The protection of Big Data is a absolute priority constantly sought by enterprises. Bad management practices may lead to data theft or impact the vulnerability of the platform.
Traditional IT security is not scalable and can't ensure an effective protection of Big Data. On the other hand, a security based on data can bring a real response to this challenge.
In this sense, we will focus on the protection of sensitive data within Hadoop, the main Big Data platform.
Our action consists in proposing a dynamic framework, which calculates the sensitivity of the data in an automated way, without any intervention of the data's owner . 
The sensitivity of data changes over time based on the scenarios provided by our scalable infrastructure. 
The intended purpose is to protect sensitive data in the Hadoop cluster, and keep it out of reach of unauthorized users.