Difference between revisions of "How Imaginet blocks spam"

From Imaginet FAQ
Jump to navigation Jump to search
(Fuzzy checksums are not really fuzzy)
Line 1: Line 1:
 
Imaginet uses a bespoke spam filtering system aimed at providing the best possible spam filtering with near-zero [http://en.wikipedia.org/wiki/False_positive#False_positive_error false-positives]. Techniques used include:
 
Imaginet uses a bespoke spam filtering system aimed at providing the best possible spam filtering with near-zero [http://en.wikipedia.org/wiki/False_positive#False_positive_error false-positives]. Techniques used include:
  
- [http://jessekornblum.com/presentations/cdfsl07.pdf Fuzzy checksums]
+
- Checking of mail against a 'fuzzy' checksum database generated from mail received by [http://en.wikipedia.org/wiki/Spamtrap spamtraps]
  
 
- [http://osbf-lua.luaforge.net/papers/osbf-eddc.pdf Advanced bayesian statistics]
 
- [http://osbf-lua.luaforge.net/papers/osbf-eddc.pdf Advanced bayesian statistics]

Revision as of 14:24, 14 August 2014

Imaginet uses a bespoke spam filtering system aimed at providing the best possible spam filtering with near-zero false-positives. Techniques used include:

- Checking of mail against a 'fuzzy' checksum database generated from mail received by spamtraps

- Advanced bayesian statistics

- URIBLs and RBLs: internal, Invaluement, SORBS, GBUDB, Whitelist ZA

- Virus and spam filtering with ClamAV

- Global and customer-operated address blacklists & whitelists

- Strict enforcement of SPF

Spam-filtered mail is delivered to a Spam Quarantine or, if there is enough of a consensus between the systems we employ, discarded.

You can help us improve accuracy by reporting misclassified mail to us, please see How to report spam to Imaginet

How you can prevent spam

How you can blacklist email addresses

What does Imaginet consider spam abuse

How to report spam to Imaginet