An organization often implements redundancy in the network to ensure continued availability in the event of a device or link failure. Of course, redundancy means that two or more possible paths exist in the network, which may be via a different set of devices. If you have a redundant network that has high utilization, ideally you want to share traffic loads over the multiple paths and devices available in the network.
Spanning tree is a protocol that was designed to deal with the loops introduced as a result of implementing redundancy in the network; however, it was never designed with load sharing in mind. The only way you can load share with spanning tree is to use multiple spanning-tree topologies. By default, Cisco Catalyst switches use a separate spanning tree topology for each VLAN. This separate topology means that you can manipulate each spanning-tree topology to ensure that all redundant links in the network are utilized.
Figure 4-11 shows a similar topology to the previous scenario (see Figure 4-5), except with a few modifications that enable you to configure STP load sharing.
Figure 4-11. Scenario 4-2 Network Topology
The following changes have been made to the topology:
Because now four different VLANs are in the network now, it is important that the network is configured to ensure it operates as efficiently as possible. To achieve the efficiency you need to load share VLAN traffic over different paths where multiple paths lead towards the traffic destination.
Several methods of implementing spanning-tree load sharing exist, each with advantages and disadvantages:
Each of these methods is now discussed.