A Covert Timing Channels Data Encryption Sceme In Cloud Simulation

Abstract

Covert Timing Channels (CTC) have become an impending network security problem as the sophistication and use of data exfiltration carried out by cyber-attacks has increased. Inter-arrival periods are used by these channels to steal sensitive data from targeted networks. Machine learning approaches are increasingly being used to detect CTCs, which use statistical-based measures to distinguish malicious (covert) traffic flows from genuine (overt) traffic flows. Given the attempts of cyber-attacks to elude detection and the expanding column of CTCs, covert channels detection must increase in both performance and precision in order to detect and prevent CTCs, as well as reduce the quality of service degradation caused by the detection process. We provide a new image-based method for fully autonomous vehicles in this research. Our strategy is based on the fact that covert channels provide communications that can be transformed into colored  visuals. Our approach is based on this observation and is meant to detect and find the malicious part (i.e., a sequence of packets) within a traffic flow automatically. Our technique lowers the drop in service quality caused by blocking complete traffic flows in which hidden channels are found by finding the covert components within traffic flows. To detect covert traffic, we first convert traffic flows into colored images and then extract image-based attributes. We use these attributes to train a classifier on a huge dataset of covert and overt traffic. We use these attributes to train a classifier on a huge dataset of covert and overt traffic. This method achieves remarkable results, with a detection accuracy of 95.83 percent for cautious CTCs and a covert traffic accuracy of 97.83 percent for 8-bit covert messages, much above the capabilities of commonly used statistical-based solutions.