These docs are written as a sequence, not a reference manual. Each chapter builds on the one before it. Start with About — especially What is Tiny-AWS? — to understand why the project exists and what you are trying to learn. Then read Linux Fundamentals in order. Those chapters explain the kernel primitives — namespaces, cgroups, filesystems, processes — that make container isolation possible. After that, read Networking Fundamentals. Those chapters explain how packets move, how routing works, and how Linux builds virtual networks with bridges and veth pairs. Finally, read Building the Cloud. Those chapters walk through the actual tiny-aws system — every service, every file — from architecture to Kubernetes.
What to expect in each chapter
Every post follows the same shape:
A short introduction that states the problem in plain language
## sections that walk through the idea step by step
Concrete commands, diagrams, or code where they help
A closing section that connects the idea back to Tiny-AWS
You do not need to memorize everything. The goal is to understand the mechanism well enough that when you open the Tiny-AWS source code, the pieces look familiar instead of magical.
How to use the code
When a chapter mentions a Tiny-AWS file — networking.rs, sandbox.rs, and so on — open it alongside the doc. The chapters are written to match the actual implementation, not an abstract textbook version of it.
If something does not click on the first read, keep going. Many ideas only make full sense once you have seen the next two or three layers stacked on top.
Linux Fundamentals — chapters 2 through 20, in filename order
Networking Fundamentals — chapters 21 through 29, in filename order
Building the Cloud — chapters 30 through 53, in filename order
The first two sections teach the primitives. Building the Cloud is where you see them assembled into a real system. It walks through every service in tiny-aws — the registry, the agent, the scheduler, storage, messaging, networking, Lambda, IAM — then puts them together: setup, first deploy, a real multi-laptop Kokoro TTS deploy, multi-machine clusters, Kubernetes on tiny-aws, and using instances as agent sandboxes. It ends with an honest look at what is missing and why each gap is a hard engineering problem.
