STP Concepts and How the Algorithm Works

That output is the shortest possible summary of a running STP: who the root is, who this switch is, and what every port is doing. Once you can read it, STP stops being mysterious. The protocol itself is a simple distributed algorithm repeated by every switch in the topology.

#The Three Steps of STP#link

1. Elect a root bridge — exactly one switch becomes the reference point for the whole topology
2. Determine the best path to the root from every other switch — this becomes each switch's root port
3. Determine which ports should forward and which should block — based on which switch is 'closest' to the root on each segment

Every decision is derived from one data structure: the BPDU. Bridge Protocol Data Units are frames sent every 2 seconds (Hello timer) by every switch, containing the sender's view of the topology. When a switch receives a BPDU with a 'better' root than its own, it updates and propagates.

#BPDU Exchange#link

BPDUs carry the sender's ID, its belief about the root, and the cost to reach that root. When a switch hears a BPDU with a lower root ID (better bridge), it accepts that as the new root and forwards its updated view. When the topology converges, every switch agrees on the same root.

#The Bridge ID and Root Election#link

The bridge ID is 8 bytes: 2 bytes of priority and 6 bytes of MAC. Lower is better. Priority defaults to 32768 but is configured in increments of 4096 — the well-known values are 0, 4096, 8192, 12288, 16384, 20480, 24576, 28672, 32768, 36864, 40960, and 45056.

#Default Timers#link

A classic STP convergence after a topology change takes 30–50 seconds: Max Age (20s) to detect, then Listening (15s), then Learning (15s), then Forwarding. That's a long outage for a user. RSTP (802.1w) cuts this to a few seconds, which is why nobody runs classic STP on new deployments.