Content filtering spam is on its way out now.

Postini says that content based filtering is so last year. The method of looking at a message's text itself and judging if it is spam or not is increasingly not going to work as spammers are increasingly only using images, URLs, and phone numbers with nor real text so to speak of.

So the spam is going to have to be blocked by looking at the sender. This is the way Microsoft's new version of Sender ID and the related SPF work. The general idea is that they look at the IP address from which a message arrived and then look to see where it claims to have been sent from.
If those don't match, and the registered IPs for that domain don't include the one that the message was sent from, then it is a flagged message.

While that all sounds good, there is going to be a lag period as companies need to come on board, registering addresses. Even then the fact that there is a doesn't mean that it is discrepancy spam - so it ideally should be combined with a series of other tests. If it fails more than one, than it is much more likely to be spam than if it is only failing a single test.

Posted by Eric at August 16, 2004 09:44 AM | TrackBack