MSTP Operation

To further illustrate how MSTP operates and converges, MSTP Topology with the CIST Root Bridge Contained within a Region displays a network with two MSTP regions. Each region contains three MSTP bridges and one MSTI. The overall network topology also contains one CIST root bridge (Switch A, which has the lowest bridge ID), one interconnecting 802.1w bridge (Switch D), and 10 full duplex, point-to-point segments. VLAN Default spans all of the bridges and segments in the network, VLAN engineering is local to its respective region, and STPD S0 is configured as the CIST on all bridges.

MSTP Region 1 consists of the following:

Three bridges named Switch A, Switch B, and Switch C
One MSTI STPD named S1 with an MSTI ID of 1
VLAN Engineering mapped to the MSTI STPD, S1
Switch A as the CIST root bridge (this is the CIST root bridge for all regions)
Switch A as the CIST regional root bridge
Switch A as the MSTI regional root bridge
Three boundary ports that connect to MSTP Region 2 and other 802.1D or 802.1w bridges

MSTP Region 2 consists of the following:

Three bridges named Switch E, Switch F, and Switch G
One MSTI STPD named S1 with an MSTI ID of 1

Note

The MSTI ID does not have any significance outside of its region so you can reuse IDs across regions.

VLAN finance mapped to the MSTI STPD, S1
Switch E as the CIST regional root bridge
Switch F as the MSTI regional root bridge
One master port that connects to the CIST
Three boundary ports that connect to MSTP Region 1 and other 802.1D or 802.1w bridges

The following sequence describes how the MSTP topology convergences:

Determining the CIST root bridge, MSTP regions, and region boundaries.
Each bridge believes that it is the root bridge, so each bridge initially sends root bridge BPDUs throughout the network. As bridges receive BPDUs and compare vectors, the bridge with the lowest Bridge ID is elected the CIST root bridge. In our example, Switch A has the lowest Bridge ID and is the CIST root bridge.

The bridges in the MSTP regions (Switches A, B, C, E, F, and G) advertise their region information along with their bridge vectors.

Segments 1, 3, and 9 receive BPDUs from other regions and are identified as boundary ports for Region 1. Similarly, segments 2, 3, and 9 are identified as boundary ports for Region 2.
Controlling boundary ports.
The CIST regional root is advertised as the Bridge ID in the BPDUs exiting the region. By sending CIST BPDUs across regional boundaries, the CIST views the MSTP regions as virtual 802.1w bridges. The CIST takes control of the boundary ports and only CIST BPDUs enter or exit a region boundary.

Each MSTP region has a CIST regional root bridge that communicates to the CIST root bridge. The bridge with the lowest path cost becomes the CIST regional root bridge. The port on the CIST regional root bridge that connects to the CIST root bridge is the CIST root port.

For Region 1, Switch A has the lowest cost (0 in this example) and becomes the CIST regional root. Since the bridge is both the CIST root bridge and the CIST regional root bridge, there is no CIST root port on the bridge.

For Region 2, Switch E is the CIST regional root bridge and so a port on that bridge becomes the CIST root port.
Identifying MSTI regional roots.
Each MSTI in a region has an MSTI regional root bridge. MSTI regional roots are selected independently of the CIST root and CIST regional root. The MSTP BPDUs have M-records for each MSTI. Bridges belonging to an MSTI compare vectors in their M-records to elect the MSTI regional root.
Converging the CIST.
The CIST views every region as a virtual bridge and calculates the topology using the 802.1w algorithm. The CIST calculates the topology both inside and outside of a region.
Converging MSTIs.
After the CIST identifies the boundary ports, each MSTI in a domain converge their own trees using 802.1w.

At this point, all CIST and MSTIs have assigned port roles (Root, Designated, Alternate, and Backup) to their respective spanning trees. All root and designated ports transition to the forwarding state while the remaining ports remain in the discarding state.

Propagating topology change information is similar to that described for RSTP.

For more information, see Propagating Topology Change Information.

For a configuration example, see MSTP Configuration Example.