[5663] in cryptography@c2.net mail archive
Re: Intel RNG
daemon@ATHENA.MIT.EDU (Eugene Leitl)
Fri Sep 17 18:26:59 1999
From: Eugene Leitl <eugene.leitl@lrz.uni-muenchen.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <14306.45564.757199.593527@lrz.de>
Date: Fri, 17 Sep 1999 14:26:20 -0700 (PDT)
To: Arnold Reinhold <reinhold@world.std.com>
Cc: Anonymous <nobody@replay.com>, cryptography@c2.net
In-Reply-To: <v04210100b408455d8bb1@[24.218.56.100]>
Are there any other advantages in a hardware PRNG other than it cannot
be overwritten? (Yes your hardware might be incorruptible but the
software layers always be). I could imagine the soon-to-arrive (you
might disagree but the writing's on the wall) CPUs with considerable
FPGA areas will make purely hardware PRNG obsolete.
Analog FPGAs might be even useful for true RNG.
Arnold Reinhold writes:
> I do not see anything "reasonable" in the excuses Anonymous
> attributes to Intel not allowing access to raw RNG bits. If Intel
> wants developers to use their RNG API they need only publish it.
> Professional programmers these days respect APIs and realize they
> risk future problems if the do not follow them.
