Path: utzoo!attcan!uunet!aplcen!samsung!cs.utexas.edu!sun-barr!lll-winken!maddog!brooks
From: brooks@maddog.llnl.gov (Eugene Brooks)
Newsgroups: comp.windows.x
Subject: Re: Security extensions to X
Message-ID: <41053@lll-winken.LLNL.GOV>
Date: 9 Dec 89 07:03:16 GMT
References: <8912072350.AA26289@kanga.lcs.mit.edu>
Sender: usenet@lll-winken.LLNL.GOV
Reply-To: brooks@maddog.llnl.gov (Eugene Brooks)
Organization: Lawrence Livermore National Laboratory
Lines: 25

In article <8912072350.AA26289@kanga.lcs.mit.edu> jim@EXPO.LCS.MIT.EDU writes:
>There is growing interest within various members of the Consortium for an
>extension that will provide at least B2 security.  Given that the DoD has
>indicated that it would like to see secure window systems within the next 2-3
>years, there will probably be an effort starting up in the next year or so.
The issue has been given some thought here, the notion however is more general
that just X and covers the whole issue of security of compartmentalized
information on a network where any node can snoop for data without being
seen.  One possibility is to persue a similar tact as that used by SUN for
secure NFS.  Use DES (or better) encrypted TCP/IP, each pair of nodes for
which secure communication must occur share a key for the link.
TCP/IP packets evidently have source and destination address fields, and one
simply encripts the data in the packet according to the agreed upon key between
the two nodes.  Each link between two nodes has a unique key, and the encription
can be as nasty as is required for the data.  Some "spy" node listening to the
channel will only get hash.

This is probably not fool proof.  There are probably a lot of headers and
other frequently occuring patterns in the data transferred by X.  A good
example would be commonly used background tiles and such.  The classic
weakness of having both the clear text and the encrypted text for a message
does a good job a compromizing the key.  It is probably sufficient for
compartmentalizing information, where all people with physical access to the
network are suitably cleared.
brooks@maddog.llnl.gov, brooks@maddog.uucp