0

I have two switches (A and B), each one divided into two "Sub-switches", i.e. two VLANs.

Port 1 to 12: VLAN 100 (untagged)

Port 13 to 24: VLAN 200 (untagged)

Now, I want to connect the switches. Of course, I could configure one port to transport both VLANs (tagged) and connect both switches. Everything should work fine then.

But if I consider both parts of the switches as independent switches (they have completely isolates VLANs), I expected that it should be possible to connect each of the "Sub-switches" to its counter-part:

Port 1 of switch A to port 1 of switch B.

Port 13 of switch A to port 13 of switch B.

If I do that, it seems I've created a loop. I do not fully understand why this should be a loop, as I am working with different VLANs. If I was using two different switches instead of one switch configured with two VLANs, the setup would surely work. But where's the difference?

Best regards Tom

TomS
  • 175
  • 1
  • 1
  • 9
  • 1
    *If I do that, it seems I've created a loop* If those ports have NO tagged VLANs, and If both switches are configured identically, there should not be a problem – Ron Trunk Dec 30 '20 at 20:42
  • It would be helpful to include the make and model of the switch you are using, as well as the configuration file you have loaded into the switch. – Bert Dec 31 '20 at 12:46

1 Answers1

-1

That would also create a loop using two physical switches. A loop is created when you have more than one path between two endpoints.

You can use one cable to connect two switches, typically referred two as a daisy chain. You could connect them with two cables, typically called a ring, but only if you have STP (spanning tree protocol) configured.

The confusion lies here (and you can verify this with a packet sniffer tool such as Wireshark): If you have an untagged port, the switch will strip the VLAN tag from the outgoing packet. So the traffic is just 'plain' network traffic without any VLAN information.

Bert
  • 2,863
  • 12
  • 13