Modifying Network Traffic Profiles to Counter Recognition

Abstract

With the rapid advancement of networking techniques, network traffic analysis has become key part of network management. Network traffic detection creates open threats to privacy and confidentiality of sensitive data. Different encryption techniques are being used to hide network information. But encryption doesn’t hide statistical properties of a packet, therefore, network traffic still can be classified by manipulating these properties. In our research, we proposed an approach in which we added additional data to WebRTC data channel associated with RTC peer connection that modified packet sizes. We modified network profile of live streams such as Audi Stream, Video Stream, and Screen Sharing Stream, that helped us to show false data of network traffic profile to fool detectors and censoring authorities to circumvent cen sorship. As WebRTC data channel is peer-to-peer (P2P) connection, so sharing data through it is also mimicking signaling server.