Posts

Showing posts with the label Spanning Tree

Can an LLM Elect an STP Root Bridge and Port Roles? A Hands-On 802.1D Test

Small, real, reproducible test run 16 November 2026 - a 4-bridge, 5-link 802.1D topology, not a benchmark claim. The deterministic ground-truth script and the model's raw reply are linked in Sources. Disclosure: this article was drafted with AI assistance; the test itself was run against a second, independent model instance so the author's own writing didn't also grade its own homework. The sixth test in this series asked a model to summarize routes correctly - this seventh test asks it to do something structurally different: elect a distributed root and assign per-port roles across four bridges, where the right answer on any one bridge depends on what every other bridge is doing. Classic 802.1D Spanning Tree Protocol has exactly three rules that decide everything - lowest bridge ID wins root, lowest root path cost wins root port, lowest root path cost wins designated port on a shared segment - and all three have a well-known tie-break (MAC address, then port/bridge ...