Computer Science ETDs

Mahdi Zamani

Publication Date

5-1-2016

Abstract

We live in an era when political and commercial entities are increasingly engaging in sophisticated cyber attacks to damage, disrupt, or censor information content and to conduct mass surveillance. By compiling various patterns from user data over time, untrusted parties could create an intimate picture of sensitive personal information such as political and religious beliefs, health status, and so forth. In this dissertation, we study scalable and robust distributed algorithms that guarantee user privacy when communicating with other parties to either solely exchange information or participate in multi-party computations. We consider scalability and robustness requirements in three privacy-preserving areas: secure multi-party computation (MPC), anonymous broadcast, and blocking-resistant Tor bridge distribution. We propose decentralized algorithms for MPC that, unlike most previous work, scale well with the number of parties and tolerate malicious faults from a large fraction of the parties. Our algorithms do not require any trusted party and are fully load-balanced. Anonymity is an essential tool for achieving privacy; it enables individuals to communicate with each other without being identified as the sender or the receiver of the information being exchanged. We show that our MPC algorithms can be effectively used to design a scalable anonymous broadcast protocol. We do this by developing a multi-party shuffling protocol that can efficiently anonymize a sequence of messages in the presence of many faulty nodes. Our final approach for preserving user privacy in cyberspace is to improve Tor; the most popular anonymity network in the Internet. A current challenge with Tor is that colluding corrupt users inside a censorship territory can completely block user's access to Tor by obtaining information about a large fraction of Tor bridges; a type of relay nodes used as the Tor's primary mechanism for blocking-resistance. We describe a randomized bridge distribution algorithm, where all honest users are guaranteed to connect to Tor in the presence of an adversary corrupting an unknown number of users. Our simulations suggest that, with minimal resource costs, our algorithm can guarantee Tor access for all honest users after a small (logarithmic) number of rounds.

Language

English

Keywords

Privacy-Preserving Technologies, Multi-Party Computation

Document Type

Dissertation

Degree Name

Computer Science

Level of Degree

Doctoral

Department Name

Department of Computer Science

First Committee Member (Chair)

Crandall, Jedidiah

Second Committee Member

Ford, Bryan

Third Committee Member

Trilce, Estrada

Project Sponsors

National Science Foundation

Download

Share

COinS