351951 event 1417690909 1475892639 Ph.D. THESIS PROPOSAL

TITLE: Secure Multiparty Computation Outsourcing Techniques for the Mobile Platform

Henry Carter
School of Computer Science
College of Computing
Georgia Institute of Technology

Date: Monday, December 8, 2014
Time: 10:00 AM – 11:30 AM EST
Location: KACB 3100


Committee:

Dr. Patrick Traynor, Advisor
School of Computer and Information Science and Engineering, University of Florida
Dr. Mustaque Ahamad
School of Computer Science, Georgia Institute of Technology
Dr. Sasha Boldyreva
School of Computer Science, Georgia Institute of Technology
Dr. Chris Peikert
School of Computer Science, Georgia Institute of Technology
Dr. Kevin Butler
School of Computer and Information Science and Engineering, University of Florida


Abstract:

Mobile phones are becoming a primary computing platform for millions of users worldwide.  Many mobile applications
provide convenient services requiring users to input GPS location, social networking data, or other potentially private 
information.  However, these applications cannot always be trusted to maintain the privacy of that user data.  To solve
this problem, Secure Multiparty Computation (SMC) provides a cryptographically secure means for evaluating a
function over encrypted inputs.  However, these constructions typically use expensive underlying operations, and
require computation and communication resources that are not available on a mobile device.

In this proposal, we present a set of novel techniques for securely outsourcing the most costly operations in an SMC
protocol to a Cloud provider.  Rather than naively trusting the Cloud to perform the secure computation in place
of the mobile device, we show that it is possible to add input and output verification to the computation to prevent
a malicious Cloud from tampering with the computation.  We demonstrate the feasibility of our protocols
through benchmark testing, and present a variety of example applications that solve practical privacy problems
on mobile devices.  Ultimately, we plan to demonstrate how these techniques can be modified and streamlined to
satisfy a range of potential real-world threat models.

]]> 221981 1788 1808 102851