A while ago I was musing on spam, and the increase in volume I've seen recently.
Well, I admit it. I think I'm part of the problem. But, then, I think that everyone running silent anti-spam software is too. While on the one hand SpamAssassin keeps spam out of my mailbox, running it doesn't keep that spam out of anyone else's mailbox, and neither does it do anything to keep the spam off my machine in the first place. IP blocking, though, does do that.
I think it's time to build a widget to help automate the generation of the "you suck, get lost" IP list from the spam that SpamAssassin identifies.
Re:Different tactic
BooK on 2002-12-18T10:14:26
The problem with blacklists is "how not to block valid emails?"...
grinder had a very good idea to hit spammers where it hurts. Quote:
And what I'm toying with is something a little more radical. Sure I'll probably sign up with an RBL or two, but what I'm going to do is I'm still going to accept everything. But. When I come acrosss mail that comes from a suspected spammer, I'm going to put sleep(120) or so between each step in the transfer dialog.If enough people start doing this (and maybe people do; I haven't looked into the matter) then spammers are going to be hit where it hurts: raw throughput. Instead of pumping out 100 000 messages per hour, they're gonna start choking on sites which take minutes to complete.If you are an innocent site, well, tough, from time to time you'll have a socket tied up on your mailer to a remote site, but as sending email isn't the be-all and end-all of your business, that's not going to be a big problem. But if you're a spammer, you're going to find that you can't bomb as many people in the same time frame as you used to. Spamming therefore becomes more uneconomical.And that, in my books, is a Good Thing.Re:Different tactic
ziggy on 2002-12-18T14:37:48
I think Dan is making a different point. We're not significantly increasing the cost for the spammer. The cost to send spam is not directly related to the volume of spam sent. Therefore, savvy users are reducing the effectiveness of current spamming techniques, so the next step in the arms race is to increase the amount of spam sent, so that enough of it gets through to make it worthwhile for the spammer.Now we're trying to remove the economy of spam, by making sure the user never sees the spam. I think hitting the Spammers in the wallet is the only place they're going to take any notice.Re:Different tactic
Matts on 2002-12-18T15:12:19
It's not the cost of sending spam I'm talking about. It's the return cost of spam. What the spammer's client gets out of spamming. If we can reduce that to zero (or as close to zero as possible) then the spammers exist no longer.
Of course what we really need is a secure email protocol that validates the sender. Unfortunately SMTP is just too pandemic now to get that to happen.Re:Different tactic
ziggy on 2002-12-18T15:33:15
Yes, and what I think Dan has noticed is that when a few savvy users use spam-blockers, the return-on-spam cost decreases, but doesn't become zero (or near-zero). Therefore, the next result is to send more spam, since the incremental cost of sending out a few hundred or a few thousand more messages is near zero.It's not the cost of sending spam I'm talking about. It's the return cost of spam. What the spammer's client gets out of spamming. If we can reduce that to zero (or as close to zero as possible) then the spammers exist no longer.I think that the idea behind spam blocking is sound, but it's predicated on virtually everyone using spam blockers. However, if anything less than a vast majority of users use spam blockers, then the paradoxical near-term effect is to increase spam.
Re:Different tactic
Matts on 2002-12-18T17:24:42
I disagree.
It takes expensive hardware to send a lot of spam. Diminishing returns means that they're going to either have to spend more on hardware and bandwidth to get out more spam, or they're going to go bust.
Besides, I'm not sure I want to eliminate spam - I'd be out of a job;-) Oh the internal conflicts! Re:Different tactic
Elian on 2002-12-18T15:30:56
I know complaining and bouncing haven't worked--the first are ignored, the second spoofed to bother some poor sucker somewhere else in the world. But making spam silently disappear on the end-user's machine doesn't help much either. The spammers don't give a damn if some of their spam gets filtered, as they expect almost all of it to be ignored anyway.The problem with the anti-spam software is it's actually taken away some of the disincentives that mail admins used to use. When AOL blacklists your IP address, none of your spam gets through and you notice that. When 90% of AOL users kick in anti-spam filters that means that 10% will not, and who cares that 90% of the work is wasted? These guys brag about million-message an hour delivery capabilities. Each one's so cheap that the silently tossed ones are irrelevant, especially as these weasels are looking at 0.0015% response rates anyway. Mail caught by spam filters is lost in the noise of mail deleted as spam anyway.
Do I like and use anti-spam software? Heck yes. But using it has, ultimately, increased the spam load on my server rather than decreased it, which is why I want a bigger solution.
Re:Different tactic
Matts on 2002-12-18T15:53:54
What I recommend is modifying qpsmtpd's spamassassin capabilities to reject rather than filter if you see a high scoring spam.
I don't recommend this to everyone, as some people have mail that SpamAssassin scores rather highly, but for regular geeks it's probably a good idea.
I do feel sympathy though. I know exactly what you mean. But I don't really have a solution yet.:-/
