One Year SpamGeek
Daniel Nashed – 3 January 2009 15:37:58
A year ago I put the first version of SpamGeek into production. SpamGeek is my own anti-spam solution based on an Extension-Manger hooking directly into the SMTP task. I decided to write my own anti-spam solution as a weekend-fun project and it is a free solution for smaller environments for up to 5 users.
There are still no commercial plans because for larger customers IBM Lotus Protector and Ironport are the best fit.
After some optimization I have around 99% spam-hit rate. I have around 2-4 spam messages per day still going thru.
Around 7% of my incoming traffic is good mail. Everything else is SPAM (see stats below).
SpamGeek gives me very detailed logging and helps me to figure out problem in the SMTP configuration of my customers and partners.
The combination of very detailed logging and a flexible rules based configuration is specially helpful for admins who want to be in full control of their SMTP traffic.
Below is a current statistic from December 2008.
Another interesting statistic is that in 80% of the cases no RBL lookup is needed to determine the message status. RBL checks are skipped when a message is already rated as "good" or "bad" thru other rules.
This can be based on simple build-in rules or the Domain/IP cache for the IP or Domain.
So after a year looking into SpamGeek and a couple of optimization steps thru lessons learned from the SPAM stats, I am very happy that I decided to build my own anti-spam solution :-)
-- Daniel
December 2008 Message Stats
7% | 933 | Accepted | ||||||||||||||||
0% | 24 | Delayed | ||||||||||||||||
2% | 265 | Greylisted | ||||||||||||||||
69% | 8743 | Blocked | ||||||||||||||||
12% | 1484 | Domino Rejected - User not found | ||||||||||||||||
1% | 150 | Tried Relaying | ||||||||||||||||
7% | 901 | Client terminated Connection | ||||||||||||||||
0% | 4 | Second Message No Data | ||||||||||||||||
1% | 77 | No Message | ||||||||||||||||
12581 |
- Comments [9]