Analysis of Molecular MUX PUFs with Stochastic Challenges

Physical unclonable functions (PUFs) are small circuits primarily used as hardware security primitives for authentication. PUFs work by generating unique signatures due to the inherent random variations in the PUF manufacturing process. This paper addresses molecular PUFs that can be used for authentication in biosecurity applications where these PUFs can be naturally implemented using bio-substrate (as opposed to electronic). This paper analyzes a novel stochastic molecular multiplexer (MUX) PUF where the input challenge corresponds to a stochastic number encoded using a fractional coding representation; thus, the challenge is not binary. Each multiplexer in the PUF is realized using six reactions. The source of randomness in the PUF is the intrinsic variations in the rate constants of each reaction. Intra-PUF and inter-PUF variations are simulated in linear MUX PUFs consisting of 8, 16, 32, 64, and 96 stages. These variations are used to compute the reliability and uniqueness of different PUF configurations. It is shown that, for the stochastic molecular PUFs to be feasible, the distributions of the rate constants require large variances.