TRNGs and PUFs: The Foundation of Embedded Security

Modern day electronic devices are connected using wired and wireless networks. These devices need ways to authenticate each other and to communicate securely in an environment where their messages may be observed by an attacker. For this purpose, they use procedures called keyexchange protocols and entity authentication protocols. In order for these protocols to work, each device needs to have access to random numbers – these are numbers that cannot be guessed or predicted by any outside observer. Random numbers are usually generated by hardware modules called True Random Number Generators (TRNGs). These modules produce random numbers from unpredictable physical processes such as thermal noise that is present in all electronic circuits. Unique identifiers of electronic circuits, similar to human fingerprints, are required for easy authentication of devices. They are implemented using hardware modules called Physically Unclonable Functions (PUFs).
TRNGs and PUFs are essential for achieving secure communication and identification. My research proposal focuses on theoretical limits of these building blocks. For example: One of the research questions that I aim to answer is: How many random bits can be generated per 1 nJ of energy. The goal is to provide theoretical benchmarks for TRNG and PUF designers to evaluate their future designs. In addition, I plan to work on development of novel TRNG and PUF types.