% Anonymous Communications
% Andrew Lewman andrew@torproject.org
% December 05, 2012

# Who is this guy?

501(c)(3) non-profit organization dedicated to the research and development of technologies for online anonymity and privacy

[https://www.torproject.org](https://www.torproject.org)

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/2009-oval_sticker_new}
\end{center}

# Three hours of this guy talking?

Let's hope not.

Ask questions; early and often.

# Agenda
 - Definitions and Concepts of Anonymity
 - What data?
 - Attacks against anonymity
 - Deployed Systems (Centralized and Decentralized)

# What is Anonymity?
\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/2llg3ts.jpg}
\end{center}

# Definitions: Anonymity

  - a set of all possible subjects
  - state of not being identifiable within anonymity set

# Definitions: Unlinkability

  - unlinkability of two or more items of interest from the adversary's perspective
    - items can be messages, people, events, actions, etc

# Definitions: Unobservability

  - state of items of interest being indistinguishable from any items of interest

# Definitions: Pseudonymity

  - identifiers of sets of subjects

# Definitions: Traffic Analysis

  - The who, what, when of traffic
  - Think of the post office

# Definitions: Steganography

  - the art and science of writing hidden messages in such a way that no one, apart from the sender and intended recipient, suspects the existence of the message, a form of security through obscurity. -- [Wikipedia](https://en.wikipedia.org/wiki/Steganography)
  - alice or bob are talking, but to whom?

# Definitions: Cryptography
  - protecting content in transit
  - does nothing to hide the traffic of items of interest

# What data to protect?

 - bits of info which put together deanonymize you
  - Names of individuals
  - location address (street, IP address, zipcode, etc)
  - operating system info
  - language info
  - amount of data sent
  - amount of data received
  - traffic timing (heartbeats)

# Anonymity Loves Company
  - make the set of users as large and coherent as possible to create a large anonymity set

# Attacking Anonymity: Timing Analysis
  - An attack used to analyze the time properties of data transfer between items of interest.
    - When was data sent?
    - How much was data sent?
    - How long did it take to send the data?
    - When was data received?
  - wireshark demo

# Attacking Anonymity: Timing Analysis

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/iograph-torproject-website.png}
\end{center}

# Attacking Anonymity: Statistical Disclosure

  - Also called an intersection attack
  - trying to identify mutually disjoint sets of recipients
  - exponential time involved per number of messages to be analyzed

# Attacking Anonymity: Tagging

  - tagging (make one item of interest unique)

# Attacking Anonymity: Traffic Confirmation

  - who sends, how often, and when
  - etherape demo

# Centralized Systems

 - cheap, easy, ubiquitous
 - PPTP, IPSec, SSL, SSH, XMPP common protocols

# Proxy and VPN Servers
  - proxy server works on your behalf
  - VPN is virtual private network
    - proxy for the network layers (layers 2 or 3 of OSI model)

# Proxy and VPN Servers

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/single_hop_relay.jpg}
\end{center}

# Trusting the provider

  - trusting the provider
  - promises, contracts, mistakes
  - some may filter or clean data before passing on to destination

# Trusting the provider

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/evil_single_hop_relay.jpg}
\end{center}

# Irrelevant provider

  - Single machine, or cluster of machines, are connected to a network
  - If the proxy provider won't cooperate, use the network around it.

# Irrelevant provider

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/data_snooping_single_hop_relay.jpg}
\end{center}

# Decentralized Systems

## Mix Networks
 - cascades (JonDos/JonDonym)
 - routes (tor)

## Similar Routing networks
 - I2P - Garlic routing, closed network, anonymity and reputation
 - Freenet - closed network, anonymity, distributed file storage and sharing
 - GNUnet - closed network, anonymity, distributed file storage and sharing

# Break?

Anyone need a bio-break for 10 minutes?

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/wwwhatranorg}
\end{center}

# What is Tor?

 - online anonymity software and network
 - open source, freely available (3-clause BSD license)
 - active research environment:

Drexel, Univ of Waterloo, Georgia Tech, Princeton, Boston University, University College London, Univ of Minnesota, National Science Foundation, Naval Research Labs, Cambridge UK, Bamberg Germany, MIT...

 - increasingly diverse toolset:

Tor, Tor Browser Bundle, Tails Live System, Orbot/OrWeb, Tor Weather, Tor auto-responder, Secure Updater, Arm, Tor2Web, and so on.

# Who uses Tor?

\parbox{8cm}{\sloppy \setbeamercolor{background}[\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/anonymousman}}
\parbox{3cm}{\sloppy
\begin{flushleft}
\begin{itemize}
\begin{small}
\item Normal people
\item Journalists
\item Law Enforcement
\item Human Rights Activists
\item Business Execs
\item Militaries
\item Abuse Victims
\end{small}
\end{itemize}
\end{flushleft}
}

# How many people use Tor?

estimated 500k to 900k daily users

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/huge-crowd}
\end{center}

# How does Tor work?

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/tor-network}
\end{center}

# How does Tor work?

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/tor-safe-selection}
\end{center}

# How does Tor work?

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/tor-safe-path}
\end{center}

# How does Tor work?

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/tor-keys1}
\end{center}

# Attacks on Tor

 - First hop can learn your IP address.
 - Last hop can watch your traffic.

# Attacks on Tor

 - Adversary can block all Tor nodes by IP address and TCP port
    - our answer is to use non-public relays called Bridges
 - Adversary can legally harass last hop; DMCA, Child Abuse Materials, Threats, etc
 - Adversary can run relays, use network to restrict access to other relays

# Attacks on Tor

 - Deep Packet Inspection

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/procera-evolved-dpi}
\end{center}

# Attacks on Tor

\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/iran-ssl-dpi-26-seconds-to-death}

# The Future: Usability

 Who are our users?  What do they understand about anonymity, Tor, and privacy online?
 Can we guide them to make smarter decisions?
 How do we educate them before they start?

# The Future: Obfsproxy & Pluggable Transports

 Obfuscating proxy for network traffic

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/obfsproxy_diagram.png}
\end{center}

# The Future: Scaling

  Why not 10,000 relays? Why not 1 million? 10 million?
  Need privacy-preserving Scalable Distributed Hash Table designs

# The Future: IPv6

  Basic support for IPv6 clients and relays works now.
  Need support for IPv6 destinations and pure IPv6 relays

# The Future: UDP

  Tor only transports TCP packets now.
  This limits usable applications
  Need to support real-time video and audio chats over Tor.

# Thanks!

\begin{center}
\includegraphics[keepaspectratio,width=\textwidth, height=.8\textheight]{images/thankyou_img}
\end{center}

Visit [https://www.torproject.org](https://www.torproject.org) for more information, links, and ideas.