High-Performance Fault Diagnosis Schemes for Efficient Hash Algorithm BLAKE

Augmenting the security of cryptographic algorithms by protecting them against side-channel active attacks (and natural faults) is essential in cryptographic engineering. BLAKE algorithm is an efficient hash function which has been developed based on Bernstein’s ChaCha stream cipher. Because of the fact that Google has chosen ChaCha along with Bernstein’s Poly1305 message authentication code as a replacement for RC4 in TLS for Internet security, BLAKE’s implementation is of paramount importance. In this paper, we present high-performance fault detection schemes for BLAKE. Specifically, for the round function, two fault diagnosis approaches are developed and analyzed in terms of error detection capability and overhead. Through our injection-based error simulations, we show that the error coverage of almost 100% can be achieved for the proposed approaches. In addition, through hardware platform benchmarks, we show that the proposed architectures have implementations which reach acceptable area/delay overheads. The proposed high-performance fault diagnosis approaches will make the hardware implementations of BLAKE more reliable.