[107719] in Cypherpunks
Re: CDR: Re: Adieu Privacy: Intel identifiziert Chips
daemon@ATHENA.MIT.EDU (Martin Minow)
Thu Jan 21 18:46:41 1999
In-Reply-To:
<Pine.BSF.3.96.990121213606.18731A-100000@pakastelohi.cypherpunks.to>
Date: Thu, 21 Jan 1999 15:25:30 -0800
To: Lucky Green <shamrock@cypherpunks.to>, cypherpunks@einstein.ssz.com
From: Martin Minow <minow@pobox.com>
Cc: Jim Burnes - Denver <jim.burnes@ssds.com>
Reply-To: Martin Minow <minow@pobox.com>
At 21:40 +0100 1999/01/21, Lucky Green wrote:
>
>Since CPU's lacking the ID features will be unable to run the modern
>applications of the future, market pressures will force manufacturers of
>Intel compatible CPU's to add equivalent identifers.
>
This will, however, depend on the success of the Open Source
movement. Also, vendors of cross-platform applications will
need to avoid -- at least in the short term -- o.s. and
hardware-specific capabilities that are not present on the
complete range of machines of interest.
On the other hand, in the medium to long term, all CPU's
of any reasonable size will be networked and, hence,
identifiable by their network ID.
I suspect that one significant reason that application vendors
are interested in a CPU-specific ID is to prevent application
privacy: your application will not run until the installer
contacts a secure server and retrieves a hardware-specific
unlocking key. To take this one step further, it's no
unreasonable to imagine applications distributed using
an instruction set that is cryptographically keyed to the
CPU id (and with the decryption taking place on-chip in
the instruction decoder).
Martin Minow
minow@pobox.com
