1 Montgomery Modular Multiplication in Hard- ware
1 Montgomery Modular Multiplication in Hard- ware
1 Montgomery Modular Multiplication in Hard- ware
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
FEI KEMT<br />
Figure 6 – 8 Sampled waveform of a clock signal for TRNG for Configuration A for temperatures<br />
<strong>in</strong> range −40 ◦ C + 30 ◦ C.<br />
The random sequences were tested by simple statistical tests def<strong>in</strong>ed <strong>in</strong> FIPS<br />
standard [57]. The test suite can reveal a bias or unbalanced distribution of zeros<br />
and ones <strong>in</strong> generated sequence by application of 4 basic tests (monobit test, poker<br />
test, runs and long runs tests). If at least one test from the set was not passed, the<br />
result is denoted as FAILED, otherwise we put OK mark.<br />
In Table 6 – 9 we summarise the results of statistical tests at different FPGA chip<br />
temperatures. It can be seen that while the configuration A has produced by some<br />
temperatures the sequences that did not pass the statistical tests, the configuration<br />
B is reliable <strong>in</strong> the whole range of temperatures.<br />
The columns with critical samples number show the number of samples <strong>in</strong>fluenced<br />
by jitter. It can be observed that <strong>in</strong> case of the configuration B, when we have set<br />
a low bandwidth of the loop filter, the number of <strong>in</strong>fluenced samples is significantly<br />
higher.<br />
We further <strong>in</strong>vestigate the number and position of critical samples for both con-<br />
figurations <strong>in</strong> dependency on the chip temperature. Crucial impact on the statistical<br />
parameters of the generated sequence have the samples with probability around 0.5.<br />
In case we elim<strong>in</strong>ate the almost constant samples, with less than 100 by jitter <strong>in</strong>-<br />
fluenced values dur<strong>in</strong>g 1000 periods, there are 4-6 and 12-13 highly critical samples<br />
per edge for configuration A and B, respectively.<br />
The Figures 6 – 10 and 6 – 11 show <strong>in</strong> details the area of ris<strong>in</strong>g edge of the sampled<br />
waveform. We can observe how dur<strong>in</strong>g the measur<strong>in</strong>g period the number of sampled<br />
116