Reliability bound and channel capacity of IBS-based fuzzy embedders

Physical Unclonable Functions (PUFs) are an alternative to conventional secure key storage in embedded devices. PUFs exploit unique sub-micron manufacturing variations in each device to generate characteristic responses. Each response is subject to noise, so a so called fuzzy embedder is required to use a PUF for reliable key reproduction. Index-Based Syndrome coding (IBS) is an information theoretically secure fuzzy embedder, introduced in 2010 [1]. This work derives a lower bound for the reliability of IBS-based fuzzy embedders, that depends on the characteristics of the PUF and the parameters of the fuzzy embedder. In addition, the channel capacity is applied as new criterion for fuzzy embedder evaluation. In this case, the channel capacity determines the maximum key size that can be embedded. The presented measure is complementary to security analysis and thus adds a new dimension for evaluating future encoding schemes for IBS-based fuzzy embedders. A case study demonstrates the practical impact of the contributions.