Fastest Spanning tree in multiple-ring topology

[azi123]

Hi everybody!

I have a topology in which there is the main ring and some nodes in the main ring have a minor ring. (see attachment please)
1- which STP configuration has the fastest convergence time? How would you configure spanning tree for all switches?
I think RSTP does the best, but Is there any other configuration to help convergence time?

2 - Does RSTP have any limitation? limitation In number of switches for instance? or incompatibility with other vendor's RSTP or spanning tree algorithm?

3- Is RSTP practical solution for this topology with a large number of switches? 45 switches as the main ring and more than 100 switches in minor rings?

4- Is there any other faster way, rather than spanning tree, to provide protection for this large network? 

Thank you all! 

[Paolo Moserle]

Hi azi, good question.

I'm really interested on it, because you can read some different version on web and talking to several colleagues in the sector there are different opinions.
Anyway I can try to help you based on my experience:

1- which STP configuration has the fastest convergence time? How would you configure spanning tree for all switches?
I think RSTP does the best, but Is there any other configuration to help convergence time?

Customization of STP in this topology can will be a nightmare, and I discourage this.

2 - Does RSTP have any limitation? limitation In number of switches for instance? or incompatibility with other vendor's RSTP or spanning tree algorithm?

This is the most interesting question, in theory RSTP not have the 7 hope (on max age default parameter) limitation like STP.

I haven't any negative experience in the interoperability with it.

3- Is RSTP practical solution for this topology with a large number of switches? 45 switches as the main ring and more than 100 switches in minor rings? 

4- Is there any other faster way, rather than spanning tree, to provide protection for this large network? 

In my opinion RSTP is not the best choice, MRP or REP can help in this type of ring topology, but 45 switch is very a lot number on the same ring. The limitation of REP is 25 and MRP is 50 I guess. However I suggest you to divide the topology in more rings.

I'm really interested if anyone can suggest anything other.

[Joseph W. Doherty]

#1 Yes, you would want a rapid-STP variant over original STP.

#2 Yes, regarding other vendors. Cisco's rapid-STP is per VLAN where as many other vendors rapid-STP is common to all VLANs. MSTP, though, can be a better choice between vendors.

#3 It might be practical although there are "better" technologies (e.g. like perhaps SPB), unfortunately your equipment might not support them especially if dealing with different vendors.

#4 On Cisco equipment, as mentioned by Paolo, REP might be "better".

[Paolo Moserle]

Thanks for your replay Joseph, but based on your experience you would recommend using of RSTP on this type of topology?
In the real world are there some limits related to the hop in RSTP?

Thanks.

[Joseph W. Doherty]

My experience with larger network topologies have been all L3, so I have no direct experience for those. of a size like yours, that are L2 relying on STP. The mid-size L2 topologies I've dealt with, having some L2 redundant paths, were often a rat's nest, very much not rings. Those and the smaller L2 topologies, generally w/o L2 redundant paths, I've supported have been had either STP or rapid-STP, the latter does work "better" (in fact, I would often upgrade such topologies from STP to rapid-STP).

As I understand it, rapid-STP doesn't have the same "diameter" limitation of STP.

[Paolo Moserle]

Thanks Joseph, for your answer!