VLANs Spanning Multiple STPDs

Traditionally, the mapping from VLANs to STP instances have been one-to-one or many-to-one.

In both cases, a VLAN is wholly contained in a single instance. In practical deployment there are cases in which a one-to-many mapping is desirable. In a typical large enterprise network, for example, VLANs span multiple sites and/or buildings. Each site represents a redundant looped area. However, between any two sites the topology is usually very simple.

Alternatively, the same VLAN may span multiple large geographical areas (because they belong to the same enterprise) and may traverse a great many nodes.

In this case, it is desirable to have multiple STP domains operating in a single VLAN, one for each looped area.

The justifications include the following:
  • The complexity of the STP algorithm increases, and performance drops, with the size and complexity of the network. The 802.1D standard specifies a maximum network diameter of seven hops. By segregating a big VLAN into multiple STPDs, you reduce complexity and enhance performance.

  • Local to each site, there may be other smaller VLANs that share the same redundant looped area with the large VLAN. Some STPDs must be created to protect those VLANs. The ability to partition VLANs allows the large VLAN to be "piggybacked" in those STPDs in a site-specific fashion.

The following figure has five domains. VLANs green, blue, brown, and yellow are local to each domain. VLAN red spans all of the four domains. Using a VLAN that spans multiple STPDS, you do not have to create a separate domain for VLAN red. Instead, VLAN red is “piggybacked” onto those domains local to other VLANs.

Click to expand in new window
VLANs Spanning Multiple STPDs
../Graphics/EX_051.svg

In addition, the configuration in the figure has these features: