Efficient Evaluation of Physical Unclonable Functions Using Entropy Measures

Physical unclonable functions (PUFs) are security primitives which are based on uncontrollable variations in the manufacturing process and can be used as secure key storage and for identification and authentication. While some work has been done on global evaluation of PUFs with respect to predictability, there is no systematic evaluation strategy to give designers insights into design problems. However, flaws in the design can give an adversary an advantage to guess the response of a PUF. Therefore, this work focuses on a systematic scheme for designers to evaluate the predictability of PUF implementation. Also, the used entropy measures are compared to state-of-the-art metrics for PUFs, convergence properties are presented and the scheme is validated using simulation results for PUFs.      «

Physical unclonable functions (PUFs) are security primitives which are based on uncontrollable variations in the manufacturing process and can be used as secure key storage and for identification and authentication. While some work has been done on global evaluation of PUFs with respect to predictability, there is no systematic evaluation strategy to give designers insights into design problems. However, flaws in the design can give an adversary an advantage to guess the response of a PUF. There...     »