1998-08-18 - Private Doorbell - Re: CRYPTO-GRAM, August 15, 1998

Header Data

From: Bill Stewart <bill.stewart@pobox.com>
To: Bruce Schneier <schneier@counterpane.com>
Message Hash: bd956d0120704b476b920adbab13194790aecdcb636a0e0d1bec3f0482a66766
Message ID: <>
Reply To: <199808151730.MAA05817@mixer.visi.com>
UTC Datetime: 1998-08-18 07:39:27 UTC
Raw Date: Tue, 18 Aug 1998 00:39:27 -0700 (PDT)

Raw message

From: Bill Stewart <bill.stewart@pobox.com>
Date: Tue, 18 Aug 1998 00:39:27 -0700 (PDT)
To: Bruce Schneier <schneier@counterpane.com>
Subject: Private Doorbell - Re: CRYPTO-GRAM, August 15, 1998
In-Reply-To: <199808151730.MAA05817@mixer.visi.com>
Message-ID: <>
MIME-Version: 1.0
Content-Type: text/plain

Bruce Schneier wrote in his newsletter
> It seems that every few months we get key-escrow repackaged with a new name.
> The latest new name is "Private Doorbell," and the spin is that the keys 
> are escrowed in the routers.  Other than the name, there's really 
> no difference between this and other key escrow schemes: 

Based on the talk by Elizabeth Kaufman from Cisco at the Bay Area
Cypherpunks meeting, I'd say you're wrong.  (I'm pleasantly surprised;
this is the first "compromise" I haven't rabidly disliked.)
(Elizabeth says not to blame her for the name, it's Cisco's PR folks' fault.)

The Private Doorbells proposal says that routers that encrypt
have a clear side and a black side, and if you want to wiretap them,
you can already wiretap the clear side, so Louis Freeh should be happy enough
(since it meets the "legitimate needs of law enforcement",
even though it the blocks no-warrant no-knock wiretaps he'd really like) 
so just give us the export permits we want and stay off our backs.
There's no special key escrow in the routers, it's just
doing what encrypting routers do already, and if Big Louie wants
to wiretap an ISP today, he can try to get a warrant and then do it.

It seems to me that there's a minor catch - routers with
multiple private-line encrypted interfaces can decrypt and
encrypt traffic without ever hauling it through an Ethernet
where it'd be easy to tap, but most of those configurations
still have an Ethernet somewhere, and any Bad Guy who's doing encryption
at the endpoints ought to be using end-to-end encryption anyway.
Bill Stewart, bill.stewart@pobox.com
PGP Fingerprint D454 E202 CBC8 40BF  3C85 B884 0ABE 4639