Download Full Text (823 KB)

Faculty Sponsor

Dr. Anyi Liu


Department of Computer Science

University Affiliation

Indiana University – Purdue University Fort Wayne


Despite of extensive research for malware and Trojan horses, covert channel (CC) attacks are still considered as one of the topmost insider attacks in information security. The most virulent feature of this attack is that it can be launched through either specially crafted content, or manipulated timing characteristics to transmit sensitive inside information to outside adversaries, all the while, its malicious behavior remains undetectable.

In a mobile cloud environment, the CC is particularly harmful and difficult to detect because malicious mobile users can freely join and leave the cloud and harvest the collected information through CCs. In addition, information can be hidden in some features, which are unique to mobile devices. Previous technologies mainly focused on designing proof-of-concept CCs on mobile devices, which are suffer from the low-bandwidth of its communication channel and easy to be detected.

We present MediaChannel, a CC generator that leaks the sensitive information to malicious mobile users and outwits most security defenses. The information can be hidden into video and audio streams, and be conveyed through WIFI or bluebooth traffic. To ensure robustness of the hidden information and higher bit-rates than most of the known CCs, we proposed a scheme that dynamically embeds information into video frames. Once the information is been embedded into the frames, it is difficult to detect and remove it. The proposed encoding and decoding algorithms ensure sender and receiver follow the same protocol and no additional information is needed to be transmitted. To facilitate robust communication, various error-correction mechanisms have been applied to MediaChannel, such that it can automatically correct errors due to signal interference. Our implementation demonstrates MediaChannel can archive the bandwidth several orders of magnitude higher than the previously studied covert channels on mobile devices. which prove Mediachannel a realistic threat to mobile users.


Computer Sciences | Physical Sciences and Mathematics

The exploits and prevention of covert channels in mobile cloud