Research On Facebook Video Identification Based On Accurate Construction Of Encrypted Fingerprint

To protect user privacy and ensure network security,major video platforms have adopted various encryption technologies to transmit video traffic.Traditional DPI-based traffic detection methods are difficult to apply to the identification of encrypted video traffic.Therefore,how to effectively identify encrypted video traffic is becoming important to not only better network supervision,but also the establishment of a healthy and harmonious network environment.At the same time,it also helps network service providers to monitor network conditions more accurately and has an important role in promoting network service quality.Existing works on encrypted video traffic recognition focuses on the research of matching recognition algorithms,but there is a lack of specific research on the characteristics of the encrypted video traffic and a comprehensive evaluation of the results of these methods in a large video dataset.As a result,the practicability of existing works cannot be guaranteed.Therefore,how to effectively identify the video from the encrypted video traffic is of great significance.To solve the above problems,this thesis proposes a new method for identifying encrypted video traffic through in-depth analysis of the key characteristics of encrypted video traffic,which mainly includes the following research contents:(1)This thesis proposes a method HHTF to restore the size of encrypted video ADU based on the TLS encapsulation features.This method can accurately restore the encrypted video ADU size.By analyzing the traffic characteristics of the encrypted video ADU in the network transmission,the main key characteristics that affect the size of the encrypted video ADU are proposed:the size of HTTP header encapsulated in the TLS layer,Length_HTTPHeader,the number of TLS fragments required by the TLS layer to encapsulate a video ADU,N_TLSFragment,the total size of an encrypted video ADU,Length_PaylaodSum.Using the above features,a linear regression model for restoring the size of the encrypted video ADU is constructed.This model uses the ground size of the encrypted video ADU as the label.The experimental results show that the method for restoring the size of the encrypted video ADU based on the TLS encapsulation features proposed in this thesis has a very high accuracy rate.(2)This thesis presents an encrypted video traffic identification method based on video fingerprints.This method is based on the two ways of HHTF restoring the size of the encrypted video ADU,and proposes two different ways of constructing the encrypted video transmission fingerprint,and uses the ground size of the video ADU to construct the corresponding plaintext fingerprint.Finally,the fingerprint sequence matching algorithm is used to identify the video represented by the encrypted fingerprint.To verify the effectiveness of the encrypted video traffic identification method proposed in this thesis,the most competitive encrypted video traffic identification method in the existing research work was reproduced and compared in the same large-scale video data set.The experimental results show that the method proposed in this thesis has obvious advantages especially in the face of a large dataset.(3)Based on the above method,a prototype system that can be used to demonstrate the performance of encrypted video traffic identification is designed and implemented.First,the overall framework of the prototype system is designed,which mainly introduces the design and implementation of the prototype system from the aspects of encrypted video ADUs size restoration,construction of video fingerprints,matching between encrypted video fingerprints and plaintext fingerprints,and user interface display.The system supports the functions of manually adjusting various parameters,displaying and saving the final results,and has strong practicability.