[comp.virus] Infected networks

padgett%tccslr.dnet@mmc.com (A. Padgett Peterson) (06/11/91)

Last week I had occasion to disinfect another large network with the
Jerusalem (not ours - an outside company). The traditional respons is
to take down the net, clean the server, and check all of the clients
before reconnection.  On reflection, this seemed inordinately
inefficient so I came up with a new methodology which I offer for
comment. Note: this works for Jerusalem, Sunday, and non-stealth
infections which infect an executable before allowing it to run -
please be aware of this limitation up front.

The method was as follows:
a) take down net & clean server
b) remove non-essential applications
c) replace essential applications with a batch file that
   1) copies a clean selfcheck program from a writelocked directory
   2) runs the self check program
   3) runs the requested application

In this case I had such a self-check program (1400 bytes) that just
checks its own length & checksum. If it passes, the program exits, if
it fails, the client machine displays a warning message and is locked
up. In this manner, the server application files are protected from
infection (are never called by an infected client). Each client gets a
new copy of the "goat" file so clean clients are not affected, and
infected clients are identified.

Admittedly, this is a special case and directed to a small number of
viruses, but they seem to be the most common.

Comments ?
				Warmly,
						Padgett

p1@arkham.wimsey.bc.ca (Rob Slade) (06/15/91)

padgett%tccslr.dnet@mmc.com (A. Padgett Peterson) writes:

> In this case I had such a self-check program (1400 bytes) that just
> checks its own length & checksum. If it passes, the program exits, if
> it fails, the client machine displays a warning message and is locked
> up. In this manner, the server application files are protected from
> infection (are never called by an infected client). Each client gets a
> new copy of the "goat" file so clean clients are not affected, and
> infected clients are identified.

I have been reviewing a product from Bangkok called Victor Charlie
that takes a similar approach.  An intriguing concept.

I hope to be able to release the review shortly.

=============
Vancouver          p1@arkham.wimsey.bc.ca   | "If you do buy a
Institute for      Robert_Slade@mtsg.sfu.ca |  computer, don't
Research into      (SUZY) INtegrity         |  turn it on."
User               Canada V7K 2G6           | Richards' 2nd Law
Security                                    | of Data Security