When a change in the topology occurs on a spanning tree protocol activated network, it has an affect on all switches in that topology.
The switch that detects a link status change on one of its links will send out a topology change notification (TCN) BPDU towards the root bridge out of its root port.
If there is an upstream switch between the affected switch and the root bridge that upstream switch will send an acknowledge back down the link to the originating switch and forward the topology change notification out of its root port towards the root bridge.
When the root bridge of the spanning tree topology receives the topology change notification the root bridge creates a configuration BPDU with the topology change notification flag set and floods it all switches.
When a switch receives this configuration BPDU with the topology change flag set, all switches change their MAC address aging timer (default 300 seconds) to the forwarding delay timer (default 15 seconds). The result of setting the MAC address timer so short is it flushes out MAC addresses for devices that did not communicate in this short window but retains all MAC addresses for devices that did communicate in forwarding delay timer size ageing window. The MAC address ageing timer is set back to its last value (300 seconds by default) once the forwarding delay timer expires.
By flushing the MAC address table this prevents the switch from sending traffic to a host that is no reachable via the port. There is a side effect from flushing the MAC address table in that there is an increase of unknown unicast flooding whilst the MAC table is built up again with updated information.
Topology change notifications can generated on a per VLAN basis, so when a TCN is generated it will affect all hosts within that VLAN. As the number of hosts increases in a network so will the number of topology change notifications generated. To view when a topology change has occured on a VLAN you can use the command show spanning-tree vlan X detail on a STP participating switch. The command will show the number of topology changes and when the last change had occurred.
When the layer 1 signalling places a port in a down state, through a link going offline as an example, STP will consider this event as a direct link failure and will react in a number of ways.
If a link has one port as a designated port and the other switches status is a blocking port, there is no interruption to network traffic. Both switches will send a topology change notification to the root switch which will result in a MAC table flush.
If in the event there is a link where one port is a designated port and the other is a root port, the failure will result in network traffic disruption since traffic is forwarding across that link.
The root switch will send a topology change notification out of its designated ports, where if the other switch had an alternative link to the root switch would receive it. This means the affected original switch has a new path to the root switch.
The original switch will need to go through the Forward Age timers (15 seconds) twice, once in the listening and the other in the learning state resulting in a 30 second disruption to network traffic.
Leave a Reply