Privacy-Preserving Distributed Information Sharing

Abstract

In many important applications, a collection of mutually distrustful parties must share information, without compromising their privacy. Currently, these applications are often performed by using some form of a trusted third party (TTP); this TTP receives all players inputs, computes the desired function, and returns the result. However, the level of trust that must be placed in such a TTP is often inadvisable, undesirable, or even illegal. In order to make many applications practical and secure, we must remove the TTP, replacing it with efficient protocols for privacy-preserving distributed information sharing. Thus, in this thesis we explore techniques for privacy-preserving distributed information sharing that are efficient, secure, and applicable to many situations. As an example of privacy-preserving information sharing, we propose efficient techniques for privacy-preserving operations on multisets. By building a framework of multiset operations, employing the mathematical properties of polynomials, we design efficient, secure, and composable methods to enable privacy-preserving computation of the union, intersection, and element reduction operations. Additionally, we address the problem of determining Subset relations, and even use our techniques to evaluate CNF boolean formulae. We then examine the problem of hot item identification and publication. Many applications of this problem require greater efficiency and robustness than any previously-designed secure protocols for this problem. In order to achieve sufficiently efficient protocols for these problems, we define two new privacy properties: owner privacy and data privacy. By designing our protocols to achieve owner and data privacy, we are able to significantly increase efficiency over our privacy-preserving set operations, while still protecting the privacy of participants.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2006
Accession Number
ADA459056

Entities

People

  • Lea Kissner

Organizations

  • Carnegie Mellon University

Tags

Communities of Interest

  • Biomedical
  • Cyber
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Anonymous Communications
  • Computations
  • Computer Science
  • Computers
  • Construction
  • Counting Methods
  • Cryptography
  • Data Sets
  • Detection
  • Identification
  • Information Exchange
  • Information Science
  • Notation
  • Simulations
  • Simulators
  • Statistical Analysis

Fields of Study

  • Computer science
  • Mathematics

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Cybersecurity.
  • Systems Analysis and Design