Brenno
I looked into the greylisting option and I'm not sure that it's going to be feasible for us; my instinct tells me we'd encounter way too many false positives using this approach.
The system is supposed to have false positives, these messages then get delayed & sending servers retry (unknown to the end recipient). On subsequent redelivery attempts the mail gets delivered & the sender is added to the whitelist automatically.
You can manually create extensive whitelists of all known senders so that mail gets delivered immediately without being delayed by greylisting.
You will be stunned at the virtually zero spam & virus messages coming through (actually not coming through).
Over a short period of time after implementation, you would have most popular senders whitelisted, so there would be no further delays for those senders.
Spammers etc usually use servers that will not retry sending, so their original messages do not get delivered as the greylisting rejects them, which typically also include virus laden messages.
See
http://wiki.contribs.org/GreylistingThe main downside of greylisting, is that local recipients (users) often expect to immediately or instantly receive email messages within a minute after they are sent, but when new senders (who are not yet in whitelists) send an email message, there can be a few hours or even a day or two delay (depending on external sending mail servers).
So users have to tolerate a (usually unknown in most cases) delay the first time they receive a message from a sender, otherwise that sender can be immediately added to a whitelist to have instant delivery (eg for urgent situations).
For regular senders there is no delay.
As I said, uses need to tolerate the way greylisting works, & accept that trade off as a means of protecting themselves & the network against viruses & spam etc.
Greylisting does work very effectively.