MSTP Configuration Guide

Introduction

MSTP (Multiple Spanning Tree Protocol) is a kind of spanning tree protocol developed on the basis of STP/RSTP. Proposed in IEEE802.1s, it inherits all the advantages of STP/RSTP while solving the problem of idle links in STP/RSTP by binding most STP instances to a few VLANs. This means that by forming multiple trees without loops, broadcast storms are resolved and redundant backups are achieved. Multiple spanning trees achieve load balancing between VLANs, with traffic from different VLANs forwarded in different paths.

Basic Concepts

MSTP Principle

The principle and calculation method of MSTP is the same as STP/RSTP. The basic idea is to prune the ring network into an acyclic tree by comparing priorities and connecting each sub-priority device in turn from the highest priority root node (Root) in the network.

MSTP Domain

Unlike STP/RSTP the concept of domain is introduced in MSTP, dividing the whole network into different domains, consisting of multiple switching devices in the switching network and the segments between them. The switches in the same MST domain have the following characteristics.

• Both activated MSTP.
• Has the same domain name.
• Has the same VLAN-to-spanning-tree instance mapping configuration.
• Has the same MSTP revision level configuration.

Multiple MST domains can exist in a LAN, and each MST domain is physically connected to each other directly or indirectly. Users can divide multiple switching devices into the same MST domain by using the MSTP configuration command.

• The configuration resulted in the three domains region 1 to 3 as shown.
• In MSTP instance 0 has a special meaning and is called CISTI (Common and Internal Spanning Tree Instance) and the rest of the spanning trees are called MSTI (Multiple Spanning Tree Instance). By default all VLANs are bound to the CISTI, ensuring the security of the entire network.
• CIST is a single spanning tree that connects all switching devices within a switching network, generated by STP or RSTP protocol calculations. The IST (Internal Spanning Tree) plus the CST (Common Spanning Tree) of all MST domains form a complete spanning tree, or CIST, which is expressed as an IST within a domain, where the switches within the domain are computationally pruned to obtain a spanning tree, and as a CST between domains, where each domain is treated as a single node and a spanning tree is obtained by pruning each domain through calculation.

From the figure it can be seen that CIST prunes the ring network into a tree within the domain separately and the entire network is pruned into a loop-free tree in the unit of domain between domains. Due to the introduction of domains and instances, there are two types of root nodes in MSTP: the total root, and the domain root. The total root is a global concept, relative to all interconnected STP/RSTP/MSTP devices with only one total root, and is also the root of the CIST. The domain root is a local concept relative to a specific instance of a specific domain. S1 is the total root and also the domain root of instance0 in region 1.

MSTP Message

MSTP uses the Multiple Spanning Tree Bridge Protocol Data Unit (MST BPDU) as the basis for spanning tree calculations, and MST BPDU messages are used to calculate the spanning tree topology, maintain the network topology and communicate topology change records. The message format is as follows

For both intra-domain MST BPDU and inter-domain one, the first 35 bytes are the same as for RST BPDUs. From the 36th byte onwards is the MSTP proprietary field. The final MSTI configuration information field is concatenated by a number of MSTI configuration information groups. The key messages in the MST BPDU are shown in the table below.

Table 1 MST BPDU key messages

field	Length	Significance
protocol ID	2	0x00 indicates the STP family of protocols, including STP, RSTP, MSTP
Protocol Version ID	1	0x 00 for STP, 0x02 for RSTP, 0x03 for mstp
BPDU type	1	0x00 for configure, 0x02 for RSTP or MSTP, 0x80 for TCN
flag	1	Identification by bit, topology change and P/A bit, role are in it
Root ID	8	CIST Root ID
Extenal Path Cost	4	CIST external path cost
Regional Root ID	8	The root ID in this region
Port ID	2	Port ID of the sending port
Message Age	2	Message ageing value
Max Age	2	Maximum message aging value
Hello Time	2	Time interval for sending BPDUs
Forward delay	2	Status switching interval
Version1 Length = 0	1	This is the end of RSTP, the next code is MSTP
Version3 Length	1	Length of Version3
MST Configure ID	51	Identifies the MST domain in which it is located
CIST Internal Root Path Cost	4	CIST internal path cost
CIST Bridge ID	8	CIST Bridge ID
CIST Remaining Hops	1	CIST remaining Hops
MSTI Configure Message	-	The configuration of the MSTI, describing the individual MSTI instances in turn. This field is not available when there is no MSTI

The code for the MST Configure ID is as follows.

