07 · Journal · Cloud ArchitectureVol. 10 · Q2 2026kleiotechnology.com

What we mean when we say cloud architecture.

AWS, GCP, Azure, hybrid. Reference architectures, migrations, and cost programs that survive the second budget cycle.

Habakkuk 2:2

Write the vision, and make it plain upon tables, that he may run that readeth it.

§ I — Cover concept

The context behind the article.

Journal 021
5 min
Image direction

Cloud Architecture
5 min
Article

AWS, GCP, Azure, hybrid. Reference architectures, migrations, and cost programs that survive the second budget cycle.

Why it belongs in the journal

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 in the first place.

§ II — Article

What we mean when we say cloud architecture.

A practice, not a checkbox

Cloud architecture is the part of a system that you cannot redo cheaply. Network boundaries, account structure, identity model, data residency, and the contract between platform and product teams are decided early and live for years.

We are brought in when those decisions need to be made well, or remade with as little disruption as possible.

What an engagement actually delivers

Most of our cloud architecture engagements produce three things:

  • A written architecture letter — prose, not slides — describing what we would build, what we would not, and why. Yours to keep, yours to challenge.
  • A migration or modernization path scoped to the next two quarters, not the next decade. Every step has a rollback.
  • A FinOps baseline with tagging, cost allocation, and a small set of dashboards a CFO can read without a guided tour.

Why "second budget cycle" is in the description

The first cloud bill is funded by a transformation budget. The second one comes out of operating expense, and that is where most cloud programs lose air. We design accounts, services, and data flows so that the second invoice cycle is explainable line by line.

What we do not do

We do not write a 200-page target state architecture for a system that does not yet exist. We do not recommend exotic managed services because they are interesting. We do not pretend that a multi-region active-active topology is a free lunch.

Where we work

AWS, GCP, and Azure as first-class targets. Hybrid and sovereign cloud where regulation requires it. Kubernetes when it earns its keep; plain managed services when it does not.


The deliverable that matters most is a smaller diagram than the one you started with.

§ III — Reading note

What the article is really about.

Operating tension

AWS, GCP, Azure, hybrid. Reference architectures, migrations, and cost programs that survive the second budget cycle. In practice, the hard part is usually not implementation syntax but aligning delivery, controls, and operator trust so the thing can survive contact with a real team.

Kleio view

We treat these articles as public design memos: short, opinionated, and anchored in systems that have to be bought, operated, and defended long after launch week.

§ III — Continue reading

Three adjacent articles.

Season