1993-06-02 - Re: Crypto anarchy in a VW? (not the bug)

Header Data

From: Jim McCoy <mccoy@ccwf.cc.utexas.edu>
To: cypherpunks@toad.com
Message Hash: 13d9971a575218245e9aa919715b8e134574ceb4aca1cf4235c6564ce163c154
Message ID: <199306021836.AA01020@flubber.cc.utexas.edu>
Reply To: <9306021056.AA05648@toad.com>
UTC Datetime: 1993-06-02 17:59:04 UTC
Raw Date: Wed, 2 Jun 93 10:59:04 PDT

Raw message

From: Jim McCoy <mccoy@ccwf.cc.utexas.edu>
Date: Wed, 2 Jun 93 10:59:04 PDT
To: cypherpunks@toad.com
Subject: Re: Crypto anarchy in a VW? (not the bug)
In-Reply-To: <9306021056.AA05648@toad.com>
Message-ID: <199306021836.AA01020@flubber.cc.utexas.edu>
MIME-Version: 1.0
Content-Type: text


Mark <mark@coombs.anu.edu.au> writes:
> 
> >I still don't see why all of the actual encryption couldn't be done in 
> >software though...
> 
> Me either, apart from TEMPEST issues...

Speed.  No software implementation will be able to match a hardware DES
chip in total throughput.  I have enough trouble dealing with the drive
transfer speeds imposed upon PC unix systems with the lame bus, but even
this could keep up if I had to run my file access through a software DES
system.  There are cards out there that can do this, and it doesn't really
make sense not to offload this to an external device.

> Linux comes with slot in file system
> modules (as detailed in a letter to Jim) that you can easily adapt to your
> own uses. Ive been playing around with this idea for a while. Adding a
> desfs(tm) (me :) to a linux kernel is not going to be that hard I think..
> (touch wood).

Yes, the other thing that pushed me to linux (besides the larger user
community) was the support for "drop-in" filesystems.  

jim




Thread