This Linux Foundation project focuses on mechanisms to provide in-network network obfuscation techniques that can confuse machine learning techniques at fingerprinting data flows.
This work follows on previous works such as Dust, Marionette, Taranet, BuFLO, etc. Traffic, even when encrypted can be fingerprinted and analyzed, violating user privacy. The purpose of this work is in the use case where multiple entities operate a network or overlay, where any node in the path can passively attempt to fingerprint flows which transit through it. Multiple methods have been implemented which have varying levels of obfuscation.
Installation and Usage instructions are located in the top-level directory for the desired approach.
For the time being, contributions can be done through pull requests. There does not have to be an issue open or associated with the pull request.
BSD-3