Table 2 MST Configure ID

field	Length	Significance
Configure ID Format Selector	1	Use 0x00 directly
configure name	32	Name
revision level	2	No specific requirements, for simplicity, use 0 in sonic
configure Digest	16	The summary information of all MSTI and VLAN mappings on this Bridge

The code for the MSTI Configure Message is as follows.

Table 3 Message Code

field	Length	Significance
MSTI Flags	1	Marking by position
MSTI Regional Root Identifier	8	Domain Root
MSTI Internal Root Path Cost	4	Intra-domain path cost
MSTI Bridge Priority	1	Intra-domain Bridge priority
MSTI Port Priority	1	Intra-domain port priority
MSTI Remaining Hops	1	MSTI’s remaining hops

MSTP Topology Calculation Principles

MSTP divides the entire network into multiple MST domains, treating each domain as a node. Each MST region is computed based on the STP or RSTP algorithm and generates CST (a single spanning tree). The MSTP vector priority is calculated as follows [Table 42] shown here

Table 4 MSTP Vector Priorities

Priority vector name	Description
Rootbridge ID	The Root Bridge ID is used to select the root bridge in the CIST. The corresponding bridge ID in the BPDU is calculated as: Priority (l6bits) + MAC (48bits)
External route path cost (ERPC)	The path cost from the MST domain root to the total root. the external route path cost saved on all switches in the MST domain is the same. If the CIST root bridge is in the domain, the external route path cost saved on all switches in the domain is 0
Domain Root ID	Also commonly referred to as the MSTI tree root, the domain root ID is used to select the root of the tree in MSTI. It is also elected by the bridge ID and is calculated as Priority (l6bits) + MAC (48bits)
Internal route path cost (IRPC)	The route path cost for this switch to reach the domain root bridge. The internal route path cost value kept by the domain edge port is greater than (the lower the priority) the internal route path cost kept by the non-domain edge port
Designated Bridge	The designated bridge for a CIST or MSTI instance is the nearest upstream switch to the domain root for this switch. If this switch is the master root or domain root, the designated bridge is its own
Designated port	The port on the designated bridge that is connected to the root port of this switch is the Designated Port. Its Port ID (Port ID) = Priority (8 bits) + Port Number (8 bits).
Receiving ports	The port priority must be an integer multiple of 16 for the port that received the BPDU message. Its Port ID (Port ID) = Priority (8 bits) + Port number (8 bits). Port Priority must be an integer multiple of 16

The minimum vector has the highest priority and the rules for comparison are as follows.

