Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!iuvax!noose.ecn.purdue.edu!ee.ecn.purdue.edu!hankd
From: hankd@ee.ecn.purdue.edu (Hank Dietz)
Newsgroups: comp.arch
Subject: Re: Randomised Instruction Set Computer
Summary: High overhead for lousy encryption
Keywords: viruses
Message-ID: <1990Jun4.155940.5547@ecn.purdue.edu>
Date: 4 Jun 90 15:59:40 GMT
References: <3131@goanna.cs.rmit.oz.au>
Sender: news@ecn.purdue.edu (USENET news)
Organization: Purdue University Engineering Computer Network
Lines: 15

In article <3131@goanna.cs.rmit.oz.au> ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) writes:
>The crazy idea, then, is to ensure that a virus can't be sure of knowing
>what bit pattern to generate in order to realise its effect.  What if we
>tried to make a computer hard to effect by giving it a randomised
>instruction set?
[stuff about using table-lookup to re-map instruction set for each program]

Can certainly be done, but table lookup isn't exactly the hardest encryption
scheme to break.  Compilers tend to generate recognizable instruction
patterns (unless they have really good optimizers), so the virus would only
need to be a bit more clever to defeat all that expensive & slow extra
hardware....  Of course, this is just my opinion.  Modal instructions might
make things more difficult, however, without the need for a full table....

						-hankd@ecn.purdue.edu