Subject: Re: Multicast ethernet question
To: Heiko W.Rupp <hwr@pilhuhn.de>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: tech-net
Date: 11/04/1997 18:59:45
"Heiko W.Rupp" <hwr@pilhuhn.de> writes:
>Jonathan Stone wrote:
>|I think the convention is that each protocols does its own checking,
>|and your driver doesn't need to at all.
>Yes and no. IP can do it for you. But if the card has already the possibility
>to filter out packets that no one wants, then there is less work to do at
>IP-layer.
Sorry, that's completely incorrect, as I explained in my original message.
The IPv4 mapping from class D addresses to Ethernet link-level
multicast addreses is a a many-to-one mapping. Ethernet hardware
filters only on the Ethernet link-level addresses. So the IP level
software *has* to check that the class-D addresses it receives are in
fact specifically requested, and not just addresses that map to the
same link-level bits. (Other parties on the same broadcast domain
might just be using a class-D multicast that maps to the same
link-level address, for example.)
>If you have to put the card into promiscous mode on a lan with much
>multicast traffic, performance will suffer.
Huh? Nobody is arguing this, but i don't see how it's at all relevant,
given the imperfect hash-based hardware filtering on the card in question.