Re: clipper pin-compatible chip

Remember folx.. ya heard it here first.. Lesse.. was
quite a while ago.. the day/or after that it was announced I believe.
when ideas come full circle I can't help but think we missed one
go-around and this is another reminder to do it.  Let's
plan ahead so we don't get trapped without an escape.
 
TTFN.
 
-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
- DrZaphod                #Don't Come Any Closer Or I'll Encrypt!   -
- [AC/DC] / [DnA][HP]     #Xcitement thru Technology and Creativity -
- [drzaphod@brewmeister.xstablu.com] [MindPolice Censored This Bit] -
-         50 19 1C F3 5F 34 53 B7   B9 BB 7A 40 37 67 09 5B         -
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

DrZaphod writes:
 
>  	Remember folx.. ya heard it here first.. Lesse.. was
> quite a while ago.. the day/or after that it was announced I believe.
> when ideas come full circle I can't help but think we missed one
> go-around and this is another reminder to do it.  Let's
> plan ahead so we don't get trapped without an escape.

Yeah, a lot of folks suggested the same thing (I didn't, so I'm
neutral on the credit issue). It was debated in scy.crypt, as well as
here on Cypherpunks.

Reverse-engineering the Clipper/Capstone/Skipjack chips to the point
of allowing drop-in replacements would entail several things:

1. Access to the Clipper chips. I don't think the AT&T phones ("Reach
out and tap someone") are available yet...reports of yield problems,
delays, unhappiness at AT&T with the pace of deliveries, etc. I don't
know if Mykotronx, or their chip supplier (VLSI Technology Inc.) is
selling chips to end users....Arthur Abraham follows this and may
know.

So, getting enough chips is the first step.

2. "Peeling" the chip is the next step, that is, getting access to the
innards of the chip. Reports are that VTI is using various
tamper-resistant technologies to make peeling the chip harder. (My
guesses: etch-resistant epoxies, increased glass passivation/scratch
protection, and judicious use of 3-layer metal processess to make
probing difficult.)

3. Remember, testing the chip and attaching logic proves is generally
not sufficient to deduce the internal layout. Emulation cannot be done
this easily. The circuitry must match up to some extent (not exactly,
but more than just partial emulation).

4. I expect the Skipjace algorithm cannot be "faked" without knowing
the algorithm, and more. Several articles on this have alluded to the
difficulties in producing plausible-looking LEEFs (Law Enforcement
Exploitation Fields) without greater knowledge of the algorithms and
the keys assigned (I believe, but I could be wrong) to the specific
chip.

5. My guess is that a reverse engineering job on Clipper would cost
more than $200K for a commercial service to do, perhaps cheaper for
some motivated grad students at Berkeley or Stanford to do. (But then
they probably are thinking about other things.) And the job would
still have a fair chance of failing, not being a good enough
emulation, etc.

I suspect this is why such a project isn't going anywhere. The cost
and expertise needed, and the basic lack of a motivation.

The lack of motivation comes from the feeling of many of us that the
talk of Clipper, of weaknesses in it, etc., is misguided, that the
proper approach is reject key escrow on general principles. Convincing
businesses and individuals to steer clear of Clipper is an easier win.

(And if unapproved crypto is outlawed, all bets are off. Emulations
are not likely to pass muster, either.)

Finally, I've heard no repudiation of the other idea many folks have
suggested, some the day Clipper was announced (me, for example), and
some very recently: encrypt your data with your own algorithm prior to
using the government's system, should it be mandated. Unless they open
your packets, they'll never know. If they open your packets and
complain, claim it was something like bits for a test of randomness,
etc. (Too many variations on this to go into now. Suffice it to say
that outlawing the sending of bits that they can't "understand" is too
hard to enforce, even with foreseseable trends.)

But I don't want to discourage DrZaphod from launching such a project
to reverse-engineer the Clipper chip....I wish him luck on this
effort.


--Tim May


-- 
..........................................................................
Timothy C. May         | Crypto Anarchy: encryption, digital money,  
tcmay@netcom.com       | anonymous networks, digital pseudonyms, zero
408-688-5409           | knowledge, reputations, information markets, 
W.A.S.T.E.: Aptos, CA  | black markets, collapse of governments.
Higher Power:2**859433 | Public Key: PGP and MailSafe available.

Actually, all you need to know is what pins do what, and what the protocol
of those pins are in terms of their communication with the rest of the
board.

My suggestion was to make a plug in chip that replaced the clipper
chip, however, that chip doesn't need to actually be able to talk to
another clipper chip.  You don't need to disassemble a real clipper
chip to be able to do that.

In order for Uncle Sam to be able to spread the clipper and not
give AT&T any idea about what it actually contains, it has to give
AT&T some interface specs.   It has to provide this information to
any company that intends to use the clipper chip.

Just as you don't need to know the microcode for an Intel 486 to
build a clone, you don't need to know what the guts of the clipper are
to replace it with another chip which doesn't have to be compatible 
with it in terms of communicating with another clipper.  It should
rather only need to communicate with another of its kind be it an
IDEA chip or an RSA chip, etc.

All things considered, a simplified approach at looking at a clipper
would be that of a filter or a pipe.  Data goes in at one end, some
other data goes out the other end.  All you need to do to replace it
with another filter that does a similar (but not identical) job is to
find how the clipper talks to the outside world, which Uncle Sam
has to provide if he wants this chip to take off the ground.  In which
case, once we have specs, we can build an interface to an IDEA chip, etc.
 
Another example: with old old Macintoshes there was a thing called a
"Killy Clip" which looked like a laundry pin that attached over a 68000
CPU and took over its bus to an accelerator card, or some other system
expansion card.  Such things were originally frowned upon by Apple and
mostly Steve Jobs, but they were popular at the time.  They worked.
And they didn't need to take the 68000 apart to the silicon.  Just
attached to its pins and took over from there.

Couldn't the same thing be done with Clipper?
A pin compatible, bus-protocol compatible, but communications
incompatible chip attached to a switch that lets you select between the clipper and the replacement?

> In order for Uncle Sam to be able to spread the clipper and not
> give AT&T any idea about what it actually contains, it has to give
> AT&T some interface specs.   It has to provide this information to
> any company that intends to use the clipper chip.

Why don't you just call Mycotronix up and ask for data sheets on the
Myk-78 and Myk-80.  I believe that they are located in Torrence CA.

Eric

> etc. (Too many variations on this to go into now. Suffice it to say
> that outlawing the sending of bits that they can't "understand" is too
> hard to enforce, even with foreseseable trends.)

	Which reminds me..

	I've recently started taking a class for the "American
Cultures" requirement here at UC Berkleey, about the "Languages of
America." In any case, I've been reading numerous articles about cases
in which people were *penalized* (mostly in school) for speaking to
their peers in their own native language as opposed to English, as
mandated by law.
	People have said, "Outlawing strong crypto is like telling you
tha you have to speak in a language they can understand." Don't be so
surprised-- it's been done before. (Not to as an extreme degree as
we're thinking about, but it *has* been done.)

I don't think the idea proposed is to reverse-engineer the Clipper.
Rather, the idea is that once you know the pin-out you can make an
electrically-compatible (and, in important ways, software-compatible)
replacement.  You'd then have a phone that does IDEA encryption
instead of Skipjack, and presumably the non-crypto parts of the phone
wouldn't know and wouldn't care.

--
| GOOD TIME FOR MOVIE - GOING ||| Mike McNally <m5@tivoli.com>       |
| TAKE TWA TO CAIRO.          ||| Tivoli Systems, Austin, TX:        |
|     (actual fortune cookie) ||| "Like A Little Bit of Semi-Heaven" |