From: "Mark W. Eichin" <eichin@cygnus.com>
Date: Thu, 17 Dec 92 15:59:41 PST
To: UFLTAI%MEMSTVX1.bitnet@CUNYVM.CUNY.EDU
Subject: re: abusing the system?
In-Reply-To: <9212172333.AA12866@toad.com>
Message-ID: <9212172358.AA07292@tweedledumber.cygnus.com>
MIME-Version: 1.0
Content-Type: text/plain


>> ps:  Any tips on tracing anonymous mail and newspostings?  I mean beyond the
>> "from" and "path" things... ie, trace to the userid...  Someone tried to forge
>> a posting in my name... (yes, that's what got me thinking :))
	Remember to look at the "Message-Id" -- on typical unix
mailers, that has the IP address encoded into it to help make it more
"unique". 
	A social point to keep in mind, though: one reason we really
*need* signed messages is because there is no real identity attached
to email. It is easy to "believe in" some identity you see on the net,
and for the most part enough of them are real that it is ok... but
I expect this to become even more of a problem than it is now without
signatures.
	A "historical" example -- at MIT, as part of Project Athena,
we have a real-time messaging system called Zephyr (for more details,
look in Usenix proceedings from some time in 87 or 88, or just look at
athena-dist.mit.edu:pub/usenix/zephyr.PS.) It optionally uses kerberos
authentication, and the recipient application will display whether a
message is authenticated or unauthenticated. People tended to ignore
this, until one of the other developers wrote a program that looked at
the database of current users, picked a pair at random, picked a
message at random, and sent it to one, from the other. (It backfired
amusingly once -- it sent a message from him, to me, saying "I'm
stopping at the coffeehouse, want me to get you anything?" to which I
responded sure... and then harassed him about it for years, until he
finally *did* bring me the M&M's I wanted. :-)
	The point was that this program didn't fake the authentication
(it did use privileged access to look at the user database, which is
not available remotely, but the messages themselves were
unauthenticated) but rather noone paid attention to it. The
"unauthenticated" flag was made more visible in a later release, I
believe... but I don't think anyone ever went as far as refusing
unauthenticated personal messages altogether. I could see that
happenning with email...
				_Mark_ <eichin@athena.mit.edu>
				MIT Student Information Processing Board
				Cygnus Support <eichin@cygnus.com>