1996-08-17 - ISPs in cyberspace - how?

Header Data

From: Vincent Cate <vince@offshore.com.ai>
To: Black Unicorn <unicorn@schloss.li>
Message Hash: b9d16323a4d3fdc33e8c4acfe7d55a00441408d58a1ded5387f9e983184fc39e
Message ID: <Pine.LNX.3.91.960816170922.18711A-100000@offshore>
Reply To: <Pine.SUN.3.94.960816111931.24799B-100000@polaris>
UTC Datetime: 1996-08-17 00:01:44 UTC
Raw Date: Sat, 17 Aug 1996 08:01:44 +0800

Raw message

From: Vincent Cate <vince@offshore.com.ai>
Date: Sat, 17 Aug 1996 08:01:44 +0800
To: Black Unicorn <unicorn@schloss.li>
Subject: ISPs in cyberspace - how?
In-Reply-To: <Pine.SUN.3.94.960816111931.24799B-100000@polaris>
Message-ID: <Pine.LNX.3.91.960816170922.18711A-100000@offshore>
MIME-Version: 1.0
Content-Type: text/plain



Black Unicorn:
> The trick is to get the ISP to exist in cyberspace, or ever will it be
> subject to the whim of the local authority.
> 
> This is a key and very important distinction.

I agree.   I think this would be a good focus for cypherpunks to think
about for awhile.  So let me try to start something.

A first step is just having a domain name (foo.com) that you can move to
different virtual hosts.  This is still vulnerable at either the Internic
or at your nameservers.  Could also be a subdomain of some provider (say
foo.c2.org), but then c2.org gets the pressure. So far I don't know of
reporters going after the Internic for allowing a name (like foo.com).  So
this could be safe for awhile.  But it is not totally secure. 

A better method would be to make our own cypherpunk top level domain (I
think Sameer talked about this some time back).  With this people would
either have to setup their nameserver to use one of the cypherpunk
nameservers or get the IP address from some other method (a web page,
ftpable file, newsgroup, searching AltaVista, or a web page with a cgi
script to do the lookup on a machine).  But with this there would be no
easy way to cut off a name, and the ISP could always relocate if their
physical location were cut off (i.e. the IP address had to change). 

Then it is just an issue of coordinating this top level domain.  Say a
public key for each new subdomain and updates are done by sending a pgp
signed message to each of the servers.  With this type of design you could
have hundreds or thousands of servers that each were updated
independently.  To shut it down would take shutting down a lot of
nameservers. 

What do people think?  

   --  Vince

-----------------------------------------------------------------------
Vincent Cate   vince@offshore.com.ai  http://www.offshore.com.ai/vince/
Offshore Information Services         http://www.offshore.com.ai/





Thread