Re: Principles of Spam-abatement

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sat, 28 Feb 2004, Tom Petch wrote:

> <extract>
> *** Anonymous Bulk Email Software
> 
> *** is a super fast bulk email software that sends out at speeds greater
> than 1,000,000 emails per hour* on a dedicated mailing server. *** has the
> capability to use Proxies and Relays and also to send directly.
> 
> Some of the features include:
> Anonymous Mailing using Proxies
> Message Randomization to bypass Spam Filters
> Speeds over 850-950K emails per hour on Turbo Mode
> Up to 1000 Threads
> Unlimited Email List Size (up to 100 Million per file)
> HTML and Plain Text Emails
> Tag Macros to personalize and randomize emails
> Custom Headers ....... more on
> </extract>
> 
> Something along the lines of 'Know your enemy' comes to mind; get hold of
> such a product, reverse engineer it, find its weaknesses and nullify it.  I
> am thinking that spam is and will remain a long-term battleground and it
> needs serious effort to counter, perhaps a Cert-like organisation, and we
> are just not putting in enough serious effort yet; perhaps the cost to us
> is not yet high enough to stir us to action.

I keep swearing that I'm not going to respond yet again, but I keep
getting drawn into it as people miss a key point... such as the real
point of this "extract".

It has been pointed out several times that this sort of battle is one
that you cannot "win" on the grounds of technology alone.  At the same
time you are reverse engineering their spam engine, they are reverse
engineering your reverse-engineered anti-spam engine, and they have the
distinct advantage that your anti-spam engine is quite likely to be open
source while their spam engine is quite likely to be completely closed
source.  Every move you make to block them, they make a move to counter
your block, and they can move faster.  Look at the "features" list above
-- what is it if not "moves" against blacklists (using proxies), keyword
and bayesian spam filters (message randomization, tag macros),
slow/blocking MTA's (lots of threads), browser and text based message
review by the user (if they get in with a plausible tag-macro-generated
subject line, the user HAS to see the actual message in order to do end
stage rejection of material that makes it through their local filter, if
any).

Move-countermove.  This is not a technology problem, it is a war, an
exercise in practical biology.  There are attackers and defenders and it
is pointless to erect some expensive Maginot Line and pretend it will
solve the problem.

Adding encryption or signature simply causes spamware vendors to add an
encryption and signature module to their code and mailing list database
and address-grazing webbots.  This in turn makes the people who sell
spamware still more money selling their "new improved version",
cheerfully payed by the spammer who continues to make all that lovely
money and can easily afford the latest version.  Add solving a "puzzle"
and you might cut down on the peak throughput -- until they add puzzle
solvers in a backing cluster, and of course that adds tremendous
MATCHING expense to every legitimate MTA on the planet.  Add a delay,
and they add more threads where the MTA can wait out your delay but
maintain net throughput in parallel connections up to the limits of the
bandwidth of their MTA POP bottleneck (likely to be much lower than the
capacity of their spamware in any event).  In all cases they continue
saturating their connection with outgoing messages, which is all they
can afford to do anyway.

SPAM is not a static problem, it is a dynamic problem, being developed
and driven by Evil systems and software engineers every bit as talented
and dedicated as their Good opponents.  As long as spammers make money,
they will find ways around and through any mere "algorithmic" defense,
because the algorithmic defense has an unpassable boundary where false
positives become an unacceptable barrier (whether or not they are
rejected at the MTA or further downstream at the MUA), and plenty of
email traffic related to legitimate commerce has a finite chance of
being a false positive by an overaggressive filter.

There is one and only one way to "stop spam" (as opposed to learning to
live with it so that it doesn't bother you -- much -- as Vernon and I
and many others do already). Change the fundamental rules of the game.
When spamming is openly illegal and/or spammers stop making money (on
average), spam will stop.  Until then, history clearly shows that as
long as there is a buck to be made, there will be those trying to make
the buck, and they will route around every obstacle you are willing to
put in their path because they can automate their attack and can scale
expenses and find products that make them money even with a 0.001%
response rate, while you have to defend one system and user at a time,
some of whom BUY the products the spammers are selling.  We are NOT
going to stop spam with protocol, software, technology, as long as it
legitimately makes money.  We won't even slow it down.  Spam as a
problem is still actively growing, in SPITE of ever-more sophisticated
defenses, driven by all that MONEY spammers are making.

There are several ways that we as a society might try to stop them from
making money.

