1997-12-14 - profit from spammers (Re: hashcash spam prevention & firewalls)

Header Data

From: Adam Back <aba@dcs.ex.ac.uk>
To: rfarmer@HiWAAY.net
Message Hash: cae7159921cc3df17c38fef72b3ec1a984190f081d848a3676e3e056a846be04
Message ID: <199712132332.XAA03725@server.eternity.org>
Reply To: <Pine.OSF.3.96.971213120112.17974C-100000@fly.HiWAAY.net>
UTC Datetime: 1997-12-14 00:32:37 UTC
Raw Date: Sun, 14 Dec 1997 08:32:37 +0800

Raw message

From: Adam Back <aba@dcs.ex.ac.uk>
Date: Sun, 14 Dec 1997 08:32:37 +0800
To: rfarmer@HiWAAY.net
Subject: profit from spammers (Re: hashcash spam prevention & firewalls)
In-Reply-To: <Pine.OSF.3.96.971213120112.17974C-100000@fly.HiWAAY.net>
Message-ID: <199712132332.XAA03725@server.eternity.org>
MIME-Version: 1.0
Content-Type: text/plain




Randall Farmer masquerading as Joe <rfarmer@HiWAAY.net> writes:
> > What's the difference between this and simply keeping track of how many
> > messages each user sends in a 24 hour period and blocking people who are
> > obviously spamming?
> 
> Spammers don't have to appear to be the same person every time, nor
> do their messages have to be identical or even similar, so blocking
> of that kind can be worked around fairly easily -- at least more
> easily than hashcash.

I think Phelix was probably viewing it from the point of view if "what
if all ISPs turned off non subscriber mail forwarding at their SMTP
hubs"?

Well that would work clearly enough.  However I think there are a
number of technical problems to acheiving this.  Firstly SMTP does not
include authentication.

Secondly doing so reduces flexibility and many users use multiple
ISPs, forward their mail to various places, etc., and this kind of
stuff gets in the way.

Really we I think should be discouraging the control freak "positive
identification to use this port" syndrome such as the ident fuck up,
(Ident is a dumb method of identifying who is on the other end of a
socket on a unix box.  It opens up a socket to the ident port on the
originating machine, and asks who is on socket x port y?  The machine
can determine this from local OS tables, and sends the info back.
This works if the user does not have root on the machine.  Fortunately
this snoopy bastardised protocol is not doing so well these days
because there are more and more people who have root on their own
machines, and because there are so many windows machines which don't
know what a protected port number is.  This is good because Ident
sucks.) and moving instead towards, "who cares you are so long as you
can't overload my machine".  Moving longer term towards "who cares who
you are so long as I profit from your connecting to this port" aka
charge for port access with payer and payee anonymous ecash.  Then
everyone is welcome to use anywhere as a mail forwarding service --
spam becomes welcomed, and encouraged, too much custom and too low
bandwidth discouraging customers, the ISP will use the profits to
purchase a few more T3s.

Hashcash is the interim solution.

In the interim it is a fact of life that there are many many open SMTP
forwarding hubs.  The lack of software to configure them otherwise,
and inertia will ensure things remaing this way for some time.

Hashcash cuts out spam to your site (if you are an ISP), or to your
mailbox (if you are a user) even in an environment of an almost
unlimited supply of open SMTP hubs, and disposable ISP accounts,
because it puts the onus on the sender to consume more resources than
you.

Authenticated the hell out of your ISP's SMTP as a forwarding hub
won't prevent your users getting spammed to death, nor will it reduce
the overall spam problem much because the spammers will just use one
of the other open SMTP servers.

Adam
-- 
Now officially an EAR violation...
Have *you* exported RSA today? --> http://www.dcs.ex.ac.uk/~aba/rsa/

print pack"C*",split/\D+/,`echo "16iII*o\U@{$/=$z;[(pop,pop,unpack"H*",<>
)]}\EsMsKsN0[lN*1lK[d2%Sa2/d0<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<J]dsJxp"|dc`






Thread