First, compare the root bridge IDs. If the root bridge IDs are the same, then compare the ERPC. If the ERPC is still the same, then compare the domain root ID. If the domain root ID is still the same, then compare the IRPC. If the IRPC is still the same, compare the specified bridge ID. If the specified bridge ID is still the same, compare the specified port ID. If the specified port ID is still the same, compare the receiving port ID. Calculation of CIST After configuration message exchange comparison, a switch with the highest priority in the entire network is first selected as the root of the CIST tree, and then the IST is computed within each MST domain by the MSTP protocol algorithm; at the same time MSTP treats each MST domain as a single switch and computes the CST between MST domains by the STP or RSTP protocol algorithm. The CST and IST form the CIST for the entire switch network. Calculation of MSTI Within the MST domain, MSTP generates different spanning tree instances for different VLANs based on the mapping relationship between VLANs and spanning tree instances, with the following characteristics. (1) Each MSTI computes its own spanning tree independently, without interfering with each other. (2) The spanning tree calculation method for each MSTI is essentially the same as for the RSTP. (3) Each MSTI spanning tree can have a different root and a different topology. (4) Each MSTI sends BPDU within its own spanning tree. (5) The topology of each MSTI is determined by command configuration (not automatically generated). (6) The spanning tree parameters can be different for each port on different MSTIs. (7) The role and status of each port can be different on different MSTIs. MSTI Spanning Tree Algorithm Implementation At the beginning, each port of each switch generates a configuration message with its own switch as the root bridge, where the root path cost is 0, the specified bridge ID is its own switch ID and the specified port is this port. Each switch sends its own configuration message out and performs the following operations upon receipt of other configuration messages. (1) When a port receives a configuration message with a lower priority than its own (the comparison of priorities is based on the vector priority comparison rules described earlier), the switch discards the received configuration message and does not do anything with the configuration message for that port. (2) When a port receives a configuration message with a higher priority than the configuration message of this port, the switch replaces the contents of the configuration message of that port with the contents of the received configuration message; then the switch compares the configuration message of that port with the configuration messages of other ports on the switch and selects the best configuration message. The steps to calculate a spanning tree are as follows. (1) Elect the root bridge. This step is performed by comparing the tree root IDs of configuration messages sent by all switches, and the switch with the lowest tree root ID value is the CIST root bridge, or the MST domain root bridge. (2) Elect the root port on a non-root bridge. Each non-root bridge designates the port that receives the optimal configuration message as the root port of its own switch. (3) Election of the designated port. In this step there are two further sub-steps as follows. First, the switch calculates a standard designated port configuration message for each port based on the root port’s configuration message and the root port’s path cost. Replace the tree root ID with the tree root ID in the root port configuration message, replace the root path overhead with the root path overhead in the root port configuration message plus the path overhead of the root port, replace the specified bridge ID with the ID of its own switch, and replace the specified port ID with its own port ID. The switch then compares the configuration message calculated by the above rules with the original configuration message on the corresponding port. If the original configuration message on the port is better, the switch blocks the port and the configuration message on the port remains unchanged, and the port will no longer forward data and will only receive configuration messages (equivalent to the root port); if the configuration message calculated by the above replacement is better than the original configuration message on the port, the switch sets the port as the designated port and the configuration message on the port is replaced with the configuration message calculated by the above replacement and is sent out periodically. (4) After the MSTI spanning tree topology converges, BPDUs are sent periodically according to the Hello timer regardless of whether the non-root bridge receives information from the root bridge. If a port does not receive BPDUs from the specified bridge (the higher-level switch it is connected to) for 3 consecutive Hello times (the default setting), then the switch considers the link with this neighbor has failed.

MSTP’s Handling of Topology Changes

The criterion for detecting a topology change in MSTP is based on whether the state of non-edge port has migrated to the Forwarding state, and if it has, a topology change occurs. Once the switch detects a topology change, it proceeds as follows.

• Starts a TC While Timer (which is twice the value of Hello Time) for all non-edge designated ports on this switch and clears the incoming MAC addresses for these ports during this time. If it is a state change on the root port, the root port is started.
• These ports where a state change has occurred send out TC BPDUs in which the TC is set until the TC While Timer times out. The root port always has to send such TC BPDUs. The other switches receive the TC BPDUs and perform the following processing.
• Clear all MAC addresses learned by the port, except for ports that receive TC BPDUs.
• Start the TC While timer for all own non-edge assigned ports and root port and repeat the process.

MSTP Configuration

The priority of MSTP configuration is: BPDU protection > BPDU filtering > MSTP edge interface (and normal MSTP calculation process).

• If BPDU protection and BPDU filtering are enabled at the same time, BPDU protection takes effect and the interface goes down after receiving the message.
• With BPDU protection or BPDU filtering enabled, the MSTP edge interface will never lose its edge interface role.

MSTP Default Setting

The default setting of MSTP is shown in the table below.

Table 5 MSTP Default Setting

Parameters	Default value
Spanning Tree Protocol working mode	MSTP mode
MSTP functionality	The global MSTP function is enabled, and the MSTP function of the port is enabled.
Priority of instance	8
Priority of the port	128
Calculation of path overhead	dot1t, the IEEE 802.1t standard
Forward Delay Time	15 seconds
Hello Time	2 seconds
Max Age Time	20 seconds
Auto edge port	Enabled
BPDU filter	Disabled
BPDU guard	Disabled

MSTP Global Switch

Table 6 MSTP Global Switch

Purpose	Commands	Description
Enter global configuration view	configure terminal	-
Enable global MSTP	mstp state enable	-
Disable global MSTP	mstp state disable	-

Configure MSTP Attributes

Table 7 Configure MSTP Attributes

Purpose	Commands	Description
Enter global configuration view	configure terminal	-
Configure MSTP forward delay	mstp forward_delay time	time: delay time, range 4-30
Configure the MSTP hello packet delivery interval	mstp hello time	time: interval time, range 1-10
Configure MSTP aging time	mstp max_age time	time: ageing time, range 10-1000000
Configure the MSTP domain	mstp name name	-
Specify the bridge MAC.	bridge mac HH:HH:HH:HH:HH:HH	-

