Re: public blacklists

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 2004-12-09 at 09:06, Ow Mun Heng wrote:
> > 
> > No, greylisting is a completely different thing. Greylisting ensures that the 
> > sending server is a proper MTA that retries when it sees a temporary failure 
> > during a delivery attempt. Most spamware does not do this, hence greylisting 
> > stops lots of spam. SURBL is looking at the message body after delivery and 
> > scoring it as likely to be spam or not based on the URLs found there. Two 
> > completely different things.
> 
> In that case, in some cases, eg: if one runs their own mail-server,
> grey-listing seems to be a better option compared to spamassassin, even
> when using SURBL.
> 
> Reason being, greylisting stops it at the MTA level, spamassassin only
> tracks it once it's already in the system.

Yes you need to have control of the MTA receiving the messages to
implement greylisting.  I would say that greylisting complements
spamassassin and vice versa.  Spamassassin is great at assigning a score
to a message, as you indicate once it is on the system.  Greylisting
eliminates a large portion of spam messages before they every get on
your system.  

I initially implemented spamassassin which after a few days of training
bayes and the addition of a few extra rule sets worked wonders.  It
tagged virtually all spam coming into the system.  However the down side
was that someone still needed to review the spam bucket for false
positives.  In addition during severe spam storms the mail server would
bog down trying to process all those messages through spamassassin. 

I then implemented greylisting which at the time I figured would only
help a little.  I was amazed to find that it actually blocked 90 to 99%
of the spam we were receiving.  And because spamassassin did not have to
process all those messages the system does not break a sweat anymore.  

Running the combination of greylisting and spamassassin has provided
virtually 100% elimination of spam.  And surprisingly the maintenance of
both is minimal.  A few scripts that are run take care of feeding bayes
and the greylisting configuration has required minimal support once the
whitelist was populated with known associates.  

I would recommend using both systems as they handle different aspects of
the spam problem.  Plus it provides some depth to the defenses.  If/when
the spammers start figuring out ways around greylisting spamassassin is
there backing it up.  But so far there has been little sign that the
spammers are bothering to work around the issue.

-- 
Scot L. Harris
webid@xxxxxxxxxx

To err is human, to forgive unusual. 


[Index of Archives]     [Current Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]     [Fedora Docs]

  Powered by Linux