April 7-10, 2025 | Toronto, Canada

Health Testing with Sample Variance (N23b)

19 Sep 2024
16:00

Health Testing with Sample Variance (N23b)

Due to the nature of the SP 800-90B requirements, entropy sources often implement only the minimum required testing, namely the adaptive proportion test (APT) and repetition count test (RCT). This talk suggests that entropy drops in raw data may be detected more efficiently using a sample variance test (SVT). Efficiency gains can result in reduced data buffering and increased entropy source throughput. The talk describes how to implement the SVT on a window of data and explores its application to both Normally and non-Normally distributed raw data samples.