Master of Science Internship project:
SRAM-PUF statistical analysis
Location: Eindhoven, The Netherlands
Within the research group of Intrinsic-ID we are looking for a motivated student (Electrical Engineering, Mathematics, or Computer Science), who is looking for an internship project during the M.Sc. phase of his/her university education. Intrinsic-ID is an innovative startup that has originated from Philips Research and focuses on a technology called Physical Unclonable Functions (PUFs).
PUFs can be seen as “biometrics of a device” or a device specific fingerprint. Variations during manufacturing lead to unique characteristics on ICs. Since the variations are uncontrollable, these characteristics are unclonable. Consider for example SRAM memory blocks that all have their own startup behavior, resulting in a kind of fingerprint, unique for each such block. In principle each SRAM cell should assume each of the two possible states (0 and 1) with equal probability during startup, but as a result of random deviations occurring in the production process one state is more likely to occur than the other. This leads to a unique fingerprint for such a memory block, which is inherently unclonable. Using data processing (error correction and entropy extraction), Intrinsic-ID converts these fingerprints into cryptographic keys that can be used to solve security use cases for authentication, secure data storage, key management, secure boot and many others.
Intrinsic-ID has built up a large database of SRAM PUF measurements from all kinds of SRAMs (various brands, compiler settings, technology nodes) under a wide variety of conditions (temperature, voltage, ageing, power-off times, power ramp-up times). The goal of this project is to conduct detailed statistical analyses on these data in order to justify certain assumptions that are made in the design of data processing algorithms for the cryptographic keys. Statistical aspects to be investigated are for example: independence of bits, bit probabilities and correlations depending on location, randomness of bits (independency over devices), and randomness/noise entropy after ageing.
Responsibilities include, but are not limited to
- Making models for the behavior of SRAM PUFs;
- Design statistical tests in Matlab/Octave to validate the models;
- Run, analyze and document test outcomes.
Required experience and appreciated qualities
- Student (Electrical Engineering, Mathematics, or Computer Science) MSc phase of university education;
- Experience with statistical analysis;
- Experience with mathematical software tools such as Matlab/Octave;
- Experience with scripting and test automation (preferable, not mandatory);
- Interest in and experience with security is considered to be a plus;
- Able to think outside of the box: find unexpected ways to solve problems or meet requirements;
- Good organizational skills and excellent problem solving abilities.
Intrinsic-ID is a high tech security company that spun out from Philips. We develop, market and sell truly unique security products based on the unique silicon fingerprint of a chip. As a leader in this field we provide the root of trust, authentication and key management solutions to blue chip semiconductor companies and OEMs in the mobile, IoT, embedded and government markets. Our headquarters are in Silicon Valley but our R&D office is in Eindhoven (The Netherlands). Currently we are seeking highly motivated individuals to help us grow our company and further expand our product portfolio to provide ubiquitous security for the connected world.
High Tech Campus 9
5656 AE Eindhoven