Encrypted Traffic Classification Based ML for Identifying Different Social Media Applications

Abstract

increasing the deployment of encryption in network protocols and applications poses a challenge for traditional traffic classification approaches. Social media applications such as Skype, WhatsApp, Facebook, YouTube etc. as popular representatives of encrypted traffics have attracted big attention to communication and entertainment. Therefore, the accurate identification of them within encrypted traffic has become a big issue and a hot topic to explore them in detail. In this context, Machine Learning (ML) approaches have shown promise in this area especially for detecting and classifying the encrypted traffic data. Therefore, this work is concentrated on the challenges and has explored the ability to use ML algorithms for social media classification from traffic traces and provides a developed solution, which is able to identify the social media sub-class.