Methods and systems for PUF reliability estimation using high-dimensional stochastic modeling
Computing & Wireless : Computing Methods
Available for licensing
- Michael Orshansky, Ph.D. , Electrical and Computer Engineering
- Ye Wang, B.S. , Electrical & Computer Engineering
Silicon physical unclonable functions (PUFs) are widely used in emerging hardware security applications such as device identification, authentication, and cryptographic key generation. PUFs generate unique randomness by exploiting inherent random process variation during modern IC fabrication. Fuzzy extractors used in PUF-based key generators include error-correcting code (ECC) blocks. In order for ECC to operate reliably in reconstructing the key under all possible working conditions, the bit error rate under environmental variations needs to be estimated with high confidence. If the error rate that a block of data exhibits exceeds the maximum assumed BER, the ECC will fail in reconstructing the key.
The cost of ECC and storage requirements for helper data grows rapidly with the error rate that needs to be handled, and thus any misprediction of error rates is highly costly. The proposed invention develops a method that improves the prediction accuracy.
The invention allows to reliably estimate the worst-case error rate thereby allowing a significant cost (in terms of throughput, power, energy, silicon area, and price) reduction of ECC blocks and the storage requirement for helper data.
IC design houses, semiconductor manufacturers
- 1 PCT patent application filed