Project Title: Measurement-based Design and Analysis of Censorship Circumvention Schemes
Project Period: September 1, 2013 - August 31, 2018
The Tor network is one of several widely-used censorship circumvention tools designed to allow users to safely and efficiently access censored content by “routing around” censorship technology through one or more relays. Using these relays makes it difficult for a censor to determine what content is being accessed by whom. Unfortunately, this difficulty also makes it hard to measure the use and performance of such tools, which in turn leads to challenges in assessing the impact of the tools, and the security and performance of proposed alternatives.
The goal of this project is to develop protocols, collect data, and deploy software to overcome this difficulty.
• Develop and deploy protocols to privately measure the use of Tor. • Design and evaluate performance and security improvements for Tor. • Design and analyze generic techniques to improve blocking circumvention.
Notes from 2014 Dev Meeting
measure relay level stats
- bandwidth/ewma on circuits, or connections, exit connections
- num circuits exiting Tor
- num hidden service rendezvous
- fraction of circs that are hidden service circs over last 24 hours
measure user statistics/model
- how many are online (countMe)
- how often online and how often they stay away (churn)
- better to not store stats in RAM for a long time
- differential privacy -> add noise so that at any point, it is safe (in RAM or not)
- anything collected about clients needs to be clear why it benefits clients
- collaborative client/guard scheme
- separate processes? simpler implementation with ec already in Tor
there is a large gap between what makes an acceptable (interesting to researchers) diffpriv system and what should be implemented and maintained in Tor. how do we bridge the gap?