From: dlv@bwalk.dm.com (Dr.Dimitri Vulis KOTM)
Date: Sat, 29 Mar 1997 16:10:29 -0800 (PST)
To: cypherpunks@toad.com
Subject: Re: remailer spam throttle
In-Reply-To: <199703291126.LAA00386@server.test.net>
Message-ID: <88Lc5D5w165w@bwalk.dm.com>
MIME-Version: 1.0
Content-Type: text/plain


Adam Back <aba@dcs.ex.ac.uk> writes:
> Dimitri Vulis <dlv@bwalk.dm.com> writes:
> > Adam Back <aba@dcs.ex.ac.uk> writes:
> > > [encrypt the message and send the recipient the key, discard the
> > > key, when they request the message they supply the key]
> >
> > That's a good idea, but it'll take up a lot of disk space at the
> > machine running the remailer.  Right now, remailers that provide
> > latency don't keep an e-mail for more than about 12 hours. Once
> > you start keeping them around for a few days (a reasonable grace
> > period for a first-time user), it's a lot more disk space.
>
> "Disk is cheap."  I just bought a 3.8Gb quantum for $300.  That's 4
> times cheaper Gb/$ ratio than last disk I bought a year ago.

I paid over $700 for a 2GB disk about 18 mos ago. Yes, disk prices
are falling through floor - giving the lie to the scum that tries to
control Usenet "spam" - but is it really worth it to have to maintain
a huge spool space to support a remailer? If you really think you can
afford a few GB of extra disk storage, start up a Usenet site of virtue.

> 1Gb encrypted message pool?  No problem.
>
> Don't have the resources for the 1Gb disk?  Your ISP will be charging
> you many times that for your leased line.
>
> If it's an issue charge postage to cover your storage and transmission
> costs.
>
> However, I suppose if you're using an ISP shell account, and have no
> leased line, you may have more space restrictions.

It should be possible to run a remailer on an average $19.95/mo shell
account.  Most of them come with a disk quota of a couple of megabytes.
I do think that keeping a lot of mail in the queue (more than is kept
by the remailers that use latency now) would make such accounts
unsuitable for remailers.

> A better idea, (for both ISP shell account and leased line version)
> would to send the recipient the encrypted data, you keep the key :-)
>
> (View it as a secret split where the parts happen to be different
> size, you keep the small one).
>
> They send you the data and you decrypt, send it back and discard the
> key.  You could store 16 million keys on your 1Gb remailer spool
> partition.  That 16 million keys would represent 16 Gb of delivered
> encrypted email.
>
> You could still cope with a fair number of messages in 1Mb remailer
> key spool in your home file space on an ISP based shell account.
>
> I think this solves remailer resource objections.

yes, this sort of solves the storage problem.

Thinking how this might work...

Alice's remailer gets an e-mail for Bob.

Alice's remailer generates 2 pseudo-random numbers, K and L, and uses
K to encrypt the message with a symmetric cryptoscheme.

Alice's remailer sends the encrypted message and L to Bob with the following
note: if you want to receive the key to decrypt this message, send back L and
acknowledge the disclaimer.

Alice's remailer retains the triple (K,L,Bob). Because it's small, it can
be kept for a long time.

If Bob sends back L, the remailer sends Bob K so he can read his message.

I see a few problems with this. I'm sure it can be improved.

1. What if Bob is another remailer, unknown to Alice?

2. What if Bob doesn't have the program to decode his message? (It's
fair to assume that everyone can fine PGP. It's not fair to assume that
everyone can find e.g. triple DES.)

3. What if the LEA's decide that the collection of triples on Alice's
computer is worth looking at, for instance, for the list of addresses.
(OK, they could be encrypted probably.)

4. What if the LEA's decide to find out how K, L are generated?

> You still have the problem of saturation of your network bandwidth.
> You can probably encrypt (it's symmetric key, say IDEA) fast enough
> for encryption overhead to be less of a problem than the saturation of
> the bandwidth of your leased line.

I think IDEA or triple DES would be fast enough for this.

> Your other problem is user acceptance -- your average user may still
> object to receiving Mb's of junk which while the contents don't offend
> him (he can't read it), the sheer volume may annoy him.

Given how some people react very negatively to what they consider
"unsolicited e-mail", I think it might be a bad idea to send them
large encrypted messages w/o confirming that they want them. But
it's also a bad idea to keep an unknown's e-mail on the remailer.
The only way out, IMO, is to *discard* an unknown's e-mail and to
tell him what positive action he can take so future e-mails don't
get discarded.

And by the way: a remailer should probably keep a list of people
who have been told about the need for a positive action, so they
won't get duplicate notices more than every week or so. If X
sends Y 10 e-mails and the remailer sends Y 10 identical notifications
with the instructions how to enable receipt, Y might get justifiable
upset.

When Alice's remailer gets an e-mail for Bob, it should do
something like:


          try to fetch Bob's public key
          from a well-run key server
            /                    \
         success               failure
           |                      |
      encrypt the e-mail      discard the
      with bob's key and      e-mail!
      pass it to Bob              |
                              was the form
                              letter sent to
                              Bob in the last
                              7 days?    \
                             /            no
                           yes              \
                          /                exit
                    send bob the
                    form letter


Wow, isn't that a perry flow chart. :-)

...
> > Suppose a LEA wants to search the computer hosting the remailer.
> > They come across a bunch of encrypted files.
> > The operator has to convince the LEA that they don't have the means
> > to decrypt the e-mails or even to figure out who they're from.
>
> Well you can't can you?

It may be hard to prove a negative to a LEO who doesn't know what
the hell you're talking about. You have a file in your spool that
was encrypted with a key that your program generated, but now you
no longer have the key? Well, tell us how the key was generated.

...
> I agree with what you're saying that technological aptitude can
> provide some metric of cluefullness.  And that cluefullness usually
> correlates with good net citizenly behaviour (eg understanding what a
> remailer is, and not legally threatening, or setting the Feds on to
> the human owner of a bot.
>
> However there exist quite a few counter examples: cluefull people who
> do not use PGP, or consider it to be too much hassle (quite a lot of
> them hang out on cypherpunks, and will actually object if you send
> them encrypted email, even though spending most of their time on
> cpunks arguing about the usefulness and essentiality of these tools
> for privacy).

Well - if generating and publishing a PGP key were a requirement
to be able to receive anonymous e-mail, I'm sure they'd do it.
Some people (like Dale) legitimately question certain uses of PGP, but
I think in this case it's worth encouraging its use.

> Also some journalists are pretty clueful (or useful to be able to send
> information too, shall we say, journalists as a group having a bad
> rep), and not many can handle PGP.

Hmm - I'm sure John Markoff and even Declan can handle PGP.
If, for example, Jon Schwartz can't handle PGP (and I don't know
whether he can :-), then he doesn't deserve any anonymous tips.
Free market in action.

Later - thanks a lot for the extensive comments.

---

Dr.Dimitri Vulis KOTM
Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps