Scientists have advanced the fight against spam
Drawing main trick against the spammers themselves, scientists have found a way to recognize unwanted promotional mailing with a 100-percent accuracy.
As the portal “Popular Mechanics”, the lion’s share of billions of spam messages that come users per day, distributed botnet, network computers ‘zombies’. With the help of special programs botnet creators gain control of part of the resources of the remote computer, and he slipped to its owner is included in the network, obeying the commands from the outside. Such networks are often surrender “in the hiring of” spammers who send them via the countless harassing messages, which (with varying success) are struggling creators of spam filters.
Modern filters can block almost 100% of incoming spam, but they also have an unpleasant feature: they are often mistaken in removing completely harmless (and sometimes very necessary) letters. The difficulty is that even one spam message is to recognize not so simple. Clever spammers use their templates that are included in the botnet computers modify according to a specified algorithm. Letter to “mutate”, and know it for some phrases sometimes simply impossible.
However, the solution created by scientists from Berkeley, can circumvent this problem. Or rather – to draw it to fight spam. Whatever the “mutated” the letter, with the help of special algorithms can be based on several of these messages to recreate the original template, and only then “fed” his spam filter to select letters, based on it. Screenings will be affected by any letters that are created from this template, on any computer on the botnet.
Programmers had to check their approach to the case. For this they have a program that unknown persons used for “involvement” of computers in “criminal activity”, in other words – in the botnet. Program installed on your computer and within minutes he had “betrayed” them about a thousand spam messages that are generated on the basis of the original (unknown programmers) template.
This thousand, but was sufficient to restore the original template and the transfer of its spam filter. Effectiveness has been tested with analysis of 1 million e-mails, both conventional and undesirable. Neatness recognition based on it Message was 100 per cent, and most importantly – “good” messages removed was not a single.
