Author

Laura Lincoln

Abstract

Suppose there is a movie you would be interested in watching via pay-per-view, but you refuse to purchase the feed because you believe that the supplier will sell your information to groups paying for the contact information of all the people who purchased that movie, and the association of your name to that purchase could hinder career, relationships, or increase the amount of time you spend cleaning SPAM out of your mailbox. Private Information Retrieval (PIR) will allow you to retrieve a particular feed without the supplier knowing which feed you actually got, and Symmetric Private Information Retrieval (SPIR) will assure the supplier, if the feeds are equally priced, that you received only the number of feeds you purchased. Now you can purchase without risking your name being associated with a particular feed and the supplier has gained the business of a once paranoid client. The problem of SPIR can be achieved with the cryptographic primitive Oblivious Transfer (OT). Several approaches to constructing such protocols have been posed and proven to be secure. Most attempts have aimed at reducing the amount of communication, theoretically, but this thesis compares the computational expense of the algorithms through experimentation to show that reduction of communication is less valuable in the effort of achieving a practical protocol than reducing the amount of computation. Further, this thesis introduces new protocols to compete with previous published protocols that derive security from additive homomorphic probabilistic encryption schemes, and explores means to increase the length of data handled by these protocols so that the media is more useful and the time to complete the protocol is reasonable.

Library of Congress Subject Headings

Computer security; Information retrieval; Data encryption (Computer science); Computer networks--Security measures; Privacy, Right of

Publication Date

2006

Document Type

Master's Project

Student Type

Graduate

Department, Program, or Center

Computer Science (GCCIS)

Advisor

Radziszowski, Stanislaw

Advisor/Committee Member

Homan, Christopher

Advisor/Committee Member

Hart, David

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: QA76.9.A25 L56 2006

Campus

RIT – Main Campus

Share

COinS