Faculty Advisor

Martin, William J.

Faculty Advisor

Sunar, Berk

Abstract

As critical government, industry, and consumer applications move online, techniques to maintain data security and privacy must be updated. Traditional encryption methods leave data vulnerable when it is searched or modified. Homomorphic encryption fills the gap, enabling such operations on encrypted data and eliminating the vulnerable decryption step. We worked with the CUDA-accelerated Fully Homomorphic Encryption (cuFHE) Library, the fastest of its kind in the public domain, to create efficient arithmetic functions. We built upon and modified the existing gate primitives, arranging them to create functions which are hardware and application agnostic. The result is a fast platform upon which homomorphic applications can be built: applications which protect user privacy and data integrity.

Publisher

Worcester Polytechnic Institute

Date Accepted

April 2019

Major

Computer Science

Major

Electrical and Computer Engineering

Project Type

Major Qualifying Project

Accessibility

Unrestricted

Advisor Department

Mathematical Sciences

Advisor Department

Electrical and Computer Engineering

Share

COinS