On Sun, Jun 15, 2003 at 06:38:28PM +0200, Mathieu Roy wrote: > > > You are betting that most ISPs are careful about spam issues, what > > > we can easily call "good ISPs" (term I did not used) or "decent > > > ISPs". You are proposing to block/tag mails that come from some > > > ISPs, > > > > No, forget about ISPs. > Replace ISPs by IP, it does not make it better. > Even worse. I got a cable connection with an IP that change rarely, > several times in a month. I use my local SMTP, not my ISP one (don't > wanna be bothered with the ISP discontinuously working services). > If a user of my ISP run an misconfigured SMTP, as open-relay, a IP that > belong to my ISP will be blocked. > A month a ago, I may get myself this blocked IP (given by a > dhcpd)... I let you imagine the next step. Would the next step be to go to the RBL's website and ask for your IP to be re-checked so that it can be de-listed? No, probably not; that would be too easy, and wouldn't leave any room for self-righteous Nazism analogies -- and what fun would that be? Conscientious administrators (as I fully expect debian-admin would be) continuously re-evaluate the effectiveness of their chosen RBLs to decide whether the benefits outweigh the costs. I don't know anyone who would use an RBL that didn't provide clear means for getting an IP re-checked. So I don't see why assuming this RBL is broken is a useful exercise. > I hope, and guess, that DNSBL does not list IP so easily*. But the way > it works is questionable to me: > You do not block a mail because, by examining its content, it > appears to be obvious spam, but because you assume that the IP > it came from can only send spam. You assume that the only goal is to prevent the spam from reaching my inbox. By the time the spam reaches my local spamassassin rules, the spammer has already stolen my bandwidth, my ISP's bandwidth, processor power on my mail server, and perhaps processor power on mail servers in between (not to mention, the resources of the open relay itself). When spam accounts for 50% of the total mail volume on the Internet (this is not an invitation for statistics showing that one's personal spam volume is lower), this adds up -- and implementing content-based spam filters is the most resource-intensive part of all. People simultaneously complain about the high cost of broadband, and the low quality of their provider's mail service. But when it comes to the fact that running a reliable mail server today costs 2-3 times what it would without spam, these same users are unwilling to make the small sacrifices that would allow locking the spammers out of the SMTP-communicating world. > I do not know how spamming companies works exactly. If I were them, I > would not buy a static IP to send spam everyday but I would buy a > dynamic IP, with a dial-up connection (IP that change at each > connection) - blocking my IP efficiently would be blocking every IP of > my ISP. The only solution would be to write to abuse@ISP to get my > contract with the ISP over. > What could be done also if I had a static IP. For someone who doesn't understand what an open relay is, or how the DSBL works, you sure have a lot of opinions about both. > But I definitely find spamassassin conceptually much better - because > it really takes a mail for what it is. It cannot be trapped. Because > if the DNSBL one day become a major problem to spammers, who knows > what kind of methods they may use to attack them. Maybe by sending the emails from their own IPs? Then we know where they're really connected from. Then we sue them. Doesn't sound so bad to me. -- Steve Langasek postmodern programmer
Attachment:
pgpTjG5YzIL1f.pgp
Description: PGP signature