Ultrafast Physical Random Number Generation Using Hybrid Boolean Networks

Eno River Technologies and Duke University propose to develop Hybrid Boolean Network (HBN) technology for ultrafast physical random number generation. HBNs are circuits composed partly of unclocked logic elements that exhibit chaotic dynamics and clocked elements that provide readout and synchronization to the system bus. It is hypothesized that the chaotic elements amplify electronic noise in the circuit, which leads to true random number generation (TRNG). HBNs were originally developed by Duke University using low-cost FPGA technology and have been demonstrated to produce random numbers that pass the NIST statistical tests at 12.8 Gbit/s. The goal of Phase I is to mature simulation tools and the theory of operation, perform experiments on HBNs in different operating environments and on different platforms to demonstrate robustness of TRNG operation, and perform system studies to determine optimal HBN architecture. In Phase II, we will develop and demonstrate a network appliance capable of delivering 40 Gbit/s of random numbers for cryptographic applications.