Study And Implementation Of Automatic Covert Channel Detection Based On Information Flow

Covert channels can bypass the check of security strategies to transfer information illegally in a quite covert manner, which poses a great threat to information security and increases the difficulty of covert channel detection. Via analyzing typical covert channel detection technologies and methods, this thesis proposes automatic covert channel detection based on information flow to detect source codes.The thesis firstly categories existing covert channel detection technologies, introduces related detection methods, and on this base, the detection method based on information flow is determined; The formal definition of covert channel is given, information flow and its properties are defined and discussed, and the detection framework is designed and described detailedly. Subsequently, to obtain complete information flows and carry out covert channel path search effectively, information flow generation rules based on Control Flow Graph and the properties of information flow are formulated, and constrained reverse information flow graph search algorithm is designed. Finally, to decrease the false warning rate of covert channels, the thesis discusses and summaries the reasons for false covert channels, proposes syntax and semantics filtering rules, and related algorithms are designed to delete false covert channel paths as well.The detection method in the thesis is tested by Linux codes, and the results obtained in each detection step are analyzed, which verifies the method in the thesis is effective.