Covert Channel Detection: Machine Learning Approaches

The advanced development of computer networks and communication technologies has made covert communications easier to construct, faster, undetectable more secure than ever. A channel is a path through which secret messages can be leaked by violating system security policy. detection such dangerous, unwatchable, hidden threats still one the most challenging aspects. This threat exploits methods that are not dedicated purposes, meaning traditional measures fail detect its existence. review introduced brief introduction definitions, types developments, with particular focus on techniques using machine learning (ML) approaches. It provides thorough common channels ML used counter them, as well addressing their achievements limitations. In addition, this paper introduces comparative experimental study for some approaches commonly in field. Accordingly, performance these classifiers was evaluated reported. concludes our information at risk, nothing said secured work required.