рабочий вариант, но скороть 10 МБит

docs/architecture/FABRIC_AREA_AND_PEER_STABILITY_MODEL.md

@@ -6,6 +6,10 @@ This document replaces the oversimplified rule "every node must keep 3
 connections" with a stability model based on failure domains ("areas"),
 multi-path reachability, and live peer memory.
 
+It operates at the `Fabric Transport` layer. Services above the transport must
+consume service channels and must not directly reason about peer topology. See
+[FABRIC_SERVICE_OVER_TRANSPORT_MODEL.md](\\nas\\MST\\codex\\rdp-proxy\\docs\\architecture\\FABRIC_SERVICE_OVER_TRANSPORT_MODEL.md).
+
 ## 1. Why the old "3 connections" rule is not enough
 
 A raw connection count is too weak as a resilience rule.
@@ -43,6 +47,9 @@ An area can be derived from:
 The area label must be part of live node metadata and endpoint candidate
 metadata.
 
+For the current fleet, area assignment should be explicit operator metadata, not
+an inference hidden only inside routing code.
+
 ## 3. Stability objective
 
 Each node should maintain a working peer set with diversity, not just count.