Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!ames!ucbcad!ucbvax!ACC-SB-UNIX.ARPA!kzm
From: kzm@ACC-SB-UNIX.ARPA (Keith McCloghrie)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Broadcast/Multicast on ARPANET
Message-ID: <8708201334.AA17184@ACC-SB-UNIX.ARPA>
Date: Thu, 20-Aug-87 09:34:32 EDT
Article-I.D.: ACC-SB-U.8708201334.AA17184
Posted: Thu Aug 20 09:34:32 1987
Date-Received: Sat, 22-Aug-87 10:06:15 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 19

Phil,

You point out a number of the ways in which the capability of
providing subnet broadcasting or multicasting in the ARPANET could
be very useful.  I agree.  However, I suggest that there would need
to be some safeguards installed to prevent the ARPANET from becoming
susceptible to the kind of meltdowns which occur in LANs when this
capability is abused.  The absence of this capability in the wide-area
currently provides a firebreak between LAN-subnet clusters, to prevent
a meltdown from encompassing the whole internet.

Possible safeguards could include : only administratively-authorized 
Host-ports should have this capability and even the Hosts connected
to those should be required to prove that abuse will not occur;
restrictions on the types of packets/destination-addresses which
are to be forwarded by gateways using this capability.  Maybe,
only multicasting should be supported (not broadcasting) ?

Keith.