A Mantrap-Inspired, User-Centric Data Leakage Prevention (DLP) Approach

The ease of sharing information through the Internet and Cloud Computing inadvertently introduces a growing problem of data leakages. At the same time, many end-users are unaware that their data was leaked or stolen since most data is leaked by operations running in the background. This paper introduces a novel user-centric, mantrap-inspired data leakage prevention (DLP) approach that can discover, present any sending of data -- both authorized and unauthorized -- to end-users and subsequently provide them the ability to stop the sending process. We implemented our own kernel module to work together with our user-space program in getting user's approval for every sending process -- giving the user full control over all outbound data sending process in their devices. With this, the end-user can always decide which data sending process should be allowed or blocked. This overcomes the limitations of current, often inflexible and inaccurate DLP solutions depending on pre-set rules and content detection. We showcase a proof-of-concept for our new way of detecting data leakages in an end user's device. This paves the way for further research covering more complex data stealing techniques, such as the use of covert channels.