We numerically study a new hardware method for network authentication, using a pair of chaotic lasers subject to the same optical injection from a third chaotic laser, in a close-loop configuration. One laser is in the secure environment, the other is in the unsecure environment, and if their parameters are matched (i.e., they are ‘twins’), in suitable operating conditions they synchronize and generate the same pseudo-random amplitude modulation. Access to the secure environment is authorized only if the two responses are found equal. As in other schemes based on Physical Unclonable Functions (PUFs), security relies on unavoidable small differences between nominally identical but physically separate devices: due to the strong sensitivity of chaos to laser parameter dispersion, finding a suitable laser to synchronize is a very difficult task for the Adversary, while the laser of the Authorized user can be selected as a twin of the one in the secure environment.
Read full abstract