censorship resistant architectures
play

Censorship-Resistant Architectures Henry Tan Micah Sherr - PowerPoint PPT Presentation

Censorship-Resistant Architectures Henry Tan Micah Sherr Georgetown University Georgetown University Censorship-Resistant Architectures 1 The Censorship Problem Internet censorship is a problem in certain areas of the world. In some cases,


  1. Censorship-Resistant Architectures Henry Tan Micah Sherr Georgetown University Georgetown University Censorship-Resistant Architectures 1

  2. The Censorship Problem Internet censorship is a problem in certain areas of the world. In some cases, censorship may be ubiquitous, e.g. government imposed censorship. Users in these areas are unable to access restricted services are punished if they are observed accessing restricted services Most existing services do not work against these powerful adversaries. Georgetown University Censorship-Resistant Architectures 2

  3. Existing Methods Anonymity Services Tor - Relay traffic through a series of anonymizing routers. VPNs Anonymizing proxy websites etc. Very effective for their designed purposes Less so against censorship Georgetown University Censorship-Resistant Architectures 3

  4. Attack Model Powerful Adversary known to control large portion of ‘nearby’ network. performs traffic analysis participates in any communication system controls some peers outside of censored region Georgetown University Censorship-Resistant Architectures 4

  5. Adversary’s Goals The adversary wins if it discovers that Alice has been communicating with Bob is able to prevent Alice from communicating with Bob Georgetown University Censorship-Resistant Architectures 5

  6. 1) Blocking off entire communication network/service If a service: is dedicated to providing anonymity and can be identified by the adversary (e.g. entry IP addresses are published, differentiable traffic) then the adversary can shutdown the entire network without repercussions Solution Build a general purpose communication system primary purpose is not censorship resistance provides cover traffic and encourages usage, e.g. in business. Georgetown University Censorship-Resistant Architectures 6

  7. 2) Traffic Analysis SkypeMorph: shape traffic as Skype traffic FreeWave: encodes data stream as audio stream Recent work shows that this may not be enough due to VBR encoding. Solution Avoid VBR encoding. Ensure encoding schemes and traffic patterns are sufficiently similar among communication types May imply inefficient traffic for data stream Georgetown University Censorship-Resistant Architectures 7

  8. 3) Public Addresses Service’s IP Addresses must be published for clients to access them. Censor can also obtain and block these IP addresses. Solution Use trusted third parties to relay traffic to published destination. Same basic idea is used in Skype for reachability. Georgetown University Censorship-Resistant Architectures 8

  9. Approach #1 - Fully Centralized Based on standard client/server model. B roker controls network, users register with B roker Georgetown University Censorship-Resistant Architectures 9

  10. Approach #1 - Fully Centralized Users are identified by usernames and broker does not reveal IP addresses. Broker disseminates public keys and binds keys to usernames. Georgetown University Censorship-Resistant Architectures 10

  11. Requirements and assumptions Broker Does not collude with Adversary Is outside of Adversary’s influence Does not necessarily reveal Alice’s identity to Bob Use Cases More useful for multi-party communication, e.g. Google hangouts Georgetown University Censorship-Resistant Architectures 11

  12. Approach # 2 - Distributed Users create certificates - < username, public key, timestamp > signed by a known and trusted root certificate authority. Georgetown University Censorship-Resistant Architectures 12

  13. Approach # 2 - Distributed Alice contacts Charlie, a known and trusted friend. Charlie acts as a relay between Alice and Bob Georgetown University Censorship-Resistant Architectures 13

  14. Requirements and assumptions Intermediary (Charlie) Does not collude with Adversary Is outside of Adversary’s influence Does not reveal Alice’s identity to Bob Communication Network Allows for one way authentication (protect Alice’s identity) Has indirection capabilities built in and available on demand Georgetown University Censorship-Resistant Architectures 14

  15. In Conclusion ... Most existing systems are ineffective against a country level adversary. Use described architectures and techniques to build a general purpose censorship resistant network. Georgetown University Censorship-Resistant Architectures 15

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend