With the ability to leak confidential information in a secret manner, covert channels pose a significant threat to the confidentiality of a system. Due to this threat, the identification of covert channel existence has become an important part of the evaluation of secure systems. In this paper, we present a state-of-the-art survey discussing the conditions for covert channel existence found in the literature and we point to their inadequacy. We also examine how conditions for covert channel existence are handled by information theory. We propose a set of necessary and verifiable conditions for covert channel existence in systems of communicating agents. We aim to provide an improved understanding of covert channel communication and to build a foundation for developing effective and efficient mechanisms for mitigating covert channels in systems of communicating agents at the early stages of software development
