users and often deliver the highest-leverage performance win on web workloads.
• Application, Database, and Query Caches
  What exactly is being cached at different layers and how object caches differ from query caches and lower database-layer reuse.