1997-02-08 - Re: anonymous remailers

Header Data

From: “Mark M.” <markm@voicenet.com>
To: cypherpunks@toad.com
Message Hash: 405e6e011d5565b26a587494cdeadffcdfb8aa437e15dac6b7420e68ba98c02c
Message ID: <Pine.LNX.3.95.970208120030.540A-100000@purple.voicenet.com>
Reply To: <199702080810.AAA13064@toad.com>
UTC Datetime: 1997-02-08 17:06:39 UTC
Raw Date: Sat, 8 Feb 1997 09:06:39 -0800 (PST)

Raw message

From: "Mark M." <markm@voicenet.com>
Date: Sat, 8 Feb 1997 09:06:39 -0800 (PST)
To: cypherpunks@toad.com
Subject: Re: anonymous remailers
In-Reply-To: <199702080810.AAA13064@toad.com>
Message-ID: <Pine.LNX.3.95.970208120030.540A-100000@purple.voicenet.com>
MIME-Version: 1.0
Content-Type: text/plain


-----BEGIN PGP SIGNED MESSAGE-----

On Sat, 8 Feb 1997, Jeremiah A Blatz wrote:

> Ummm, if you run your own remailer, and don't get lots of people to
> use it, then traffic analysis will reveal that you are the sender
> quite quickly. It will pretty much make everything in the chain before
> your remiler useless. If you send your message through remailers a, b,
> c, and d like this:
> 
> you -> a -> b -> c -> d -> alt.drugs.and-other-various-horsemen
> 
> and only you use c, then your effective chain is:
> 
> someone who could only be you -> d -> alt.drugs.and-other-various-horsemen

This is assuming that it is a reasonable assumption that all traffic going
through remailer c originated from the owner.  If there is one non-corrupt
remailer in the chain before c, then this would not be a valid assumption
because traffic from the owner would be indistinguishable from traffic sent
by anyone else.  If the remailer has low traffic, the solution is, of course,
to make it higher traffic.  Chain a bunch of messages that get sent to
/dev/null through the remailers, being sure to include c somewhere in the
middle of the chain.  If Mixmaster is used, then it would be virtually
impossible to differentiate between "real" messages and messages destined for
/dev/null.  It would be a little easier with Type I since the size of the
ciphertext decreases after each hop.  This all assumes that encryption is being
used, of course.


Mark
-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQEVAwUBMvyzcCzIPc7jvyFpAQHRpQf+LusfAS8dhDczpYTHGjgIRo38gPHeDdVn
+qmmikRNravoEiPD9GIrZ4OeYKOs6zykvhWuMoTtsVi/a7p1HZyWzl5A5KkxofUv
nLOUoPriQ9Ps8fzc3B31G5nwj5d6Es7nnfZbGk1dV5KS5bN7fyu9umBeFiW7jNcj
eTf8GmFH7Rxi5aoUc0uMMR/YffMNl0fHo+wooPNnTBMppLouTIr9iQdCxDOJ7eJc
QAFyEXYWtRP8AqrnB0/pVAXUtrnui+Ev1waOkMYKbWuiQ8tkHbLAvcmpAVnD67jX
4f3ZQkhXG6C4VbYF3fTlL0ujZgRal0csG0X4u6x/5ID4Blle9hwtIQ==
=BqMW
-----END PGP SIGNATURE-----






Thread