My L3 switch has ~30 interfaces. If you look at the output below there are two MAC addresses showing in the results of the show interface command. The hardware address is the same for all interfaces. The physical address is different for each interface.

Assume that an IP address was configured on interface ce1/1. When ce1/1 receives an ARP request for the IP configured on it, which MAC should be used in the ARP reply?

What are the potential issues if I use the hardware address rather than the physical interface address? Is there any standard talking about this situation?

rtr1#sh interface
Interface ce1/1
  Scope: both
  Flexport: Breakout Control Port (Active): Break Out Enabled
  Hardware is ETH  Current HW addr: 4c76.2541.a6c1
  Physical:4c76.2541.a6f2  Logical:(not set)

  Interface ce1/2
  Scope: both
  Flexport: Non Control Port (Active)
  Hardware is ETH  Current HW addr: 4c76.2541.a6c1
  Physical:4c76.2541.a6f3  Logical:(not set)
  • 1
    At the very least, what is the switch model?
    – Ron Maupin
    Jun 5, 2018 at 6:52
  • It is DELL Z9100 Jun 5, 2018 at 7:38
  • In what way would you use (or be able to choose) which address is used? What's the actual problem you are you trying to solve here?
    – Teun Vink
    Jun 5, 2018 at 8:17
  • I'm a networking professional. I know How ARP works. Here I'm not asking anyone to solve a problem. 1) Some vendors use the same mac (HW addr) in ARP reply for all interfaces. 2) some vendors use individual physical addresses of Interfaces for respective network segments. which one of these 2 approaches is best..? Jun 6, 2018 at 9:16
  • If you have loop connecting 2 VRFs in same device, approche 1) won't work. Is there any documnetation and Use cases that supports approch 1) ..? why some vendors chose approch 1)..? Jun 6, 2018 at 9:18

2 Answers 2


ARP (Address Resolution Protocol) is used to resolve a layer-3 address to a layer-2 address. An ARP reply from a layer-3 interface would use the configured layer-2 (MAC) address assigned to the layer-3 (IP) interface for the layer-3 address you are trying to resolve.

I don't see where you have a layer-3 address assigned to either interface in your question, so you would not be using ARP to resolve a layer-3 address to a layer-2 address for either of those interfaces.

Switches typically only have layer-3 addresses assigned for switch management (or routing for layer-3 switches on layer-3 interfaces), and the layer-3 addresses play no part in the layer-2 switching. Even layer-3 switches are primarily layer-2 switches, and the layer-2 interfaces do not use ARP because they have no layer-3 address needing to be resolved. Layer-2 uses MAC address tables to determine where to forward frames, not ARP tables, and the switch is a transparent device at layer-2.

  • I've edited the question, please check this once. Jun 6, 2018 at 9:20
  • The answer is actually the same. Each layer-3 interface will have a unique layer-2 address. That means it will use the physical address.. The HW address is probably for use with layer-2 STP, which doesn't run on a layer-3 interface.
    – Ron Maupin
    Jun 6, 2018 at 14:28

The MAC that will be used for an ARP request will be the one associated with interface that has the ip address on it, when a switch has. L3 capability and you make an Interface of it a L3 interface it will will work its way from L3 to L1 so it will reply according to ARP defined RFC , now to your question if there will be an issue if you use the "hardware address" which I would rather call it the chassis address : 1) do you have an option to configure this? If you do what will be the purpose of it? , chassis or hardware addressesses are more for taking decissions on spanning tree , or to link that mac adrress to a management ip.

Your Answer

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

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