begin  quoting m ike as of Wed, May 18, 2005 at 07:59:24PM -0700:
[attribution lost]
> > Depends on how often people change their email address, or how
> > often you need to add/remove people from your list.  A lot of
> > people have a pretty static set of friends and relatives that
> > they can communicate via other means.  It's the geeks who end
> > up having relationships with people they never meet...
> >
> as a scientist, emails from strangers are important to me.

So a whitelist solution is bad. And hundreds of fake emails that someone
may associate with you is bad, too.
 
> > A few "don't send email to this address I mean it" can be caught by
> > the harvesters without (hopefully) confusing grandma or a long-lost
> > friend.
>
> I cannot tell if you are arguing that it is impossible to pollute a 
> spammer's address book by any significant factor.  That may 
> be the case.  But it would turn the tables, with spammers forever 
> trying to clean up their address lists.

Gabe gave a great response.  But I'll answer anyway...

Why would spammers try to clean up their address lists?

Sending email to a bogus address doesn't cost 'em anything. Since a vast
majority of 'em are forgers, bounces don't affect 'em either.

> > I can flood your system with spam to thousands of emails to fake email
> > addresses in the hopes of hitting your real email address... if I use
> > a zombie net, I can cripple your machine (or mail server) without
> > really trying. In fact, the more fake addresses you have, the harder
> > your machine would be hit.
> >
> in the end, how would the spammer benefit from this? 

Well, a 0.01% response rate is a success.

If you're not a potential customer, and you deploy hundreds of bogus
email addresses that result in your machine being crippled, it's no
skin off the spammer's nose, and as an anti-spammer, it's basically a
bonus.

> > Got an estimate as to how big that is, and how long it would take to
> > do the matching?
>
> I think it is very very feasible to match M = 500 emails per day against 
> N * M * D, where N is the number of fictitious addresses one owns,
> and D is depth, in days, of the cache.

Times U, for users.

So a successful medium ISP with, oh, 1000 users gets to archive

   100 * 500 * 30 * 1000 = 1500000000 messages.

Let's say each message is 1k each. Then we're dealing with 1,500,000,000
kilobytes, or 1,464,843 megabytes, or 1,430 gigabytes, which is roughly
a terabyte.

[snip]
> > I don't recall if RBLs expire IPs after a period of time (compromised
>
> As you've probably gather, I'm not to hip on these things. But I thought
> that anything in an email could be spoofed.  That is, what prevents a
> spammer from making it look like the spam came from my IP, which is
> valid, of course :) :) with the result that my IP gets added to the RBL.

More likely that you're running a zombie box and that gets added to 
an RBL.  But that's okay, that's what the RBL is *for*.

> > "There ought to be just one" is generally a sign that someone is trying
> > to exercise control -- and rarely is that exercised in your favor.
>
> Then you must love the de-centralized aspect of the tactic that i 
> proposed?  To implement it requires one person, the user. It runs 
> locally. Though i guess the code would be somewhat centralized.

It's fatal flaw is that it doesn't actually deter spammers.

-Stewart "Although honeypot accounts are appealing." Stremler

Attachment: pgpK0P5GfUhOA.pgp
Description: PGP signature

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to