Spanning Tree Speculation

daveba123

According to my studies I've come to the following conclusions:

When root cost is used to determine the root port or the designated port, it seems that only the outbound interface towards the root switch is factored. For example, if you look at the attached diagram, it appears that only SW2's Gi0/2 port and SW3's Gi0/1 are used, but SW2's Gi0/1, and SW3's Gi0/2 are ignored.

My speculation is, am I misinterpreting the ignoring of ports in that actually it's the true segment link speed that is cost. In other words, if one one side was Gigabit and the other was Fast Ethernet, would the 100Mbps speed be calculated as cost instead of the Gigabit cost?

Thanks for the help! I'm very puzzled....

David

MrXpert

I hope I'm understanding you correctly. had to read your post a couple of times

Assuming SW1 is the root bridge, then it sends out a BPDU with cost=0 out its designated ports.

SW2 receives it on its gi0/2 and adds path cost=4 to it.RPC=4
SW3 receives it on its gi0/1 and adds path cost=4 to it. RPC=4
SW3 forwards the BPDU onto SW2. SW2 receives it on gi0/1 and adds 4 to the cost. RPC=8 on gi0/1 on SW2.
SW2 forwards on BPDU to SW3. SW3 receives it on its gi0/2 and adds 4 to the cost. RPC=8

SW2s best lowest root path cost is out gi0/2. SW3 will use gi0/1 as its root port.
Between SW2 and SW3 they finally have to calculate which port will go into FWD mode and which one will be BLK. Seeing as they both have the same RPC, tie breaker is whoever has the lowest MAC address will be put into FWD state.

daveba123

Thanks, MrXpert. I understand now.

The cost is not calculated from a particular interface, but is calculated from the link speed between the 2 switches. For example, if one side of the link is Gigabit and the other is Fast Ethernet, the speed of the link is 100Mbps because 100Mbps is the fastest speed the link can support. Therefore the cost is 19 regardless of which side the Fa or Gi port is on.