At 08:32 17-10-2006, Dave Pooser wrote:
You think so? By my count, my server is transmitting roughly 80 bytes of
data (HELO, MAIL FROM:, RCPT TO: and QUIT); even with overhead from RBL
checks on your side that shouldn't contribute to any load. It's not like an
evil spammer could carefully synchronize it so that millions of mail servers
would all try to do callouts at exactly the same microsecond, after all.
It doesn't have to be at exactly the same microsecond. UCE and Virus
infected emails commonly forge the sender's address. It can be much
more than 80 bytes and connections when sender verification becomes widespread.
Is sender verification being used to solve the "Accept and Bounce"
problem? If so, it's better to reject at the SMTP phase.
Have you actually seen a server DOSed by sender callouts, ever? I never have
Yes.
Regards,
-sm
