With the arrival of big data era

sensitive data protection becomes the major challenge facing telecom operators. Sensitive data identification is the prerequisite for data protection. Deep packet inspection can detect and identify the flow in real time with loose coupling

make it become the big data platform of sensitive data identification an important way. However

the high storage and computing resources demand become restricted its application and development. This paper puts forward a sensitive data identification based on the feedback

through establishing the session table

real-time records of the state

sharply reduce the amount of data need to depth detection

improve the efficiency of detection.