Security evaluation of Biosec

In 2003, Cherukuri et al. proposed Biosec, a novel key distribution protocol where two (implanted) sensors that do not share any secret on beforehand can establish a key through the use of cryptography in combination with biometrics. Their solution makes use of an error correction code to tolerate some errors in the biometrics, and is the first to use patient physiological signals to securely transport a shared session key. In this paper, we evaluated the security of their protocol and identified several weaknesses. More in particular, we observed that the key is reused for multiple cryptographic primitives, and that, as stated by the authors, the biometrics randomness level may not be sufficient to fully conceal the key. In addition, we show that the effective length of the key is shorter than the one claimed by the authors due to the use of error correction codes.