How Katharion analyzes messages

What is spam?

A multi-layered approach

Katharion uses a multi-layered approach to message analysis, using a wide variety of tests to accurately detect junk mail. These tests include established, widely deployed techniques, as well as several proprietary next-generation approaches developed by Katharion that leverage the company’s experience in processing billions of messages.

Ensuring accuracy

In order to avoid legitimate messages being erroneously identified as spam, Katharion is designed so that no single test by itself will tag a message as spam. At the same time, the large number of tests, combined with Katharion’s wide breadth of spam identification techniques, makes it extremely difficult for junk mail to pass through undetected.

Combining proven and next-generation techniques

Katharion incorporates traditional spam detection techniques such as:

continuously updated heuristic rules that identify common phrases or patterns in spam message body text or headers;
message signature analysis to compare the ‘fingerprint’ of incoming email messages to known junk mail;
blacklists of mail servers, domains, or networks known to be used by spammers;
message authenticity checks based on header analysis and other data;
a probabilistic engine using Bayesian algorithms to assess the likelihood of a message being spam; and
company-specific whitelists and blacklists that allow or reject messages based on the sender

Katharion has also pioneered industry-leading techniques such as real-time message source analysis; graphics analysis; advanced authenticity checks; dynamic feedback-based rules optimization; user-based message profiling; and reputation analysis.

Together, these elements provide a proven, highly effective system for accurately identifying junk mail.

Next: How spam is handled...