§ JOURNAL · CLOUD COMPUTINGJournal 013

A good cloud architecture is often smaller than expected.

Many teams do not need more services. They need fewer boxes, clearer ownership, and cloud infrastructure that can still be understood before lunch.

READ
5 min
§ I — COVER

The context behind the article.

Journal 0135 min
VISUAL DIRECTION

Many teams do not need more services. They need fewer boxes, clearer ownership, and cloud infrastructure that can still be understood before lunch.

BRIEF

Many teams do not need more services. They need fewer boxes, clearer ownership, and cloud infrastructure that can still be understood before lunch.

WHY IT MATTERS

This entry exists to make the operating logic visible: not just the system we would build, but the constraint, tradeoff, or failure mode that forced the architecture to matter.

§ II — ARTICLE

The text.

Smaller is often better

The default instinct in cloud architecture is to split: more services, more containers, more independent deployments. But splitting has costs that are easy to underestimate and hard to reverse.

A well-structured monolith — or a small set of well-bounded services — is often more reliable, easier to operate, and cheaper to run than a distributed system that was split too early.

When to stay together

  • One team owns the code: splitting adds network calls without organizational benefit
  • Data is tightly coupled: if two services need a shared transaction, they probably should not be separate services
  • The team is small: distributed systems require distributed expertise

The modular monolith

The best of both worlds is often a modular monolith: a single deployable unit with clear internal boundaries. Clean code boundaries without network overhead. Simple deployment without orchestration complexity.


Do not split your architecture because you might need to scale. Split it because you have proven that you need to scale different parts independently.

Season