On Fri, Oct 17, 2008 at 08:25:40AM -0300, Jose Fragoso wrote: > Hi, > > Thanks for the tip on using submission, SSL or TLS > ports. That solves many of my problems. > > But I still think that dynmically allocated IPs should be > treated somehow differently by SPAMD greylisting process. > > My point is that if a remote SMTP server goes through the > greylisting process and ends up getting its IP address > whitelisted, that should not be inherited by the next > owner of that IP address. > > I know it may be difficult (if not impossible) to identify > whether an IP address is part of an address poll of some > DSL or cable provider (maybe there is a list kept somewhere > in the world of such ranges). > > I know for sure one these ranges here in Brazil. And I see > a hell of a lot of spam passing through SPAMD, just because > some of these IP addresses got whitelisted by an earlier > well-behaved temporary owner. > > So I would like to know if someone has come up with an > interesting idea on dealing this issue.
This, alas, I'd say violates the basic assumptions for greytrapping. A bit. Most of the times at least here in Sweden, you get the same IP address every time. It is only when the ISP has to it gives you a new. And shortening the invalidation time for greytrapped addresses would essentially force the SMTP clients to pass greytrapping every time. My ISP does not allow outgoing SMTP from clients so I have to use their mail hub. That is also a solution, especially when combined with SPF (Server Policy Framework), provided MTAs use it. But it is not a solution at your end, it is a would be global solution. Maybe pf OS fingerprinting can pinpoint XP and Vista machines and tarpit mail from them. Or are there serious SMTP clients on those platforms? > > Best regards, > > Jose > > -- > Be Yourself @ mail.com! > Choose From 200+ Email Addresses > Get a Free Account at www.mail.com -- / Raimo Niskanen, Erlang/OTP, Ericsson AB
