From: tcmay@got.net (Timothy C. May)
Date: Mon, 1 Apr 1996 15:43:14 +0800
To: cypherpunks@toad.com
Subject: Tamper-Resistance in VLSI
Message-ID: <ad8483aa060210049afd@[205.199.118.202]>
MIME-Version: 1.0
Content-Type: text/plain


At 12:06 AM 4/1/96, Mike Duvos wrote:

>However, I will observe that whenever technology is put forth for
>criticism on this list, there are always a few people who insist
>upon maintaining that anything can be easily defeated.
>
>"All you have to do is <blank>" they exclaim, where <blank> may
>be replaced by "Quantum Factoring", "SQUIDs", "Scanning Tunneling
>Microscopy", "NP=P", "The EPR Effect", "Nanomachines", or some
>other exotic notion which would be lucky if it had even achieved
>a laboratory demonstration under carefully controlled conditions
>much less a practical application to the problem in question.
>
>Common to all such claims is a gross underappreciation of the
>engineering difficulties involved, in this case those related to
>reading logic states buried in a densely integrated digital
>device without destroying them.  Something that isn't easy to do
>even if the device has been designed specifically for the purpose
>of permitting such observation in a laboratory environment.

While I agree that reverse-engineering/analyzing the internal states of
VLSI devices is much harder than some are claiming, it is not the case that
a chip must have been designed with this in mind for it to be possible.
When, then, is it possible, and when is it not? There is no simple answer;
I'd have to look closely at the device, its packaging, how many layers of
metal are involved, the size of the target node to be measured, and a raft
of other things.

>Such distractions, unfortunately, are why good physics rarely
>gets discussed in sci.physics, and why discussions on this list
>about nuclear bomb design, tampering, and hacking frequently take
>off in the crackpot direction.
>
>The bad eventually drives out the good, and few of the competent
>posters are going to continue to comment on a thread which has
>degenerated into the "You don't know anything.  Mr. Squid can
>read your smart card and your brain waves too" level of
>interaction.

Well, this is my third post tonight on this thread. I admit that it has
little to do with practical list issue (but then, what really does?).
Howvever, this happens to be an area of primary expertise for me (device
physics, voltage contrast, SQUIDs, sensing small charges, and
tamper-resistance), so I'm making comments to correct the various
misapprehensions here.

As to tamper-resistance, there is some exciting work being done on
"fingerprinting" of chips, some of which has been publically presented. I'm
under an NDA on some of this, but I can say that the cost of
reverse-engineering a smart card chip or satellite decoder chip is about to
take a quantum leap upward.

--Tim May

Boycott "Big Brother Inside" software!
We got computers, we're tapping phone lines, we know that that ain't allowed.
---------:---------:---------:---------:---------:---------:---------:----
Timothy C. May              | Crypto Anarchy: encryption, digital money,
tcmay@got.net  408-728-0152 | anonymous networks, digital pseudonyms, zero
W.A.S.T.E.: Corralitos, CA  | knowledge, reputations, information markets,
Higher Power: 2^756839 - 1  | black markets, collapse of governments.
"National borders aren't even speed bumps on the information superhighway."