One is for everybody on the planet to refuse to buy anything sold via
SPAM.  Hmmm, not too likely that THIS is gonna happen, right?  Lots of
folks complain like hell about SPAM 90% of the time, but when they see
the RIGHT piece of SPAM, the one that is selling something they actually
want, they buy it.  One person's SPAM is another's golden opportunity to
enlarge this or that with safe herbal products or special exercises.  In
a large enough universe of people, somebody buys some of almost anything
sold, and spam-sellers make money because spam is so cheap to send.

One is add a direct and unreasonably nonlinear "expense" to sending
SPAM.  Lots of schemes have been proposed here that are either totally
ineffective if one actually does the arithmetic or punish the innocent
as much or more than the spammers -- adding "computational costs" per
message, inserting delays of any sort, deliberately making expensive
mail servers LESS efficient so we have to buy MORE of them to acheive
the same degree of service (hmm, great idea that, sure to be very
popular).  Adding an up front "fee" for sending mail is certain to be
similarly infinitely popular with the millions of users who are tired of
living with the ever escalating price of paper mail. Also, look how
effective real mail costs are at stopping PAPER spam.  Direct mail
advertising costs roughly $1 per piece (total cost) to send, yet the
ratio of direct mail advertising to real mail in my mailbox remains 2 or
3 to 1, easily.  A lousy 1% response rate makes the advertisers money,
in most cases, for the kinds of products that are sold this way.  As
long as they make money, they will continue to fill my mailbox on the
odd chance that I might eventually be part of that 1% and buy something
they sell.

One scheme alone puts additional expenses "only" on spammers and not the
innocent (or rather, pays the cost out of tax revenues distributed
relatively painlessly across the entire population).  Pass laws
prohibiting spam and fine the hell out of spammers.  Fine the hell out
of ISPs that are a point of origin for egregious spam -- get them to
police their own network.  Enforce the acceptable use agreements upon
which the internet backbone is already based, again with real economic
consequences -- disconnection of the ISP's network and all its clients
from the backbone, for example -- unless and until they maintain a
spam-to-legitimate traffic ratio less than 10%, 5%, 1% averaged over any
month.  It's not like spammers don't have an absolutely obvious network
traffic signature -- who ELSE sends out "broadcasts" of thousands of
nearly identical TCP port 25 messages per hour, up to order a million
per day, from a generally unregistered address?  It's just that the ISP
makes a big chunk of MONEY from that high-bandwidth-purchasing client
and isn't about to say oh YOU there, could you please stop spamming and
paying us for all that bandwidth you are using while doing so?

Direct email advertising being a form of interstate commerce, get the
federal government to regulate it (as is their constitutional privilege
and duty) by requiring interstate or international spammers to be
"certified" as compliant with new, strict rules (including the setting
of the IETF "evil bit" on all spam traffic;-).  Get states to match the
law for in-state traffic.  Pass a right to privacy law making it illegal
to sell lists of email addresses or anonymously collect them for use or
for sale.  Pass a privacy law permitting only "opt-in" collection of
email addresses by corporations or other entities for their own,
strictly limited, use.

Or my favorite:  TAX spam.  If you want to but the fear of God into
spammers (or anybody:-), sic the IRS on them.  As far as I can tell, one
can tax nearly anything, certainly any activity associated with
commerce.  Impose a 1 cent/message tax on SPAM (to be increased as
necessary to keep spam unprofitable), and then do "lazy enforcement"
starting with the biggest spammers and working your way down.  Make the
ISPs liable for delinquent taxes (they were sent on their lines, after
all).

Where mere law might fail to daunt spammers armed with lawyers defending
their right to free speech and a "lack of seizable assets", the IRS has
a reputation of being unrelenting and impossible to shake once they are
on your tail, and don't let a little thing like no assets bother them at
all -- they'll cheerfully settle for a chunk of any of your FUTURE
assets and garnishee your income until 2104 if that's what it takes for
you to pay them off.  They have a whole different standard of proof, too
-- much closer to spammers having to prove their INNOCENCE than the IRS
having to prove they are guilty.  Armed with a court order issued by a
real judge and based on consumer complaints, they can easily snoop that
megamessage-per-day traffic stream while working out what the spammer
owes.  Then they can just seize the spammer's house, dog, computer, car,
and give them the choice between jail and bankruptcy and a lifetime of
indentured servitude.

On the other hand, perhaps this is too cruel even for spammers...;-)

Perhaps this wouldn't work.  Who knows.  Control by law has worked
unbelievably well for phone-spam, though -- I get vanishingly close to
ZERO phone-spam calls since I put my number on the national DNC list,
and I will personally phone my government representatives every day
(right around dinner) to let them know just how I feel if they do not
defend it tooth and nail against the perfectly understandable desperate
attempt of the now-dying phone-spam industry to have the law thrown out.

Seems worth it to give it a try.

   rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb@xxxxxxxxxxxx





[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]