OFF: anon.penet.fi

[M Holmes]

Paul Mather writes:

> Actually, the anonymous remailer network does take into account traffic
> analysis.  There is a network of remailers, and a message entering the
> system will get randomly circulated through the network before being
> delivered to its intended recipient.  Additionally, there can be a
> random delay at each remailer stage, so that the next message out of the
> remailer is not necessarily the one that just went in.

Excellent.

It occurred to me that an improvement which would prevent the
scientology problem would be an, err, anonymous anonymous remailer.

I haven't thought through protocols but what I had in mind was something
working on the Web as if it were some kind of WORMOS system. Basically
the remailer program would wander the web randomly looking for somewhere
to execute itself. It would collect encrypted messages sent to it and
send them out with as much blinding for traffic analysis as possible,
then relocate again. It would also accept encrypted commands, as
necessary from its owner, perhaps by the dead letter drop method
previously discussed.

The advatage would be that it wouldn't be clear where the program was or
in what legal jurisdiction and who was in fact controlling it. This
would make it more difficult for anyone to take injunctions to obtain
disclosure.

> Paul.

FoFP