ExpFault: An Automated Framework for Block Cipher Fault Analysis

Abstract

Intentional or unintentional faults in computation or control flow of modern computing systems can be catastrophic if they are exploited maliciously. One of the primary targets of malicious faults are the modern cryptographic primitives—especially the block ciphers. Designing precise countermeasures against such threats is desirable, but that requires comprehensive knowledge about the attack space of the cryptographic algorithm as well as the implementation. Although manual efforts for vulnerability analysis of fault attacks is quite widespread, it becomes infeasible with the fact that the number of existing ciphers is quite large today. Formidable size of fault spaces and their diversity further highlights the need for automation in this context. Since fault attacks for block ciphers are critically dependent on the mathematical properties of the cipher, a comprehensive understanding of the attack space thus should begin with automated algorithmic analysis. This chapter lays the foundations for automated fault analysis of ciphers and presents a framework called ExpFault for automated algorithm-level fault analysis.