5

I'm asking you today for helping me out with demistifying a question I have regarding the native vlan and a trunk link between a regular L2 switch and a router.

Let me explain quickly the scenario here:

We have a seperate network whose switch is a regular HP2824 hooked up with a SM fiber uplink coming from another building (the corporate network/internet). The trunk port 20 of this switch is hooked up to a regular router in our engineering network (isolated). This router serves us to route between multiple vlans/subnets we have for different sectors in the factory.

There is that trunk port going to the HP switch which confuses me a bit. I'm a cisco guy so i speak access/trunk links, not tag/untag. I'm getting used to it more everyday. On the HP switch, port 20 is untag vlan 8. vlan1 is not used on the HP side. So i'm assuming the native vlan of that port is 8.

On the router side, the link is considered an edge port (not configured as trunk) and its pvid is 69. In the factory, when you hook something in the vlan 69, you can pass through the corporate side and get internet access/ip phone, etc. Which makes sense. On switches that have computers needing an internet access, you must tag/allow vlan 69 to passthrough and untag the said port in 69 for their NIC.

I am wondering how the traffic is handled between the switch and the router? There is obviously a native vlan mismatch which leads to double-tagging/hopping. How in the 802.something world is this working? Router knows the destination network so it discards the vlan tag (which there isn't any) and routes traffic to next hop?

I'm wanting to change that router port to a trunk (because that's how to nature of things works in networking isn't? trunking two equipments together), leaving the native vlan 69 but i'm wondering what would happen on the other side with that port 20 leaving untag on vlan 8 ?? Should I tag port20 with vlan 8 accordingly?

(Need to mention there is no traffic restriction or acls on the network)

Thanks for reading me and taking the time to help!

Happy holidays!

1 Answer 1

3

Hope I understand what you are asking here (your writing is long and my memory is short :))

The HP untagged VLAN on a tagged interface functions similarly as the native VLAN you configure on Cisco switch trunk interface (actually tag/untag makes more sense than access/trunk technically speaking), the VLAN 8 in your case.

On your router interface facing HP switch, it is running as access/untagged port. So all ingress traffic on that router interface is VLAN 69; meantime router will remove VLAN 69 tag of the egree traffic to switch and when traffic arrives on switch, it will be VLAN 8. However if you have traffic marked with other VLAN, like VLAN 10 and send the traffic to router as the interface facing router is tagged/trunk port, router will not be able to understand it as the edge interface is expecting something without VLAN tag.

There is no native VLAN mismatch just like you connect a laptop to a trunk interface with certain native VLAN configured and switch does not complain about mismatch, right?

Hope this helps.

4
  • Indeed I wanted to give as many information as I can :) You're right saying tag/untag makes more sense because this is what's happening technically. In HP world, a trunk is a port aggregation. So I guess the router is stripping vlan "tag" information before forwarding the traffic to the next port? When the packet hits the interface (port 20 of the HP) it attaches the native vlan 8 for this side?
    – NewCT96
    Commented Dec 22, 2014 at 15:40
  • Yes, vlan tag is removed when leaving the router interface facing switch. 1) VLAN 8 tag is added when arrives at switch if no tag included; 2) VLAN tag info maintained if traffic already have the tag.
    – m1xed0s
    Commented Dec 22, 2014 at 15:49
  • That makes sense. I'm wondering of this works behind the scenes. Weither it's fabric or visible in any sort. I think a wireshark session would give me that answer. Have a good day.
    – NewCT96
    Commented Dec 22, 2014 at 15:54
  • Tagged / untagged is 802.1Q terminology. Usually you assign a different subnet for each VLAN, 10.0.8.0 for VLAN 8 and 10.0.69.0 for VLAN 69. Assume that the HP device port is untagged on VLAN 8 and tagged on VLAN 69; Cisco port access VLAN is 69 and VLAN 8 is trunked. The traffic will pass on both VLANs, but you have a PVID mismatch. The purpose of VLANs is limiting bcast domains, and if the VLAN tagging doesn't match (as in my example above) you will have bcasts from VLANs 8 and 69 spilling into each other. That's not a desirable situation, so standard (good) practice is to match the tagging. Commented Dec 22, 2014 at 23:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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