toplogy with redundant links

[hmc2500]

I saw this topology in a network diagram somewhere and was wondering how this would work on a layer 2 level. Is this recommended? Would spanning tree automatically resolve it? I've had redundant links in real life scenarios but spanning tree would not shutdown the redundant links. In packet tracer this works. 

[Flavio Miranda]

Hi

 This is in fact a very common design most between Internet gateways and Cores, but usually L3 connection is used. But yes, this could be used in a layer2 topology and spanning tree will disable the necessay links if correctly configured to do so.

[hmc2500]

Could you elaborate more on the configuration? Would I have to modify costs on some links? Or should it be simple with spanning tree automatically enabling/disabling ports? 

[Joseph W. Doherty]

"Would I have to modify costs on some links? Or should it be simple with spanning tree automatically enabling/disabling ports?"

You should only need to modify STP "information" if you want to determine the L2 topology it should form; including for various failures.  Also, BTW, STP doesn't actually enable/disable port, it logically blocks them for ordinary data traffic.  BPDUs are still sent and received on blocked ports (as also might some other "special" traffic, like CDP?).

[M02@rt37]

Hello @hmc2500,

Full mesh topology with redundant links can be recommended for networks that prioritize high availability and fault tolerance, especially in smaller or less complex environments. However, larger networks may benefit from alternative topologies that strike a balance between redundancy, scalability, cost, and complexity. It's important to assess your specific network requirements and consider the trade-offs before deciding on a particular topology.

In this case, yes, STP is our friend.

[Joseph W. Doherty]

What you show is a full mesh topology.

As @Flavio Miranda already described, it could be used for L2 or L3.  L2 would need something like STP to eliminate loops.

Such a topology would be highly unusual to use for L2, and somewhat unusual for L3 (because many topologies tend to be hierarchal and/or more than "two" is unusual solely for redundancy).

[hmc2500]

I am asking this because in a real case scenario this did cause a loop for me. In packet tracer it worked fine. 

[Joseph W. Doherty]

"I am asking this because in a real case scenario this did cause a loop for me. In packet tracer it worked fine."

Possibly there was a issue with your "real case" configurations, also possibly why @Flavio Miranda wrote "if correctly configured"; or some other failure reason.  I.e. STP, on real equipment, should be able to preclude the multiple loops, such as PT appears to do for you.

[MHM Cisco World]

only Loop happened if I am right you config port-channel and here PO not work and sure L2 loop will happened 

[David Ruess]

As many have mentioned whether its L2/L3 the loop will be prevented by either STP for L2 or Routing protocols and their loop prevention mechanism. You mentioned it worked in PT but not your real life scenario. PT uses Cisco gear. That leads me to assume 1 of 2 things. STP was disabled on your devices (real gear) or you are not using Cisco real gear. Some vendors do not have spanning tree enabled by default and needs to be turned on. Could this be the case?

-David

[Joseph W. Doherty]

BTW, another possible issue you can bump into with STP is when you mix Cisco and non-Cisco switches.  Most vendor's STP is "common", i.e. one instance regardless of number of VLANs.  Cisco's STP is per VLAN.