Covert Timing Channels Detection Based on Image Processing Using Deep Learning

With the development of the Internet, covert timing channel attacks have increased exponentially and ranking as a critical threat to Internet security. Detecting such channels is essential for protection against security breaches, data theft, and other dangers. Current methods of CTC detection have shown low detection speeds and poor accuracy. This paper proposed a novel approach that used deep neural networks to improve the accuracy of CTC detection. The traffic inter-arrival times are converted into colored images; then, the images are classified using a CNN that automatically extracts the image's features. The experimental results demonstrated that the proposed CNN model achieved better performance than other detection models.