0

While working on a network using VTP (without pruning yet), I noticed the STP topology (for VLAN 10 let say as an example) was including all switches even the ones not having any interface in the VLAN 10.

Is this having an impact on the convergence time for VLAN 10 knowing the network is using rapid-pvst? I would say no as RSTP seems insensitive to network size while propagating TC. No risk neither to flood more unicast (due to mac address tables being flushed) as those switches are not having any interface tied to our VLAN10.

If I enable VTP pruning, will the switches not having VLAN 10 interface will still create a RSTP instance for "their" local STP topology? Asking this because they will still have VLAN10 in vlan.dat but no trunk link allowing VLAN10 traffic to exchange BPDU for that VLAN? They will thus consider themselves as Root bridge for that topology instance?

I finally ended up with more than one question but that will help me to better understand how VTP and RSTP actually interconnects.

Official doc from Cisco about VTP (https://www.cisco.com/c/en/us/support/docs/lan-switching/vtp/10558-21.html) highlights some potential risks using VTP in parallel to STP but I'm not able to figure out what are they (if using RSTP, not STP).

2

After days of trials and researches, I managed to get a clear answer that I want to share here also.

VTP pruning does NOT reduce the number of STP instances running on each switch; it just removes replication of broadcast, multicast, unknown unicast frames for the pruned Vlan but STP/RSTP BPDU traffic is not impacted by VTP Pruning.

| improve this answer | |
  • VTP propagates the knownledge about VLANs in the network. It doesn't manage redundant links like STP. – Zac67 May 1 at 15:37
  • Thanks Zac for the comment. What I meant is that VTP actually pushes all VLAN in all switches of the VTP domain; as a consequence, the STP topology may be much bigger than needed. So are the trunk bandwidthes. Although pruning could be an option to reduce bandwidth, it cannot be one for reducing the wide STP topology generated by VTP. – mouch May 1 at 16:59
  • VTP doesn't interfere with RSTP is what is was trying to say (RSTP is VLAN-agnostic). You're correct though in that VLAN propagation does cost (some) bandwidth for superfluous broadcast forwarding - depending on the network size, that isn't normally a problem. – Zac67 May 1 at 19:25
  • On top of that, a wide STP topology is more likely to suffer from topology change. Imagine having a pretty small VLAN (used on 2-3 switches) compared to the physical network (more than 100 say); I would expect to be protected from topology changes occuring on other switches than the 2-3 where my VLAN spans. With VTP enabled; that is not true as the VLAN is now declared everywhere and even the VTP pruning option is not helping on this precise topic. Note: I'm talking about Cisco's PVRSTP here. – mouch May 3 at 15:38
  • 1
    RPVST+ is a bit different in that it creates a separate spanning tree for each VLAN. Each tree needs to converge on its own - more VLAN propagation = larger trees = longer reconvergence. In your question you were referring to RSTP which is VLAN-agnostic, so it's the direct opposite: always just a single tree, no matter how many or how large the VLANs. – Zac67 May 4 at 12:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.