What Helper Data Really Leaks: A Practical Approach to Estimate the Min-Entropy in PUFs Using Their Response Mass Function
TimeWednesday, July 13th6pm - 7pm PDT
LocationLevel 2 Lobby
Event Type
Networking Reception
Work-in-Progress Poster
DescriptionHelper data algorithms reliably extract secrets from Physical Unclonable Functions.
The necessary helper data can leak information, though.
The first approach to assess the remaining min-entropy was limited to homogeneous bias or correlation, not both.
A second one extended this to only local bias without correlation, but was limited to short code lengths.
This work presents a new approach based on convolving histograms.
It provides a good bound and good approximation given arbitrary bias, more realistic correlation effects, and practically relevant code sizes.
Experiments on real-world data show the efficiency and efficacy of the new methods compared with state-of-the-art ones.