Spanning-Tree Protocol
Why is it important to methodically place the primary root bridge ?
- Speed up STP convergence
- Minimize path cost
- Predictable layer 2 pathing
- Reduce network failure
By choosing the design of the tree, you can understand the convergence easier, because it’s predictable if there is a failure and a new root bridge is elected. You also take away the chance of an access layer switch from determining the entire layer 2 path of the network.
It is Cisco best practice to create Root bridges at the core level. This can be done by lowering the priority or making the switch the root primary. The core bridges, which can be the aggregate layer – distribution – should have a slightly higher value, so they can take over and win the root bridge election if the core has a failure. This makes the STP network behave in a more deterministic manner.
Many networks take spanning tree for granted and accept the default configuration for STP. This leaves all switches in the environment using the default root bridge priority of 32768. If all switches have the same root bridge priority, the switch with the lowest MAC address will be elected as the root bridge. This can cause trouble when a new switch is added into the topology. Many networks have not been configured with a single switch to have a lower root bridge priority, which would force that core switch to be elected as the STP root for any or all VLANs.
This is an issue because it is possible that a small access-layer switch with a low MAC address could be the STP root. This situation would add some performance overhead and make for longer convergence times because of the root bridge re-election. This can cause intermittent issues through the network due to convergence while re-election is occurring within the tree.
CISCO: “The network operation, manageability, and application performance could become very complex with traditional Layer 2 technologies such as spanning-tree. The framework of the distribution-layer system must be designed with consideration of Cisco recommended best practices, which significantly reduce network complexities, increase reliability, and accelerate network performance. To build a strong campus network foundation with the three-tier model, the distribution layer has a vital role in consolidating networks and enforcing network edge policies“
NOTE: If you are using standard default STP configuration close to 50% of your uplink ports will be blocking and passing no traffic. By careful design and placement, you can configure (rig-STP-elections) and design a system where a trunk is blocking for one-vlan, but forwarding for another. There is a fine balance, like with any good network architecture and design, by making the fewest modifications from STP defaults as possible, so that you don’t create headaches.
**You can also eliminate STP and push Layer 3 down as close as you can to the access layer. Using etherchannel’s will increase network bandwidth and resilience. Create static access ports with portfast and BPDU Guard on them.
Leave a Reply