In case it's interesting, here are some server stats that show what changed when we reconfigured our mail server to reject, rather than bounce mail that's addressed to non-existent addresses. Oct. 24th was the first full day.
Per-Day Traffic Summary
date received delivered deferred bounced rejected
--------------------------------------------------------------------
Oct 19 2003 35685 83174 0 57609 5933
Oct 20 2003 42200 86496 0 52786 6271
Oct 21 2003 44088 81862 0 50929 7271
Oct 22 2003 62220 101111 1 63215 10339
Oct 23 2003 58282 79700 7 45251 34140
Oct 24 2003 31205 13760 1 561 74616
Oct 25 2003 40750 10150 1 79 125762If you're wondering why the numbers don't seem to add up, it's mostly because one message can be addressed to multiple people, often when it's spam (or it's to a mailing list!). Thus, delivered, bounced and rejected totals can be higher than the number received. Also, every incoming message is processed twice -- once to scan for spam and viruses, a second time to deliver it. Thus, rejecting junk reduces our processing load tremendously. Yesterday, we processed about a quarter of a million fewer messages than we would have with the previous configuration! These numbers astound me... and I realize what an enormous resource hog spam creates for the major ISPs and such. I think we've seen a big increase in spamming just lately.
The domains that our top mail senders use are the big broadband consumer ISPs -- comcast.net, rr.com, attbi.com. Makes me realize why some people refuse mail from addresses in their dhcp ranges. We don't, since I believe that anyone should be able to operate a server at home.
In terms of sending messages, brin-l is at the top of our list. That doesn't surprise me; our work has to do with incoming mail, not outgoing, and brin-l is the most popular list we host.
Nick
-- Nick Arnett Phone/fax: (408) 904-7198 [EMAIL PROTECTED]
_______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
