Architecture

This is the system-map family of the Explanation quadrant: it describes how plexsphere fits together at a level above any single bounded context. Read these pages when you need a mental model for the whole platform — control plane vs. data plane, trust boundaries, where state lives — rather than the contract of a specific subsystem. Like all explanation, it discusses why, not how; for the task itself see how-to, for the bare contract see reference.

What is here

• Positioning — the one-picture view: an ecosystem of services on top, plexsphere as the sovereign control plane in the middle, and any cloud, on-prem, or edge substrate at the bottom. The engineering-facing companion to the landing-page diagram, with the capability-to-context map.
• Architecture overview — the two-binary control-plane runtime, the shared platform packages, and the trust model that connects them.
• Storage topology — which datastores hold which rows, how Postgres and SpiceDB stay aligned, and the per-Domain residency rules that constrain every read and write.
• Signed event bus roadmap — where the streaming substrate is today and where it is going.

When to read this section

• You are planning a change that crosses a bounded-context boundary and want to know which seam to use.
• You are onboarding and need a single map of the system before diving into individual contexts.
• You are debugging an integration issue and need to confirm which process owns which piece of state.

For the per-context invariants, ubiquitous language, and aggregate boundaries, see the contexts quadrant. For the operator-facing how-to guides, see how-to.