On 2013-01-28 at 15:25 -0500, Lawrence K. Chen, P.Eng. wrote:
> But, we're doing a poor job on educating our users to not respond to
> phishing and having their accounts compromised or that its not funny
> to mark emails forwarded from their accounts in hotmail/gmail/etc. as
> spam just because they want it to go away. Also to not mark forwarded
> spam as spam.... Just as Merit disabled the spam button in Zimbra for
> us. A department mailing list was going into spam, I later heard that
> a bunch of people in the department joking that they were marking the
> messages as spam. This was also causing problem for Athletics...who
> flood the students with emails quite regularly. But, student tickets
> are electronic only....and those were getting lost in other student
> spam folders.
Get the mailing-lists into a different email domain from the
transactional mails for things like tickets. You don't need different
machines or IPs. Set up DKIM for each, so that the big email providers
can reliably associate the domain for reputation purposes.
Net result: it's fairly likely that those who send out spam ("hey, we're
important") will still be filtered, but their bad actions are less
likely to cause problems for others, such as those needing tickets.
tickets.ksu.edu would have one DKIM key.
You'd try to get the lists into lists.ksu.edu (see how lopsa.org does
this too) but you would then alias [email protected] to
[email protected] inside the mail-system and tell the MLM that
the latter domain is canonical, so that the mails go out with the new
domain. As long as you keep List-Id: the same, and any other markups,
most people won't have to do anything to accommodate the change, and you
can warn people in advance of the migration for others who have filters.
These days, you really want to try to isolate mail flows by
characteristics of the mail; "human people, direct from them" vs
"mailing-lists" vs "transactional mail in response to something
initiated by the user".
Reputation is key. You *really* should read "Sender Reputation in a
Large Webmail Service" by Bradley Taylor, presented at CEAS 2006. Hrm,
for a moment I thought that the paper was not linked to from:
http://research.google.com/pubs/author70.html
but it's just yet another of those Google+ unlabelled icons where you're
supposed to infer the meaning of that symbol in relation to the context:
the globe symbol to the right links to the PDF:
http://www.ceas.cc/2006/19.pdf
-Phil
_______________________________________________
Discuss mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
http://lopsa.org/
