From: "Douglas R. Floyd" <dfloyd@io.com>
Date: Fri, 16 Aug 1996 06:59:49 +0800
To: schmidt@pin.de (Stephan Schmidt)
Subject: Re: Jurisdictionless Distributed Data Havens
In-Reply-To: <Pine.LNX.3.94.960815140139.27838A-100000@blau.pin.de>
Message-ID: <199608151830.NAA10012@xanadu.io.com>
MIME-Version: 1.0
Content-Type: text/plain


> 
> 
> Some annotations:
> 
> 1. The easiest way to make such a distributeted data havens (DDH)
> would be to use a distributed unix file system that
> doesn't distribute files but chunks of bytes. (Striping)
> 

Like AFS?

> Advantage: 
> - All normal services would work: ftp,http,...
> - Copying, deleting and modifing files.
> - Easy to install and use.
> 
> Problems:
> Because each side can supply all data (collecting on
> the fly from other DDHs), the site holder
> could be responsible for the data.
> This could be prevented by collecting and
> assembling data at the client side (e.g. using
> JAVA).

How many people trust clients?  If we wanted clients, we would use WebStor
from Mcaffee.  I don't trust any clients specific to one task, and would
rather use generic E-mail/ftp/www.

> 
> 2. When the DDHs are distributed around the world in
> a lot of different states, it could be very
> difficult for any government to get any evidence for
> "illegal" data on one site.

In some countries, when the government jails someone, it doesn't matter
what evidence they have.  Not everyone has a justice system that at least
makes an attempt to give a fair trial.