Architecture Guide
System Guide
This section is the narrative path through the repo's current architecture. Read it in order if you want the full model, or jump straight to the part you are changing.
00. System overviewPurpose, MVP goal, system shape, major boundaries, and explicit deferred decisions.01. Context and boundariesExternal systems, BEAM versus native rules, and where backend-specific logic must stop.02. OTP apps and supervisionHow the umbrella is partitioned and how supervisors are expected to isolate runtime concerns.03. Failure domainsRecovery units and restart boundaries for association, UE state, cell groups, and gateways.04. Southbound contractCanonical IR, health and capability contracts, gateway sessions, and native adapter expectations.05. ranctl action modelThe repo's single mutable control surface and how all operator flows are expected to pass through it.09. OAI runtime bridgeCurrent Docker Compose-based OAI DU bridge and how runtime evidence is gathered.12. Target-host deploymentInstall, preflight, remote execution, and evidence fetchback for real lab hosts.14. Debug and evidenceThe shortest operator path from failure to artifacts, summaries, transcripts, and fetched evidence.
Suggested reading order
- Start with 00-system-overview.
- Continue through 05-ranctl-action-model for the mutable control surface.
- Read 09-oai-du-runtime-bridge and 12-target-host-deployment if you care about real host rollout.
- Keep 14-debug-and-evidence-workflow open when running changes.