From: "Vladimir Z. Nuri" <vznuri@netcom.com>
Date: Thu, 30 Oct 1997 10:26:32 +0800
To: cypherpunks@cyberpass.net
Subject: new intel verification circuit & crypto
Message-ID: <199710300215.SAA13323@netcom7.netcom.com>
MIME-Version: 1.0
Content-Type: text/plain



John Dvorak has an interesting column in a recent PC
magazine on the possibility that Intel has created
some new circuits in its new chips that would allow
alteration by software. this is to deal with a 
"verification crisis" in which the complexity of
their large chips is outstripping their ability
to test for flaws and make them perfect. according to
Dvorak the scheme would allow Intel to release a 
software program that would actually alter the
circuitry of the chip.

information at this point is very limited; Dvorak says
that Intel doesn't want to announce the scheme for fear
that it may not work in an embarrassing scenario.

he says that the scheme is protected using cryptographic
algorithms, so that hackers would be prevented from altering
the chip. he raises the spectre of a virus that could
do permanent, irreparable damage to the chip.

however, I have some big red flags going off in my brain
the more I think about this. it seems to me that no matter
what kind of cryptography is used on the chip, there is
the possibility of reverse engineering via dissassemblers
& debuggers etc. if the software is distributed widely.
hackers have long experience with this, and I suspect if
a fix was released, the code could be cracked quite 
readily.

the only other possibility is that Intel would require
people to bring the chip into a service dealer. but this
would decrease the utility and convenience of the fix
feature, and I doubt they could protect the distribution
of the software in this case either.

anyway, an interesting topic of conversation for this
list, no?