Optimal Auction Design with Malicious Sellers

Abstract

The presence of hardware Trojans in Integrated Circuits (ICs) poses a serious security concern for the Internet-of-Things (IoT) and the mitigation of such threats has started to attract growing attention. This paper addresses the problem by designing a novel optimal auction mechanism that can optimize the utility of a buyer (system designer) acquiring ICs in a market where the sellers (manufacturers) can maliciously insert Trojans in sold ICs. The paper also proposes, as an integral component of our designed auction mechanism, the concept of redundant acquisition of ICs from multiple sellers to obtain reliability in terms of the buyer’s system operation goals and enhance his utility as well as characterizes the optimal amount of redundancy the buyer should have in the acquisition process. The proposed auction mechanism can adapt to the imperfections in the process of testing acquired ICs. The optimal fine that should be imposed on a malicious seller in the auction upon detecting a Trojan in his sold IC is also characterized. Numerical results are presented to gain important insights into the proposed auction mechanism.