Spanning Tree UplinkFast: Achieve Quick Convergence Today!

Introduction to Spanning Tree Uplinkfast

Spanning tree uplinkfast is a feature in Cisco switches that keep leaf-node or access switches their ports in the root port while other ports are in blocking states. Data is sent to the root switch through the root port. When the root port becomes unavailable, then the blocked port will act as a root port immediately.

When we enable uplinkfast command on a switch, it will be for the entire switch’s interfaces and vlans. It works by keeping track of possible paths to the root switch. That’s why we don’t apply this command on the root switch. Uplinkfast is added to the conventional STP (802.1D) to reduce convergence time, in case of link failure going to the root switch. We apply the link fast on the access switch or leaf-node switch if it has at least one port in the blocking state.

The main purpose of the uplinkfast is to reduce convergence time. If the root port fails, the alternate port becomes the root port after a 30-second delay.. This delay is due to the listening and learning states.

If there are multiple paths are exist that lead to the root switch. These paths are called uplink groups. However, the link with the lowest cost from the uplink group will be elected as the root port.

Lab Topology for Uplinkfast

Take three switches and connect a PC to each switch, as shown in the lab topology screenshot. Also, configure the switches in the following way;

  • Configure VTP domain marwat and enable VTP mode server on all the switches
  • Create vlan10 on switch1
  • Configure a trunk between all three switches
  • Enable portfast on each interface which connects the PC to the switch
  • Configure access port to the PC’s interface and assign them vlan 10
  • Verify trunk ports and spanning-tree protocol on all the switches
spanning tree uplinkfast lab

Check the spanning tree protocol on switch1;

show spanning tree of switch1

Switch1 is the root bridge, and is running STP 802.1D. Let’s check STP on Switch2;

 show spanning tree of switch2

At the last check spanning-tree protocol on switch3;

 show spanning tree switch3

Switch3’s Ethernet 0/1 serves as the root port and connects directly to the root switch (Switch1), while Ethernet 0/2 remains in a blocking state.

If any root port of switch becomes down, and if it has another port towards the root switch, then it will block the other port. The block port will not go up immediately, but it will wait for 30 seconds before forwarding the state. This wait of 30 seconds is in the case if the switch is running conventional STP (802.1D). Let’s check by example.

Ping from VPC4 (192.168.1.1) to VPC6 (192.168.1.3), and in the same time we will shutdown root port (interface 0/1) of switch3;

ping pc-1

It was successful in pinging, but it fails when we shutdown interface ethernet 0/1 of switch3. After a few seconds (30 seconds), it started pinging to VPC6. This is delay of 30 seconds is because of the listening and learning state of conventional STP (802.1D).

But, if we enable uplinkfast feature on switch3, and if we shutdown any interface of switch3. Then, the block interface will up immediately, and will not wait for listening and learning state which is normally 30 seconds.

Enable or no shutdown interface ethernet 0/1 of switch3.

Apply the spanning tree uplinkfast feature on switch3.

enable spanning tree uplinkfast on switch3

This time again, we will ping from VPC4 to VPC6, and will examine any delay in the packets. And after it, in the same while, we will shutdown interface ethernet 0/1 of switch3 and will check the result.

 ping from vpc4 to vpc6

This time, there is no delay in the packets, due to the spanning-tree uplinkfast feature. It skips the listening and learning states, and enables an interface immediately to the forwarding state without any delay.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *