CT-ISG: Cryptographic Foundations for Next-Generation Security Applications

Princeton University, Princeton NJ

Investigators

Abstract

CT-ISG: Cryptographic Foundations for Next-Generation Security Applications Boaz Barak This project tackles some of the foundational challenges to the advent of the networked society and pervasive databases pose to cryptography. Specifically, it studies the existence of cryptographic protocols that allow mutually distrusting parties to achieve a common goal with a minimal sacrifice in privacy. This includes as a special case protocols for electronic elections, electronic auctions, privacy-preserving data mining and many more applications. A central issue in such protocols is the effect on security of executing multiples instances of one or several protocols over a network. It is known that in some cases such interaction can lead to subtle but fatal security flaws, and so constructing protocols whose security is robust with respect to such interactions is a central question in this cryptography. This project uses recent and new ideas and techniques to attack this problem. More concretely, the research involves studying the tradeoff between trust and interaction. There are known protocols with robust security assuming the existence of completely trusted third parties. This project focuses on reducing the assumption of a trusted party, and aims to discover to what extent this can be done without compromising security. The goal is to get both positive results- protocols obtaining robust security under weaker trust assumptions than was previously known, and negative results- showing that certain security goals are impossible to achieve without some trust assumptions. This will allow the users of secure applications to know what is and is not achievable in terms of trust and security. This project also tackles the main sources of inefficiency in current constructions of robustly secure protocol, in order to find out to what extent the current efficiency bottlenecks in such constructions are inherent.