A nonlinear-feedback shift register (NLFSR) is a shift register whose input bit is a non-linear function of its previous state.
For an n-bit shift register r its next state is defined as:
,
where f is the non-linear feedback function. [1]
Nonlinear-feedback shift registers are components in modern stream ciphers, especially in RFID and smartcard applications. NLFSRs are known to be more resistant to cryptanalytic attacks than Linear Feedback Shift Registers ( LFSRs).
It is known how to generate an n-bit NLFSR of maximal length 2n, generating a De Bruijn sequence, by extending a maximal-length LFSR with n stages; [2] but the construction of other large NLFSRs with guaranteed long periods remains an open problem. [3] Using bruteforce methods, a list of maximum-period n-bit NLFSRs for n ≤ 25 has been made as well as for n=27. [4] [1]
New methods suggest usage of evolutionary algorithms in order to introduce non-linearity. [5] In these works, an evolutionary algorithm learns how to apply different operations on strings from LFSR to enhance their quality to meet the criteria of a fitness function, here the NIST protocol, [6] effectively.
A nonlinear-feedback shift register (NLFSR) is a shift register whose input bit is a non-linear function of its previous state.
For an n-bit shift register r its next state is defined as:
,
where f is the non-linear feedback function. [1]
Nonlinear-feedback shift registers are components in modern stream ciphers, especially in RFID and smartcard applications. NLFSRs are known to be more resistant to cryptanalytic attacks than Linear Feedback Shift Registers ( LFSRs).
It is known how to generate an n-bit NLFSR of maximal length 2n, generating a De Bruijn sequence, by extending a maximal-length LFSR with n stages; [2] but the construction of other large NLFSRs with guaranteed long periods remains an open problem. [3] Using bruteforce methods, a list of maximum-period n-bit NLFSRs for n ≤ 25 has been made as well as for n=27. [4] [1]
New methods suggest usage of evolutionary algorithms in order to introduce non-linearity. [5] In these works, an evolutionary algorithm learns how to apply different operations on strings from LFSR to enhance their quality to meet the criteria of a fitness function, here the NIST protocol, [6] effectively.