Simulating Private Information Retrieval on Amazon Web Services

Abstract

As our modern lives have gradually moved more and more online, companies and state actors have taken it upon themselves to gather and analyze our behavior online, and as these actors have gradually shown just how much they know about a private user, or a group of users, a concern for privacy has grown accordingly. A virtual private network service could help anonymize a user, but the providers of services usually log what services they provide, which can provide identifying information. Research in privacy measures have thus become a larger topic in recent time. Private information retrieval allows a user to query a database without revealing to the server any information about the information queried, and if effective enough, could provide perfect privacy to everyone. In this thesis, we examine a state-of-the-art efficient private information retrieval scheme and study every step in the protocol in a simulation implemented on Amazon’s cloud computing services.