Systematic Low Leakage Coding for Physical Unclonable Functions

Physical Unclonable Functions (PUFs) derive unique secrets from internal manufacturing variations in integrated circuits. This work shows that key generation with PUFs is a practical application of the generic information theoretic problem of secret key agreement with a compound source. We present an improved secure sketch construction with our new optimal syndrome coding scheme for PUFs, Systematic Low Leakage Coding (SLLC). Our scheme provides inherent information theoretic security without the need of a hash function or strong extractor, and optimal asymptotic performance concerning maximum key size and minimum helper data size. The secrecy leakage is bounded by a small epsilon that goes to zero for sufficiently good PUFs. The reference implementation for an ASIC application scenario shows that our scheme does not require the 47% hardware overhead for the hash function that is mandatory for the state-of-the-art approaches.     «

Physical Unclonable Functions (PUFs) derive unique secrets from internal manufacturing variations in integrated circuits. This work shows that key generation with PUFs is a practical application of the generic information theoretic problem of secret key agreement with a compound source. We present an improved secure sketch construction with our new optimal syndrome coding scheme for PUFs, Systematic Low Leakage Coding (SLLC). Our scheme provides inherent information theoretic security wit...     »