STP of PVST+ merges with STP of IEEE 802.1D or 802.1Q

Unanswered Question
Feb 2nd, 2009
User Badges:

Reference to the link in the url below:-


http://www.cisco.com/en/US/tech/tk389/tk621/technologies_tech_note09186a00801d11a0.shtml



Can you describe what is the meaning of the statement below:-


If the BPDU of the root in VLAN 1 is better than the BPDU of the root in VLAN 2 then there is no blocking port in the VLAN 2 topology. The BPDU of VLAN 2 never makes a “full circle” around the topology; it is replaced by the VLAN 1 BPDU on the B-C link, because B runs only one STP merged with VLAN 1 STP of PVST+. Thus, there is a forwarding loop. Fortunately, switch A sends PVST+ BPDUs of VLAN 2 (to the SSTP address that is flooded by switch B) towards switch C. Switch C will put port C-B into a type-inconsistent state, which prevents the loop.



1) Why the BPDU of the root VLAN 1 is better?


2)Why The BPDU of VLAN 2 never makes a “full circle” around the topology; it is replaced by the VLAN 1 BPDU on the B-C link, because B runs only one STP merged with VLAN 1 STP of PVST+?


3)In what condition switch A sends PVST+ BPDUs of VLAN 2 (to the SSTP address that is flooded by switch B) towards switch C. Switch C will put port C-B into a type-inconsistent state, which prevents the loop.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (2 ratings)
Loading.
Giuseppe Larosa Tue, 02/03/2009 - 00:32
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

Hello Mohd,

the example use a specially built scenario.


Switch B is only capable of 802.1D STP and has a single instance STP running over Vlan1.


Switch A and Switch C are Cisco switches that implement PVST+.

But there is an important difference:

the link between switch A and switch B is a 802.1Q trunk, the link between switch C and switch B is an access port in Vlan 2.

Switch C will send out of port to Switch B 802.1D standard BPDUs not cisco proprietary STP.

This implies that Switch B will understand these messages.

On the other side Switch A detects the presence of a legacy switch over an 802.1Q trunk. It will use standard STP 802.1D frames over native vlan 1 and will use PVST+ proprietary frames :


are sent to the PVST+ MAC address (also called the Shared Spanning Tree Protocol [SSTP] MAC address, 0100.0ccc.cccd), tagged with a corresponding IEEE 802.1Q VLAN tag.


Switch B treats Cisco PVST+ frames as user multicast traffic and will propagate out port to Switch C the ones that are related to vlan 2.


Switch C will receive PVST+ frames for vlan-id 2 over an access port.


Switch B will merge vlan 1 and vlan 2 topologies because it is receiving on both links standard 802.1D BPDUs that it can understand and compare.

So if Vlan1 root bridge (Switch A) has a lower BID then Switch C in vlan2 SwitchB will send out port to C switch the best BPDU (the one received from SWA) and as an end result the BPDU sent by SwitchC stops there.


But switchC by receiving PVST+ frames over the link (they are multicast traffic for switch B) can put the link in inconsistent state cutting the loop.



1) they are making an hypothesis that SWA BID in vlan 1 (priority+ mac address) is lower then SWC BID in Vlan2.


you can build this using the PVST+ commands like

SWA:

spanning-tree priority vlan 1 8192


SWC:

spanning-tree priotity vlan 2 16394


this puts in the conditions


2) explained above:

actually without inconsistency checking Switch C would move the port to Switch B as its root port and would stop to claim to be the root bridge for vlan 2 and so it will go to only listen to better BPDUs on link B to C.

For the incosistency check the link is blocked for vlan2.


3) automatically on an 802.1Q trunk non native vlan BPDUs (for each allowed VLans on trunk and in STP forwarding state) SSTP using Cisco proprietary frames are used even betweeen two Cisco switches


Hope to help

Giuseppe



noxkrugger Tue, 02/03/2009 - 02:07
User Badges:

Thank you very much for the explanation.. really helpful even i'm not fully understand the theory.


btw, what is the best setup to replace the built scenario?

Giuseppe Larosa Tue, 02/03/2009 - 07:40
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

Hello,

the simplest way to have a working scenario is to have all trunk links to the 802.1D legacy switch


in this way the problems described in the document don't happen.


Hope to help

Giuseppe



Actions

This Discussion