multiple MACs on a port

is it allowed by ethernet-standards that a switch has multiple (actually two) MACs on the same port? the port is part of a VLAN with a L3-Interface i.e. an IP-Address assigned. I am seeing the following effect:

When an IP-packet is send to and through the switch it answers arp-requests with its first MAC-address, the answer packet however comes from the second MAC-address back to the client.

AFAIK it doesn't matter from wchich MAC an IP-packet is returned, as the IP-layer doesn't care.

(btw. its an Extreme Summit 48si Switch configured for soft rate limiting and Extreme said its a feature of the chipset)

Reply to
Loading thread data ...

matthias wrote in part:

Certainly should. switches were originally very expensive and used for network segmentation. Hubs were hung off their ports.

It depends on the sophistication of the switch electronics how many MACs it can search before resorting to hub behaviour -- pushing the frame out on all ports.

A switch should not alter sending or receiving MACs. A bridge might and a router will. A proper router has one MAC per port. Many SoHo routers are really a dual ported router (still more than a bridge) with an integral switch attached to the LAN side.

-- Robert

Reply to
Robert Redelmeier

A network transparent bridge is capable of generating packets that have whatever MAC address is needed forward each packet. A PC with two ethernet cards and running Linux can be set up to do this.

Reply to
Al Dykes

the switch is transparent - so the MAC may be tied to a particular port or not - if the packet can go between ports internally, then how could you tell by observing traffic?

lots of bigger switches have a block of addresses to use - 256 on some Nortel gear i worked on and more on some others.

They tend to get assigned to logical functions, and depending on the switch there may be more than 1 associated with a port. eg maybe 1 for the "bridge" MAC allocated to a port, but a different MAC used for L3 on that port. The Nortel Passport 8600 model is sort of a bridging function and then a logical router function behind that, and the MACs seem to correspond to those logical functions.

The standards seem to be much more about how MACs behave and dont really care.


The switch logic often uses the MAC to decide whether a packet should be routed or not - if the dest MAC is the logical function for the IP routing engine on that port then it gets routed, otherwise it is handled differently.

Reply to
stephen Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.