From: shamrock@netcom.com (Lucky Green)
Date: Thu, 4 Jan 1996 05:17:56 +0800
To: cypherpunks@toad.com
Subject: Re: Foiling Traffic Analysis
Message-ID: <v02120d08ad107b4e9236@[192.0.2.1]>
MIME-Version: 1.0
Content-Type: text/plain


At 7:02 1/3/96, Fred Cohen wrote:
>You seem to be missing an important point about foiling traffic
>analysis.  It is essentially the same problem as the covert channel
>problem and its solution has the same challenges - it consumes a great
>in the way of resources.  In order to eliminate traffic analysis, you
>essentially have to always use the full bandwidth available (although
>you can have pseudo-random burst behaviors).  This in turn means that
>instead of gaining the low cost resulting from sharing bandwidth, you
>end up having far more utilization and (depending on what portion of the
>world does this) increasing the price of the resource.  So it costs a
>lot more and uses a great deal of bandwidth.

You are correct. A network of encrypted links that allways move packets at
full bandwidth is the basis of Wei Dai's Pipenet. If anyone ever codes
this, I am willing to sponsor a node. Other nodes may be set up if some
payment mechanism using Ecash is integrated with the system.


-- Lucky Green <mailto:shamrock@netcom.com>
   PGP encrypted mail preferred.