1996-05-20 - Re: crosspost re remailers

Header Data

From: Bill Stewart <stewarts@ix.netcom.com>
To: Rich Graves <llurch@networking.stanford.edu>
Message Hash: dc8efbcc185452779c6627070ddf8621803dab4832e75b210aae0d7000707520
Message ID: <199605200149.SAA10873@toad.com>
Reply To: N/A
UTC Datetime: 1996-05-20 07:08:16 UTC
Raw Date: Mon, 20 May 1996 15:08:16 +0800

Raw message

From: Bill Stewart <stewarts@ix.netcom.com>
Date: Mon, 20 May 1996 15:08:16 +0800
To: Rich Graves <llurch@networking.stanford.edu>
Subject: Re: crosspost re remailers
Message-ID: <199605200149.SAA10873@toad.com>
MIME-Version: 1.0
Content-Type: text/plain


At 01:00 AM 5/16/96 -0700, Rich Graves <llurch@networking.stanford.edu> wrote:
>The entry points into the system, though, have value. You need to be able
>to locate and trust them. Remailer reputations are valuable. Otherwise,
>you're liable to send your message into the NSA-remailers-are-us system.
>You need a web of trust among remailers at the very least, which means
>some level of exposure (at least by "social analysis" by observing the
>relationships among the various remailer nyms).

There are two ways around this problem - chaining remailers, and encryption.
As long as one remailer on the chain from your source to your destination
is not compromised, you can send mail encrypted to each remailer in the chain
(i.e. send Alice's Remailer the message E(Alice, E(Bob, E(Carol, message))))
and the message won't be compromised, though it may be blocked.

For reply-style remailers, chaining is much harder, at least if you use
a non-stealth encryption system like PGP or RIPEM which reveals the recipient's
public key in the headers, and if the remailers don't include a public key
to encrypt the message contents with as well as a key to encrypt the
connection to the next remailer with.  (Otherwise, if the sender
encrypts a message to NymA@Alice, and Alice encrypts it with Bob's key and
sends it to Bob, Bob decrypts it, encrypts with Carol's key, and sends.... 
Zed decrypts it, encrypts it with your key, and sends it to you,
then any node compromised by Bad Guys will see a message encyrpted to NymA.

>Chaos within the system is good. Moving remailers around could be good,
>provided that a service location infrastructure is established.
>Raph's list is a good start, but it needs to be more automatic and dynamic --
>which to me (perhaps wrongly) suggests formalization, which means points
>of failure. 

It would be much easier if there were a DNS hack that lets you connect to
dns.remailer.net:registry, which takes your IP address and serves it as
remailerN.remailer.net until you log off (using a short DNS expiration time).
This does provide a large number of interesting attacks on the system -
denial of service is easy (make lots of connections, filling all ports),
and it's easy to add subverted remailers (just connect!)  It also doesn't spoof
reverse lookups unless your system is able to handle multiple IP addresses
(which requires cooperation from your ISP's routers, unless remailer.net
is also running a packet laundry, which increases its targetness.)
Signed DNS responses would help some attacks, once that's standardized.

Does anyone have any ideas on how to do this correctly?

#					Thanks;  Bill
# Bill Stewart, stewarts@ix.netcom.com, +1-415-442-2215
# goodtimes signature virus innoculation







Thread