Hi Cameron, Am Fr, den 08.10.2004 schrieb cls@truffula.sj.ca.us um 18:07: > [This message has also been posted.] > In article <2N0u0-7Xz-1@gated-at.bofh.it>, Kilian Krause wrote: > > > > --=-+rYNsJkiW3Vja8Xh+ktl > > Content-Type: text/plain > > Content-Transfer-Encoding: quoted-printable > > Please don't do that. Quoted-printable is broken. how come you quote <2N0u0-7Xz-1@gated-at.bofh.it>, but my mail was "<[🔎] 1097231985.18184.19.camel@ganymede>" as can be seen at: http://lists.debian.org/debian-isp/2004/10/msg00034.html ? > > > > Hi, > > > > I just came across a quite interesting idea.=20 > > Greylisting naturally makes only sense if the remote end is no MTA, but > > a direct TCP connection. All MTAs will come back and thus greylisting > > will only introduce latency which most admins will see as neccessary > > tradeoff for reduced spam. Yet, why not use the knowledge of DNSBL to > > tell you which are the dialup-ranges of ISPs? > > I'd like to know which DNSBL knows where "the dialup-ranges of ISPs" > are. It seems to me the biggest and worst ISPs withhold that > information as some kind of trade secret. I've been collecting the > ranges at SBC and the rest, one at a time, for years, as we receive > spam from them. Never found any DNSBL that tagged more than > about 20% of them. Admins at Verizon and Charter have each told me > they don't even have a reliable list of their own dynamically > assigned IP ranges for internal use. Well, nobody said the coverage is 100%, but are you sure it's that bad? As I said, this was the *IDEA* of taking advantage of DNSBL for something it wasn't meant to be originally. Using DNSBL for denying *ALL* access and thus forcing legitimate users to use an ISP MX they don't like, doesn't very well fit my mailer policy. But having a greylisting trigger on those hosts and thus stopping them in case it's not a legitimate MTA sounds like a very good option to me. Of course this doesn't replace the sa-exim, dspam or crm114, but it should help lower the load. Nobody said this is the magic weapon everybody uses and nobody found yet, it was just me wondering if these two pieces can't be matched together for the final goal. (As I said above, I'm against massive DNSBL use on the hardcore scale, so why not try with greylisting?) So do you still reckon that DNSBL are too far out for using them as a greylisting success probability predictor? -- Best regards, Kilian
Attachment:
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil