Unconditional Security using (Random) Anonymous Bulletin Board
Albert Yu, Purdue University, United States; Hai H. Nguyen, ETH Zurich, Switzerland; Aniket Kate, Hemanta K. Maji, Purdue University, United States
Session:
Secure Multiparty Computation
Track:
14: Secure Communication and Computation
Location:
Ypsilon IV-V-VI
Presentation Time:
Wed, 10 Jul, 11:50 - 12:10
Session Chair:
Shun Watanabe, Tokyo University of Agriculture and Technology
Abstract
In a seminal work, Ishai et al. (FOCS–2006) studied the viability of designing unconditionally secure protocols for key agreement and secure multi-party computation (MPC) using an anonymous bulletin board (ABB) as a building block. While their results establish the feasibility of key agreement and honest-majority MPC in the ABB model, the optimality of protocols with respect to their round and communication complexity is not studied. This paper enriches this study of unconditional security in the ABB model in multiple ways. - We present a key agreement protocol with a novel combinatorial insight to offer a 200% throughput over the (FOCS–2006) study; i.e., using the same number of messages, we can (almost) double the bit-length of the agreed key. We also prove the near optimality of our approach. - We offer unconditionally secure protocols for the (random) string oblivious transfer functionalities. We present a 1-round chosen message random string oblivious transfer and show how to extend it to a non-interactive (random) string oblivious transfer protocol and a 2-round chosen message string oblivious transfer. - We prove a 1-round lower bound for BEC under certain conditions. Central to our technical contributions is the abstraction of a distributional variant of the random ABB functionality. Investigating the concrete efficiency of founding MPC from this primitive leads to fascinating new mathematical challenges in well-established MPC models, which will be of broader interest to the community.