In recent months, we’ve seen spam, phishing, and other email attacks increasingly use messages that are, as far as users can tell, identical to legitimate messages. In many cases, this is because the attackers did modify legitimate messages and use them for their attacks.
These attacks are difficult to detect using content-based systems because of this strong resemblance to legitimate messages. Any filter that is designed to detect these attacks has a good chance of having an unacceptably high false positive rate as well.
However, we have used the correlation abilities provided by the Smart Protection Network to effectively detect the messages. While these messages may be similar in content, using Big Data techniques we are able to determine other key differences between legitimate and malicious messages and identify the latter.
If anything, our correlation-based methods renders these techniques used by attackers counterproductive. Unlike traditional solutions, these techniques actually work better with emails that are harder for users to spot. Sophisticated attacks that imitate legitimate messages “stand out” with our new methodology, making it easier for Trend Micro to block these attacks. The “stealthier” an attack is to human eyes, the easier it is to spot using Big Data.
The details of this methodology will be published next week in a white paper. This methodology highlights Trend Micro’a ability to combine Big Data analytics and our existing threat expertise to create new methods to protect users and create solutions to today’s security threats.