Nothing in Postfix prevents you from developing stateful policies
where repated 'good' clients become longer-term allowlisted, and
repated 'bad' clients become longer-term denylisted, for some
subjective definitions of 'good', 'bad', 'long' and 'short'. In the
case of botnet spam, this will make little difference because those
bots rarely come back dsuring their limited lifetime.
I would not build that logic into Postfix, just provide the data
that policies can be based on.
'Big data' approaches can be worthwhile when there actually is
enough data. However, unless you are a service provider then there
just isn't enough to be menaningful. You'd have to pool your data
with others while avoiding privacy leaks.
Wietse
_______________________________________________
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org
