Vulnerability Mining Method for the Modbus TCP Using an Anti-Sample Fuzzer

. 2020 Apr 5;20(7):2040. doi: 10.3390/s20072040

Algorithm 1: Anti-Sample Algorithm

input:

ζ

←the double-layer RNN model;

[{\hat{Y}}_{i}^{(0)}, {\hat{Y}}_{i}^{(1)}, \dots, {\hat{Y}}_{i}^{(t - 1)}]

←generated message;
output:

{\hat{Y}}_{i}

←the generated PDU message
▽based on the generated message

[{\hat{Y}}_{i}^{(0)}, {\hat{Y}}_{i}^{(1)}, \dots, {\hat{Y}}_{i}^{(t - 1)}]

, model

ζ

outputs its probability distribution matrix

θ

(1)
do $P_{f u z z} \leftarrow r a n d o m (0, 1)$ , $θ_{j} \leftarrow \max_{ξ} (θ)$
(2)
${\hat{Y}}_{i}^{(t)}_{}^{+} \leftarrow \underset{ξ}{\arg \max} (θ)$ , ${\hat{Y}}_{i}^{(t)}_{}^{-} \leftarrow \underset{ξ}{\arg \min} (θ)$
(3)
if $θ_{j} \geq P_{f u z z}$
(4)
then ${\hat{Y}}_{i}^{(t)} \leftarrow {\hat{Y}}_{i}^{(t)}_{}^{-}$
(5)
else ${\hat{Y}}_{i}^{(t)} \leftarrow {\hat{Y}}_{i}^{(t)}_{}^{+}$
(6)
${\hat{Y}}_{i} \leftarrow {\hat{Y}}_{i} + {\hat{Y}}_{i}^{(t)}$
(7)
$t \leftarrow t + 1$
(8)
do while ${\hat{Y}}_{i}^{(t)}! = 256$ ▽The end flag stops the loop
(9)
end while