Skip to main content
EURASIP Journal on Information Security

Table 1 NIST 800-22 statistical test result. We evaluated the random numbers by using the SP800-22 statistical tests [1] for reference. We used an NIST statistical test suite called “sts-2.1.2” [26] and the recommended parameter settings of AIS 20/31. We selected a MN-TRNG and acquired random numbers at the room temperature (25 °C) and the rated voltage (1.8 V)

From: ASIC implementation of random number generators using SR latches and its evaluation

 

Data set 1

Data set 2

Statistical test

p value

Proportion

p value

Proportion

Frequency

0.352760

0.9897

0.981009

0.9869

Block frequency

0.198458

0.9814

0.065912

0.9944

Cumulative sums*

0.017970

0.9888

0.100602

0.9879

Runs

0.392050

0.9888

0.423587

0.9879

Longest run

0.054577

0.9860

0.193532

0.9870

Rank

0.013864

0.9897

0.773966

0.9888

FFT

0.405158

0.9935

0.428692

0.9832

Non-overlapping template*

0.009459

0.9897

0.763056

0.9786

Overlapping template

0.725810

0.9851

0.067077

0.9907

Universal

0.571253

0.9879

0.731471

0.9907

Approxi. entropy

0.163454

0.9916

0.144099

0.9934

Random excursions*

0.095691

0.9985

0.120217

0.9807

Random excursions varian*

0.097473

0.9923

0.035873

0.9985

Serial*

0.034210

0.9935

0.557853

0.9925

Linear complexity

0.051110

0.9860

0.750162

0.9888

  1. Two data sets (data sets 1 and 2) were evaluated and the size of each data set was about 230 bits. The asterisk denotes that a test consisted of several sub-tests and that the minimum p value is shown. Data set 1 passed all the tests and data set 2 passed the tests except for non-overlapping template matching test. The values of failed tests are shown in italic type
Back to article page