Configure MSTP Instance

Table 8 Configure MSTP Instance

Purpose	Commands	Description
Enter global configuration view	configure terminal	-
Create an instance and enter the instance configuration view	mstp instance instance-id	name: Name of the instance
Bind VLAN	vlan vlan-id	vlan-id: vlan id
Set instance priority	priority priority	Instance priority, with a value range of [0,15] and a default value of 8. The smaller the value, the higher the priority.
Set port priority in the instance.	interface priority priority interface [ethernet|link-aggregation] interface-name	Interface priority, with a value range of [0,15] and a default value of 8. The smaller the value, the higher the priority.

Configure MSTP Edge Port

The auto-edge interface feature is enabled by default on all Layer 2 ports. An interface with auto-edge enabled does the following things when its state changes from down to up: waiting for 3 seconds and trying to receive BPDUs; If no BPDU is received, it becomes an edge port and will lose edge port role the first time it receives a BPDU thereafter.

Table 9 Configure MSTP Edge Port

Purpose	Commands	Description
Enter global configuration view	configure terminal	-
Enable MSTP auto edge for the interface.	mstp auto-edge enable interface {ethernet|link-aggregation} interface-name	-
Disable auto edge port.	mstp auto-edge disable interface {ethernet|link-aggregation} interface-name	If you do not assign it as a manual edge port, it will lose edge port role.
Enable MSTP manual edge for the interface.	mstp edge-port enable interface {ethernet|link-aggregation} interface-name	Manual edge port is disabled by default on all Layer 2 ports.When a manual edge port is enabled on an interface, it immediately becomes an edge port until it receives a BPDU message.The priority of manual edge port configuration is higher than auto edge port, that is, when manual edge port is enabled, auto edge port configuration is invalid.
Disable MSTP manual edge port.	mstp edge-port disable interface {ethernet|link-aggregation} interface-name	-

Enable BPDU Filtering

This command will enable MSTP at interface level and determine whether the interface participates in MSTP protocol operation. The interface to be operated should be a Layer 2 port, and MSTP should be enabled before configuration. By default, MSTP is enabled on all Layer 2 interfaces. After this function is enabled, the interface will not participate in MSTP protocol operation, and there will be a risk of loops, therefore, please take caution.

Table 10 Enable BPDU Filtering

Purpose	Commands	Description
Enter global configuration view	configure terminal	-
Enable BPDU filtering for the interface.	mstp bpdu-filter enable interface {ethernet|link-aggregation} interface-name	-
Disable BPDU filtering.	mstp bpdu-filter disable interface {ethernet|link-aggregation} interface-name	-

Enable BPDU Guard

When the interface is enabled with BPDU message protection, it will be admin down if received a BPDU.

Table 11 Enable BPDU Guard

Purpose	Commands	Description
Enter global configuration view	configure terminal	-
Enable BPDU guard for the interface.	mstp bpdu-guard enable interface {ethernet|link-aggregation} interface-name	-
Disable BPDU guard.	mstp bpdu-guard disable interface {ethernet|link-aggregation} interface-name	-

Display and Maintenance

Table 12 MSTP Display and Maintenance

Purpose	Commands	Description
Show MSTI status	show mstp bridge	-
Show the status information of interface in CIST	show mstp brief [{ethernet|link-aggregation} interface-name]	-
Show detailed status information of ports in CIST	show mstp interface [{ethernet|link-aggregation} interface-name]	-
Show summary information of the MST domain configuration currently in effect on the device	show mstp mstconfid	-
Show MSTP status information	show mstp status	-
Show the MSTI status of a given MSTP instance	show mstp tree instance-id	-
Show detailed information on MSTI port status	show mstp treeport {ethernet|link-aggregation} interface-name instance instance-id	-

Typical Configuration Example

MSTP Function Configuration

1. Networking Requirements In a complex network, due to the need for redundancy and backup, network planners generally tend to deploy multiple physical links between devices, with one serving as the primary link and the others as backup links. This inevitably results in loops, which, if present in the network, may cause broadcast storms and corrupt MAC table entries. For this reason, loops can be prevented by deploying the MSTP protocol, which blocks redundant links in the Layer 2 network and prunes the network into a tree to eliminate loops. MSTP runs on the three devices. To implement load balancing between VLAN40 and VLAN41, MSTP introduces multiple instances. MSTP can set up VLAN mapping tables to associate VLANs with spanning tree instances.
2. Topology

