06-07-2010 04:39 PM - edited 03-06-2019 11:27 AM
Everyone
I was hoping we could start some discussion with the differences of the following protocols for spanning-tree.
I really started to get confused after I am just now learning STP. I do understand that these allow for faster convergence.
However - why do they all have to exist? What was wrong with the 802.1D
If this is something to advanced to get into - then perhaps a link to good documentation will surfice.
Thanks in advance to everyone who can participate here.
Joe
Solved! Go to Solution.
06-07-2010 06:02 PM
Joseph,
Just in general...
PVST+ is per-VLAN spanning tree (which is the default for most cisco switches).
It means that you will run an spanning-tree instance per VLAN.
This is useful when you need different layer 2 behaviors per VLAN, for example you can have different root bridge on different VLANs (so that spanning tree does not have to run as a whole on the layer 2 domain, but can run a different instance per-VLAN)
RSTP is rapid STP.
It is an enhancement to STP.
RSTP does not work with timers as regular STP (which takes up to 30-50 seconds to converge due to the transtion to all its states)
Regular STP can use port-fast for ports not connected to other switches, but all ports connected to other switches need to transition from blocking to listening, learning and finally forwarding.
RSTP optimizes this by using P2P lilnks and taking up to only 2 seconds to converge.
RPVST+
Is basically when you mix the two above.
You have an instance of rapid STP running per VLAN.
Also, some use MST which is another variance of STP which can group several VLANs to be part of a single MST region (and behave like RSTP inside that region).
MST is useful because if you have 1000 VLANs, normally you don't need to have 1000 STP/RSTP instances!!
You can instead have one instance with VLAN 1-500 and another instance with VLANs 501-1000 (just to give you an example)
You can get started here:
http://en.wikipedia.org/wiki/Spanning_tree
Federico.
06-08-2010 12:16 AM
Joe
Just to add to Federico's excellent reply.
They all exist because they are all still needed. Rapid STP / Rapid PVST+ is not supported on some of the older switches that may still be running in Enterprise networks.
And PVST, PVST+, Rapid PVST+ are Cisco proprietary because running an instance of STP per vlan is a Cisco specific thing.
Jon
06-07-2010 06:02 PM
Joseph,
Just in general...
PVST+ is per-VLAN spanning tree (which is the default for most cisco switches).
It means that you will run an spanning-tree instance per VLAN.
This is useful when you need different layer 2 behaviors per VLAN, for example you can have different root bridge on different VLANs (so that spanning tree does not have to run as a whole on the layer 2 domain, but can run a different instance per-VLAN)
RSTP is rapid STP.
It is an enhancement to STP.
RSTP does not work with timers as regular STP (which takes up to 30-50 seconds to converge due to the transtion to all its states)
Regular STP can use port-fast for ports not connected to other switches, but all ports connected to other switches need to transition from blocking to listening, learning and finally forwarding.
RSTP optimizes this by using P2P lilnks and taking up to only 2 seconds to converge.
RPVST+
Is basically when you mix the two above.
You have an instance of rapid STP running per VLAN.
Also, some use MST which is another variance of STP which can group several VLANs to be part of a single MST region (and behave like RSTP inside that region).
MST is useful because if you have 1000 VLANs, normally you don't need to have 1000 STP/RSTP instances!!
You can instead have one instance with VLAN 1-500 and another instance with VLANs 501-1000 (just to give you an example)
You can get started here:
http://en.wikipedia.org/wiki/Spanning_tree
Federico.
06-07-2010 06:04 PM
Federico
Orale! Thank you very much again for your insight and knowledge!
I will continue to venture into the world of Spanning Tree.
Gracias!!!!
Joe
01-21-2020 03:34 AM
Does RSTP and Rapid PVST performed well ? Means.. Suppose In my network I have three cisco switches, two are running on RSTP and middle one is running on Rapid PVST.
In my case... with above scenario spanning tree is not working well, one of the RSTP switch and middle Rapid PVST switch both are becoming root for all vlans and BLK/Descarding is not happening. Can someone support on it. As Cisco SF300 does not have PVST mode and in production cannot convert PVST mode switches into RSTP.
01-21-2020 09:06 AM
You generally need to run the same version of spanning tree on all of your network devices. The notable exception being that if you have RSTP and STP, RSTP will see an STP region as a single "device" and give it exactly one functional connection to the RSTP part of the network, to allow you to leave legacy STP devices in service while upgrading to RSTP.
Why are your 2nd and 3rd devices not using RPVST+?
01-21-2020 11:29 PM
Thanks for your reply and suggestions.
Actually we have SF300 switches also in network along with catalyst switches and SF300 does not have PVST or rapid PVST option. Hence we are suffering a lot to tuning the Spanning-tree in network. Is there any no way except remove SF300 and put Catalyst switches ?
01-22-2020 06:05 AM
You originally said that the other 2 were running RSTP. If they're running legacy STP, you need to remember that they'll form their own STP "group," with only one active link to the RSTP "group." My recommendation would be to connect these to the RSTP core and not cross-connect them to each other, in which case you shouldn't notice any difference. If you're concerned about redundancy, that link to the core can be redundant, either with STP keeping one active, or as a port-channel so that the load is multiplexed across multiple links.
06-08-2010 12:16 AM
Joe
Just to add to Federico's excellent reply.
They all exist because they are all still needed. Rapid STP / Rapid PVST+ is not supported on some of the older switches that may still be running in Enterprise networks.
And PVST, PVST+, Rapid PVST+ are Cisco proprietary because running an instance of STP per vlan is a Cisco specific thing.
Jon
06-08-2010 12:28 AM
Jon
Thank you for response as well.
Do you think that anytime in the near future they will narrow it down?
Joe
06-08-2010 04:24 AM
joealbergo wrote:
Jon
Thank you for response as well.
Do you think that anytime in the near future they will narrow it down?
Joe
Joe
Unlikely i think although PVST+ and 802.1d should become largely unused due to Rapid-PVST+ and RSTP. If your switches support Rapid-PVST+ then there really is no reason to run PVST+ anymore.
Note also that Rapid-PVST+ is backwards compatible with PVST+ so you can run both in your network if you have switches that only support PVST+. If you do this then it becomes important to restrict the vlans that are allowed onto the PVST+ switches to only those needed. Any vlan that is present on both a PVST+ and Rapid-PVST+ switch has to revert to 802.1d timers so you lost the benefit of Rapid-PVST+ for those vlans.
Jon
06-08-2010 12:38 PM
Jon
I am going to have to look more into these protocols.
Joe
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide