MPP and Beyond

I’m a Blackberry addict and there is nothing more frustrating then jumping for the stupid thing as it buzzes itself into a fit, expecting to find my next multi-million dollar deal, only to find some random text spam with unreadable images. It makes my friends laugh when I do blackberry acrobatics to chase the thing down as it buzzes away, hoping for something big, and I find this junk that I’m supposed to filter anyway.

So the other day, as my friend was laughing his head off about the beer I spilled to grab my bb, I was angered to find yet another bunch of jibberish spam on my bb. I told my friend about my frustration, which made him laugh even harder as I had just wasted a perfectly good beer for spam, and he asked me a great question - What’s the deal with that random text spam anyway?

Before we get there, let’s take a look at some random text spam..

rgently onus mention squire, monotone of dark glasses in hotly by cassette. clerical prudently planetary ancestral,? gun the people, coarse. music the plaster is dental wayward delusion the amateur filibuster was
ichen white water a forgiveness and ain’t undertaking at ceramics as!!! sleeplessness cuticle kilt gush launchpipe but? insemination and drowsily rollicking a

——-

Pretty random, so what’s the point? There are many types of spam filters, some look at the origin, authenticity and construction of an email message, some take fingerprints of email and compare them against databases of spam prints, while others analyze the content of email to find ’spammy’ content. There are many methods to analyze content and one method uses a token scheme that assigns a token to each word in an email. As more email passes a system a profile of normal email is built based on token modles and abnormal email can be found based on this token profile. By throwing in tons of random words there is a good chance that you will increase the number of token matches for an email, thus decreasing the likelihood that it is a spam message.

The text is in the email is only to confuse token systems, the real message is embedded in an image. Since it is really difficult for content scanners to read text in images, and your typical stupid email client will display images for you, you will see the spam message in a graphic which the filters can’t analyze proceeded by text that confuses some token based systems.
This has been a highly effective form of spam that is difficult to track with content analysis so it can pass many spam systems. A fingerprint system or verification of origin test can stop this type of spam in many cases and these techniques are the most effective with this type of spam. Of course there are great content scanners that can find text in images, but this is our experience.
So, this is what’s up with random text spam, hope it was useful.

Technorati Tags: antispam, random text, phishing, fraud, spam