Do you know a unicast protocol for the Ethernet control plane?

[hostettle]

Hi,

Does anyone know a protocol for the Ethernet control plane which has a unicast destination address?

MVRP, MMRP, MSTP, RSTP, all these protocols have a multicast reserved destination address.

Perhaps we have to look non-802.1Q control plane protocols.

Best regards,

Michel

[Peter Paluch · Accepted solution]

Hi Michel,

Unicast-based control plane protocols in Ethernet appear to be extremely rare, mostly due to the fact that they are supposed to work automatically, with no user intervention, hence they can not assume any static destination MAC to be configured - so they resort to using multicasts or broadcasts instead.

In any case, think of LOOP frames sent by Catalyst switches to detect self-looped ports. In these frames, the source and destination MAC address are set to the unicast MAC of the egress port.

I am trying to think of another Ethernet management protocol - I'll be sure to post it here when I recall any. I wonder if any of the OAM protocols, especially the one providing the loopback/ping test is unicast-based.

Best regards,

Peter

[hostettle]

Hi Peter,

The LOOP frame is a really good case. It's a non-802.1x protocol which is also client of the MAC sublayer;

I also have to check if OAM protocols are unicast-based. But... I had a personal constraint more, finding a control plane protocol unicast-based, which is also CLIENT of the MAC sublayer.

OAM protocols are located inside the MAC sublayer, below the switching matrix.

Best regards,

Michel

[hostettle]

Hi Peter,

> I wonder if any of the OAM protocols, especially the one providing the loopback/ping test is unicast-based.

In G.8013 (07/2011) section 7.3:

"The Ethernet loopback function (ETH-LB) is used to verify connectivity of a MEP with a MIP or

peer MEP(s). There are two ETH-LB types:

• Unicast ETH-LB.

• Multicast ETH-LB".

> In any case, think of LOOP frames sent by Catalyst switches to detect  self-looped ports. In these frames,

> the source and destination MAC  address are set to the unicast MAC of the egress port.

As I said above, it's a good case for my little study.

The LOOP frame, from Cisco, was certainly interesting and important before 2004.

Since 802.3ah-2004 we have the OAM remote loopback (in link OAM, and not network OAM as ETH-LB).

Best regards,

Michel