RE: [EXTERNAL] Re: Interested in wireless ?

["Manfredi (US), Albert E" <[e-mail]>]

-----Original Message-----
From: ipv6 <[e-mail]> On Behalf Of Alexandre Petrescu

Le 02/06/2020 à 14:04, Pascal Thubert (pthubert) a écrit :
>> 
>> There's a difference between the MAC address used (0x333300-blah)
>> that is multicast and the effective processing by the network.
>
> I agree there is that difference.
>
> The question is then whether we should call MAC-layer broadcast the
> effective processing by the network?
>
> For my side, I think the effective processing by the network is the fact
> that messages are received by everyone and local filters are used.  It
> is 'broadcast' in nature, but it is not the IEEE broadcast.

Agreed. The important distinction being, MAC broadcasts (all 1s MAC DA) have to be filtered at layer 3, whereas MAC multicasts are filtered by the Ethernet NIC, for instance, at layer 2. I don’t think it's fair to say that MAC multicasts are truly broadcast.

>> For all I know, the IEEE multicast protocols are pretty much never
>> deployed, and only the first bit that says broadcast is of actual
>> interest. IOW the IETF has thrown the problem of handling multicast
>> for a large number of groups over the fence to IEEE. The end result
>> is a broadcast to all nodes. True on both IEEE std 802.3 and 802.11.

I was wracking my brain trying to remember what the MAC layer multicast join protocol was called. Anyway, it does appear that it's "never actually used." However, what is used is the IGMP snooping hack, which prevents MAC layer multicasts from being flooded to all switch ports, by default. Layer 2 switches behave like layer 3 busybodies, listening to see which ports have IP multicast group membership. So in fact, that makes layer 2 switches avoid having to flood multicasts to all ports.

Bert