### Abstract

Physical Unclonable Functions (PUFs) are a resource for generating and sharing secret keys.

The mutual information between two respective observations of the same PUF gives an upper bound for the achievable secret-key rate of a secret-sharing scheme that relies on this PUF.

This mutual information can be increased by including side information about the source.

We show for a given statistical model of the SRAM-PUFs, how side information can be estimated from multiple observations of an SRAM cell. Finally, we calculate the achievable increased secret-key rate given the estimated side information.

The mutual information between two respective observations of the same PUF gives an upper bound for the achievable secret-key rate of a secret-sharing scheme that relies on this PUF.

This mutual information can be increased by including side information about the source.

We show for a given statistical model of the SRAM-PUFs, how side information can be estimated from multiple observations of an SRAM cell. Finally, we calculate the achievable increased secret-key rate given the estimated side information.

Original language | English |
---|---|

Title of host publication | Proceedings of the 2017 Symposium on Information Theory and Signal Processing in the Benelux, 11-12 May 2017, Delft, The Netherlands |

Place of Publication | Delft |

Publisher | Technische Universiteit Delft |

Pages | 125-132 |

Publication status | Published - 2017 |

Event | Symposium on Information Theory and Signal Processing in the Benelux - Delft, Netherlands Duration: 11 May 2017 → 12 May 2017 |

### Conference

Conference | Symposium on Information Theory and Signal Processing in the Benelux |
---|---|

Abbreviated title | SITB2017 |

Country | Netherlands |

City | Delft |

Period | 11/05/17 → 12/05/17 |

### Fingerprint

### Cite this

Kusters, C. J., Ignatenko, T., & Willems, F. M. J. (2017). Behavior of temperature dependent SRAM-PUFs, and consequences for secret-key capacity. In

*Proceedings of the 2017 Symposium on Information Theory and Signal Processing in the Benelux, 11-12 May 2017, Delft, The Netherlands*(pp. 125-132). Delft: Technische Universiteit Delft.