All things related to multicast

[Bill Nickless <>]

-----BEGIN PGP SIGNED MESSAGE-----

At 04:20 PM 7/7/2003 -0400, Manfredi, Albert E wrote:
>I think that to make IP multicast compelling, ISPs have to be convinced 
>that it's to their advantage. One important piece MIGHT be to come up with 
>a scheme that can accurately count group membership. IGMP does NOT provide 
>such a technique as of now.

In the context of RFC 966, that's an excellent criticism.

Recall that the work described in RFC 966 was originally 
designed/implemented to support the V distributed system on a Metcalfe 
Ethernet.  Quoting:

    On a philosophical point, there has been considerable reluctance to
    make open use of multicast on local networks because it was
    network-specific and not provided across the Internet.  We were
    originally of that school.  However, we recognized that our "hidden"
    uses of multicast in the V distributed system were essential unless
    we resorted to dramatically poorer solutions - wired-in addresses.
    We also recognized, as described in this paper, that an adequate
    multicast facility for the Internet was feasible.  As a consequence,
    we now argue that multicast is an important and basic facility to
    provide in local networks and internetworks.

In the context of a Metcalfe Ethernet supporting a distributed system, 
group membership operations can be entirely contained in the host.  Each 
host merely discards data on the shared bus that is not addressed to them, 
or the Host Group addresses for which they have no interested processes.

In the context of multicast flooding between Metcalfe Ethernets, the 
requirement is to discover where there is zero vs. one (or more) Host Group 
Members on the subnet.  IGMPv1/2 were designed to meet this requirement.

In the context of today's IEEE 802 networks, with full-duplex links 
supporting a single host on each IEEE 802.1 bridge port, additional group 
membership information is required.  IGMP snooping/spoofing attempts to 
discover the set of IEEE 802.1 bridged segments with interested receiver(s).

My position on IGMP snooping/spoofing is probably clear to all.  :-)

But to address your point, I agree that we should look at more 
sophisticated group management techniques--at the link layer and the 
internetwork layer.

At the risk of repeating myself, the IP over Infiniband and IP over IEEE 
1394 (Firewire) specifications do a much better job of multicast group 
membership management than is commonly done on IEEE 802 networks today.

===
Bill Nickless    http://www.mcs.anl.gov/people/nickless      +1 630 252 7390
PGP:0E 0F 16 80 C5 B1 69 52 E1 44 1A A5 0E 1B 74 F7     


-----BEGIN PGP SIGNATURE-----
Version: PGP 8.0.2

iQCVAwUBPwnsGKwgm7ipJDXBAQFb3AP/W3i1SMaX+DC17qPYwlR45nvgGXe6kTs6
zAHtMdUika8Sav/Q/AXPpdnIPR+w33QULGn7f3rP8CfmeYxTHAcHxZxCc/SepMWu
BftaMEpVhtDokAEdsU9w2saz52FlW5+I9h6/eLTluGglU7ELbSeW4ZSGuJWMJUcQ
wQx4HfySbe8=
=nwRW
-----END PGP SIGNATURE-----