Physically Unclonable Functions (PUFs) are security primitives that provide trustworthy hardware for key-generation and device authentication. Among them, in contrast to dedicated PUFs, intrinsic PUFs are created from existing hardware components that exploit their variability through software. In this work we focus on GPUs and present G-PUF, a PUF implemented entirely in software on CUDA and hence does not require hardware modifications. Our results show that G-PUF has comparable characteristics to SRAM and DRAM PUFs in terms of uniformity 55.61% and reliability 90.09%.

Multi-Processor System-on-Chips (MPSoCs) are popular computational platforms for a wide variety of applications due to their energy efficiency and flexibility. Like many other platforms they are vulnerable to Side Channel Attacks (SCAs). In particular, Logical SCAs (LSCAs) are very powerful as sensitive information can be retrieved by simply observing system properties that depend on the victim's software execution on the MPSoC. Unfortunately, many of the current protection mechanisms are either platform dependent or are effective only against a reduced set of attacks. In this work, we present Guard-NoC, a secure Network-on-Chip (NoC) architecture able to protect MPSoCs against a wide variety of LSCAs. The secure NoC employs three application-independent strategies to hide and isolate sensitive information: i) blinding the execution time of operations; ii) masking the execution time of operations; and iii) dual communication strategy (i.e., use packet and circuit switching simultaneously). Our results show that our secure NoC is resilient against practical LSCAs and leaks almost no information while having a minimal area and power overhead.