ASSP outgoing spam protection

 


If you ENABLE the outgoing spam protection ASSP deluxe will apply automatically following ASSP settings ;

ASSP 1.x
URIBLLocal
  ON, enable URIBL checks for local senders
ScanLocal ON, enable clamAV scan of local senders
   vsValencePB (clamAV local scan scoring) will be increased to guarantee a block
blackReLocal ON, enable blackRe checks for local senders
BayesLocal ON , enable Bayesian checks for local senders
   bayslocalValencePB (Bayesian score for local email) will be increased to guarantee a block
ASSP 2.x
Same as ASSP 1.x plus HMM check for local

Each local sender which will fail outgoing checks will be blocked with an error message. The block will be logged in ASSP log too .

ASSP 1.x
If you want allow a good sender to pass the Bayesian local check you should add the local domain or email sender in
ASSP web interface > Bayesian menu > Skip Bayesian for this local senders* (noBayesian_local)

ASSP 2.x
If you want allow a good sender to pass the Bayesian/HMM local check you should add the local domain or email sender in
ASSP web interface > Bayesian/HMM menu >
Skip Bayesian and HMM Check for this local senders* (noBayesian_local)
 

ASSP 1.x
Before enabling the "outgoing spam protection" however there are negative points to be considered ;

various filter which can apply for remote senders, cannot be applied for local sender  (i.e. mx, ptr, RBL check, greylisting and similar checks)

only following checks can be applied

-> a bayesian check (can block the email)
-> a blackRe check (only a score which should not block the email immediately)
-> a ScanLocal check , the email will be checked and blocked if contains a virus signature
-> an URIBL check  (only a score which should not block the email immediately)

So since a good/complete "scoring" strategy with multiple filters such as for incoming email could not be applied to outgoing email (only bayesian,blackre,scanlocal and URIBL) , and since the bayesian filter could cause false positives (blocking rarely good senders which send email), I recommend you to enable outgoing spam protection only if strict required .
Or you may activate the "outgoing spam protection" and each time a good sender reports difficult to send email you can put his @domain in ASSP web interface > Bayesian > Skip Bayesian for this local senders* (noBayesian_local) . Or you may activate the "outgoing spam protection" , only for some suspected user using ASSP web interface > bayesian > Do Bayesian for this local senders only* (yesBayesian_local)

ASSP 2.x
Some consideration as above however you may have much less false positive (good local sender blocked incorrectly) thanks to HMM filter which is much more efficient than Bayesian.