3. Procedure

#Configure VLANs

Switch A

sonic# config terminal
sonic(config)# vlan 40
sonic(config)# vlan 41
sonic(config)# interface ethernet 40/0
sonic(config-if-0/40)# switchport access vlan 40
sonic(config-if-0/40)# switchport trunk vlan 41
sonic(config)# interface ethernet 41/0
sonic(config-if-0/41)# switchport access vlan 40
sonic(config-if-0/41)# switchport trunk vlan 41

Switch B

sonic# config terminal
sonic(config)# vlan 40
sonic(config)# vlan 41
sonic(config)# interface ethernet 40/0
sonic(config-if-0/40)# switchport access vlan 40
sonic(config-if-0/40)# switchport trunk vlan 41
sonic(config)# interface ethernet 41/0
sonic(config-if-0/41)# switchport access vlan 40
sonic(config-if-0/41)# switchport trunk vlan 41
sonic(config)# interface ethernet 42/0
sonic(config-if-0/42)# switchport access vlan 40
sonic(config-if-0/42)# switchport trunk vlan 41

Switch C

sonic# config terminal
sonic(config)# vlan 40
sonic(config)# vlan 41
sonic(config)# interface ethernet 40/0
sonic(config-if-0/40)# switchport access vlan 40
sonic(config-if-0/40)# switchport trunk vlan 41
sonic(config)# interface ethernet 41/0
sonic(config-if-0/41)# switchport access vlan 40
sonic(config-if-0/41)# switchport trunk vlan 41
sonic(config)# interface ethernet 42/0
sonic(config-if-0/42)# switchport access vlan 40
sonic(config-if-0/42)# switchport trunk vlan 41

#MSTP configuration

Switch A

sonic# config terminal
sonic(config)# mstp enable
sonic(config)# mstp name region1
sonic(config)# mstp instance 40
sonic(config-mstp-instance-40)# vlan 40
sonic(config-mstp-instance-40)# priority 1
sonic(config)# mstp instance 41
sonic(config-mstp-instance-41)# vlan 41

Switch C

sonic# config terminal
sonic(config)# mstp enable
sonic(config)# mstp name region1
sonic(config)# mstp instance 40
sonic(config-mstp-instance-40)# vlan 40
sonic(config-mstp-instance-40)# priority 1
sonic(config)# mstp instance 41
sonic(config-mstp-instance-41)# vlan 41
sonic(config-mstp-instance-41)# interface priority 2 interface ethernet 0/42
sonic(config-mstp-instance-41)# interface priority 4 interface ethernet 0/40

Switch B

sonic# config terminal
sonic(config)# mstp enable
sonic(config)# mstp name region1
sonic(config)# mstp instance 40
sonic(config-mstp-instance-40)# vlan 40
sonic(config-mstp-instance-40)# priority 1
sonic(config)# mstp instance 41
sonic(config-mstp-instance-41)# vlan 41

4. Verify the configuration.

sonic# show mstp
role Flags: Root - Root, Desg - Designated, Altn - Alternate, Back - Backup, Mstr - Master, Disa - Disabled
state Flags: disc - Discard/Blocking/Listening,  lear - Learning, forw - Forwarding
Spanning-tree Mode: mstp
vlan   mst instance  port_role_state
-------  --------------  ------------------------------------------------------------------
Vlan40   40      Ethernet40(Desg)(forw) Ethernet41(Desg)(forw)
Vlan41   41      Ethernet40(Root)(forw) Ethernet41(Desg)(forw)

Send the stream to verify Send mutual traffic between ports to see how traffic is sent and received

vlan	No.	Traffic forwarding results
40	1	Ports 41, 40 on Switch A forwarding traffic in Vlan 40
	2	Port 41 on Switch B forwarding traffic in Vlan 40
	3	Ports 40,42 on Switch B not forwarding traffic in vlan 40
	4	Ports 40, 41, 42 on Switch C forwarding traffic in Vlan 40
41	5	Ports 41, 40 on Switch A forwarding traffic in Vlan 41
	6	Ports 40, 41, 42 on Switch B , all forwarding traffic in Vlan 41
	7	Port 42 on Switch C forwards traffic in Vlan 41
	8	Ports 40, 41 on Switch C not forwarding traffic in Vlan 41