The IP over ATM Mailing List Archive by date

["Timothy J. Smith" <tjsmith@vnet.ibm.com>]

>> At the moment neither of these ideas exist in the IPmc draft, as its
>> primary focus is to provide an solution to a constrained multicast
>> problem. Whether they go into the draft depends on discussion here.
>>
>> cheers,
>> gja

Grenville, my thanks for pulling the discussion back on line.

We have had a number of interesting threads that have
spun off of the Broadcast topic, but I think the key is the
context of your IPmc draft.  It assumes no link layer broadcast
(else why would we need 1577 in the first place) which seems to
be reasonable for a target of the 3.0/3.1 UNI.  The IPmc draft
extends the ARP Server's role to handle IPmc in a way that
keeps each LIS separate and passes traffic only to those stations
participating in the mc group.  I was looking at the draft and
hoped to have it extended in a similar way to handle IPbc.

The topic of scalability leads to the need for some sort of
IP Broadcast and IP Multicast servers.  This topic is touched upon
in your Appendix C - Hidden Multicast Servers.  I expect the
IP Broadcast case magnifies the need for surrogates in this area.

TJS
p.s. I've been out vacationing, but I haven't lost interest