How does Facet work?

Facet is a censorship circumvention system that enables the clients in a censored regime to watch YouTube, Vine and Vimeo videos in realtime. The basic idea of Facet is to send videos from these sites as the video content of a videoconferencing call, in the case of our prototype, a Skype call, between a Facet server and a client. Like all proxy steganography systems, Facet relies on the assumption that the censor is unwilling to indiscriminately block all or most sessions of the cover protocol (Skype) to avoid collateral damage.

The Advantage of Facet:

A. Facet is Emulation Consistent: because the video is transmitted over an actual two-way Skype call, there is no difference between implementations to allow identification.

B. Facet is Channel Consistent: because we transmit videos over a channel intended to transmit videos, any disruption to a Facet session would cause the same level of disruption to a regular Skype call.

C. Facet is Content Consistent: Arbitrary videos may have different characteristics from typical video conferencing calls, leading to detectable differences in packet sizes. We implement a state of art binary classifier and show that unaltered YouTube videos sent over Skype are distinguishable from typical Skype calls. To defeat this, we introduce video morphing, in which the Facet server frames the requested video within a randomly selected videoconference call. This increases the false positive rate of a classifier that can recognize 90% of Facet calls to nearly 40%.

D. No deployment at client side: for Facet clients, there is no need to install any client software (which is often blocked), or to pre-share secrets with the server. This property makes Facet easier to use and maintain, since software updates only need to be ap- plied by servers outside of the censored region.