[EFM-OAM] Loopback question
General comment on the current OAM draft. Forgive me if I'm bringing up
an old argument; I wasn't present, of course, when this was settled.
The loopback method chosen, to do a verbatim loopback and not swap the
MAC addresses, certainly has the advantage of simplicity for the passive
station. On the other hand, unlike all of the other protocols to date
(except for 802.3X Pause), it demands that you have a direct, physical
link between the MACs and not a virtual one.
If you look at all the virtual technologies available for Ethernet, past
and present: ATM LAN Emulation, Ether-over-MPLS, Ether-over-L2TPv3, IETF
VPLS, etc., most all of them involve some degree of learning MAC addresses.
Echoing verbatim will leave any such technology with the wrong idea of
what MAC address is where. It will seriously exercise any virtual
medium's MAC address learning capability while the loopback is taking place.
That's what I mean by "OAM demands physical media".
Just making sure that this factor has been discussed. The inability to
work over virtual media could limit the long-term viability of OAM.
-- Norm