home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.security.misc
- Path: sparky!uunet!pmafire!mica.inel.gov!ux1!news.byu.edu!news.mtholyoke.edu!jbotz
- From: jbotz@mtholyoke.edu (Jurgen Botz)
- Subject: Mail forging
- Message-ID: <BxrJCL.9sC@mtholyoke.edu>
- Sender: news@mtholyoke.edu (USENET News System)
- Organization: Mount Holyoke College
- References: <1992Nov10.201503.6288@gw.wmich.edu> <BxJ80A.KAr@mtholyoke.edu> <4225@bcstec.ca.boeing.com>
- Date: Sun, 15 Nov 1992 15:07:32 GMT
- Lines: 47
-
- In article <4225@bcstec.ca.boeing.com> vanzwol@bcstec.ca.boeing.com (Ted Van Zwol) writes:
- >OK, so if you won't say how to do it (possibly reasonable enough...)
- >Can you say how to prevent it?
-
- With the current set of Internet standard protocols, there is no way of
- preventing forged mail. The technology to do that exists with public-key
- crytography and electronic signatures, but since there are not yet any
- widely implemented standards for using this technology on the net, it is
- not practical for an individual to do so.
-
- >Can you say how to detect it?
-
- Detection is trick... if the message was a forgery of a "local" message on
- a Unix system, you can probably detect it from the sendmail logs; if they
- message were authentic it would not have passed through SMTP. If the
- forgery was one where a user spoofed a user on a remote system, you might
- need cooperation of the sysops of other systems along the route for
- reliable detection, making it practically very difficult if the forger
- was clever.
-
- Note that the former covers the case Mr. caps-and-exclamation-marks was
- asking about; it is quite easy to detect the case where a user of a multi-
- user Unix system tries to spoof a message from "root" (however, detection
- will be from the logs, not the message itself.)
-
- >Is it truly that trivial and foolproof? Or does it just make the
- >header *look* like it came from someone else (read root) to the
- >non-attentive eye?
-
- If done correctly, the header of a forged message will be *exactly* identical
- to the header of the message had it been authentic. If there's any way of
- identifying the forgery in such a case it's via logs of network activity,
- not the message itself. And yes, if you have a resonably good understanding
- of what a correct message header /should/ look like, it's quite trivial to
- execute such a forgery.
-
- Note however that in many cases of actual forgeries the headers might
- provide a clue to the fact that the message is a forgery because a lot
- of the forgers do /not/ really understand what a correct message header
- looks like, or conversely, they produce a "correct" looking header but
- the real system would have produced different or less correct one.
- --
- Jurgen Botz | Internet: JBotz@mtholyoke.edu
- Academic Systems Consultant | Bitnet: JBotz@mhc.bitnet
- Mount Holyoke College | Voice: (US) 413-538-2375 (daytime)
- South Hadley, MA, USA | Snail Mail: J. Botz, 01075-0629
-
-