Trustworthy Proofs for Sensor Data using FPGA based Physically Unclonable Functions
Urbi Chatterjee1,a, Durga Prasad Sahoo2, Debdeep Mukhopadhyay1,b and Rajat Subhra Chakraborty1,c
1Secured Embedded Architecture Laboratory, IIT Kharagpur, India
aurbi.chatterjee@cse.iitkgp.ernet.in
bdebdeep@cse.iitkgp.ernet.in
crschakraborty@cse.iitkgp.ernet.in
2Bosch India (RBEI/ETI)
dpsahoo.cs@gmail.com
ABSTRACT
The Internet of Things (IoT) is envisaged to consist of billions of connected devices coupled with sensors which generate huge volumes of data enabling control‐and‐command in this paradigm. However, integrity of this data is of utmost concern, and is promisingly addressed leveraging the inherent unreliability of Physically Unclonable Functions (PUFs) w.r.t. ambient parameter variations, using the concept of Virtual Proofs (VPs). Advantage of these protocols is that they do not use explicit keys and aim at proving the authenticity of the sensor. Since the existing PUF‐based protocols do not use the sensor data as a part of challenge (i.e. input) to PUFs, there is no guarantee of uniqueness of PUF's challenge‐response behavior over multiple levels of ambient parameters. Few of these protocols needs to sequential search in the challenge‐response database. To alleviate these issues, we develop a new class of authenticated sensing protocols where the sensor data is combined with the external challenge by utilizing the Strict Avalanche Criterion of the PUF. We validate the proposed protocol through actual experiments on FPGA using Double Arbiter PUFs (DAPUFs), which are implemented with superior uniformity, uniqueness, and reliability on Xilinx Artix‐7 FPGAs. According to the FPGAbased validation, the proposed protocol with DAPUF can be effectively used to authenticate wide variations of temperature from –20°C to 80°C.
Keywords: Authenticated Sensing, Double arbiter PUFs, FPGA, Physically unclonable functions (PUFs), Reliability, Trustworthy proofs, Virtual proofs (VPs).
