[c-nsp] Are multicast MAC addresses allowed in the source field?

[Tomas Daniska]

> -----Original Message-----
> From: Benny Amorsen [mailto:benny+usenet at amorsen.dk]
> 
> Microsoft were by far not the first to do this, and I still believe that
> it is a brilliant solution to a difficult problem, even though we do not
> use it.

Maybe, still changes nothing on the fact that if they do it, they can't blame other vendors for not supporting. Being Cisco, I'd choose the RFC-compliant way, too. At least for the fact they could have changed it again just as soon as I put an effort to support their tweak...


And I don't find it brilliant - why on earth I should multiply an unicast request? That's a bandaid for a concept that's bad from scratch, not a proper solution.

> It is highly worrying if the 6500/7600 breaks this for layer 2 traffic.
> If we provide an EoMPLS link to a customer, it better be transparent.
> Support for this will definitely go into our next requirements document.

Actually, it's not for L2 switched traffic, but for L3 switched SVI-to-SVI, coming and leaving the same distributed etherchannel, on different LC's. That's the way I understood the combination of the conditions for the drop to occur.

The reason was a RFC-compliance check by the egress LC's ASICs.
 
 

--

deejay