dspam implementation?

anything with anti virus, and a good spamfilter success, and low recourses is welcome :D
 
Before anyone accuses me of hijacking a thread, let me point out that Scott, the original poster, decided to put this thread into the SpamBlocker3 subforum, which makes this post fair game ;) .

I ran a test on the catch rates of the latest (private beta) version of SpamBlocker3 over the past few days and I find that I'm finally getting good on blocking spam. For example, switching to zen.spamhaus.org has almost tripled the catch rate without increasing any known false-positives.

Here are some specs from one of my servers (note this is from a private beta, not yet available):
all email blocked by blocklists: 05-21-2007: 5046
all email blocked by blocklists: 05-22-2007: 5350
all email blocked by blocklists: 05-23-2007: 6116
Click to expand...
and
all rejected/delivered email: 05-21-2007: 48622 / 9839 | 79% rejected
all rejected/delivered email: 05-22-2007: 38656 / 9252 | 76% rejected
all rejected/delivered email: 05-23-2007: 40412 / 8481 | 79% rejected
Click to expand...
And perhaps even more important, on my personal addresses spam is now less than a third of what it was with previous versions of SpamBlocker3.

I have one more set of changes to test, and I think by Monday we should know what we want to use.

I can honestly see SpamBlocker3 going live in a week or or, and I hope that since it's included in DA, admins will give it a chance.

Jeff
 
jlasman said:
Before anyone accuses me of hijacking a thread, let me point out that Scott, the original poster, decided to put this thread into the SpamBlocker3 subforum, which makes this post fair game ;) .

I ran a test on the catch rates of the latest (private beta) version of SpamBlocker3 over the past few days and I find that I'm finally getting good on blocking spam. For example, switching to zen.spamhaus.org has almost tripled the catch rate without increasing any known false-positives.

Here are some specs from one of my servers (note this is from a private beta, not yet available):

and

And perhaps even more important, on my personal addresses spam is now less than a third of what it was with previous versions of SpamBlocker3.

I have one more set of changes to test, and I think by Monday we should know what we want to use.

I can honestly see SpamBlocker3 going live in a week or or, and I hope that since it's included in DA, admins will give it a chance.

Jeff
Click to expand...


nice work jeff
look forward to the release.
 
I'm throwing in my vote for DSPAM for SpamBlocker 4 too. It is HIGHLY efficient if it is trained properly.

Here's the statistic for my own accounts for the curious:

Overall accuracy (since last reset) 98.363% (SPAM messages caught + Good messages delivered) / Total number of messages

Spam identification (since last reset) 95.491% (Spam catch rate only)

Spam ratio (of total processed) 35.039% Total SPAM messages (both caught & missed) / Total number of messages

I have been using it for 6 months now and I train it religiously. It makes a huge difference. You get a nice web panel to manage the quarantine email.

Other workers at my job have an overall accuracy between 94% to 98%.
 
I'm still a bit confused about DSPAM.

How does it integrate with exim.conf?

Jeff
 
How does it integrate with exim.conf?
Click to expand...

Sadly, I have no experience with Exim and DSPAM. The little I know(since I wasn't the one to do the setup) is with Postfix and DSPAM.

Our current setup is that we have a DSPAM+Postfix server. Mail goes through there first and is then relay to our final mail server. DSPAM act as a filter and will hold in quarantine the SPAM. (Which can be accessed through a web interface) There is nothing preventing you from having DSPAM on the same server. We used different server because it was easier for testing.

If you want, I could send you our DSPAM and PostFix config files. Maybe it would point you in the right direction? I don't have enough experience with Exim but maybe I'll have a look at it myself once I have some spare time.
 
You must log in or register to reply here.
Back
Top