From: Eric_Weaver@avtc.sel.sony.com (Eric Weaver)
Date: Tue, 26 Jul 94 14:18:49 PDT
To: wb8foz@nrk.com
Subject: CYPHERPUNKS TO THE RESCUE
In-Reply-To: <m0qSrax-000Gn6C@nrk.com>
Message-ID: <9407262117.AA23816@sosfc.avtc.sel.sony.com>
MIME-Version: 1.0
Content-Type: text/plain


   From: wb8foz@nrk.com (David Lesher)
   Date: Tue, 26 Jul 1994 18:51:19 +0000 (GMT)

   A challenge /response may make sense crypto-wise, but not $$-wise.
   The car would then need a receiver too, & the house a transmitter.
   More things to buy & break.

   A one-way solution is needed to make it fly here.

Okay, here's my "bright" "idea"...

A 32-bit counter, a 32-bit somewhat-random "salt", a 32-bit fixed
authenticator and a 32-bit checksum, two DES blocks.  The transmitter
just counts up each time the button is pressed, and the whole thing is
DES'd in CBC mode with the symmetric key or what have you.

The receiver decrypts, verifies the checksum and perhaps the
authenticator and just checks for the count to be greater than the
last time it received a signal.  This handles replays and doesn't
require exact sync between remote and base.

The receiver can have a reset-to-zero button inside so the owner can
push it and click the remote to re-sync if somehow the receiver gets
skipped way ahead.  This DOES, however, require different counters and
authenticators for different remotes.

It can also be done with a one-way hash if the salt is omitted and the
receiver can try, say, the next few dozen sequence numbers against the
received string.

Counterexamples, anyone?

Eric Weaver  Sony AVTC  3300 Zanker Road, MS 4B1  SJ CA 95134  408 955-4904
& Chief Engineer, KFJC 89.7  Foothill College  Los Altos Hills, CA 94022