A generalized approach to estimation of memoryless covert channel information leakage capacity

Estimating the amount of information that is leaked by covert channels is a necessity to comprehend and mitigate the severity of attacks exploiting these channels. Having such an estimation in design-state provides an opportunity for designers to adjust their systems to minimize information leakage. In this paper, we propose a methodology to estimate the worst-case information leakage (or capacity for the information leakage) through various memoryless covert channels – both analog and digital ones – exhibiting on–off keying structure. In that respect, we first model the communication channel as a deletion–insertion channel to account for the information losses due to software activities. Then, we derive the effective noise in covert channels as the combination of jitter noise caused by signaling time variation and Additive White Gaussian Noise (AWGN). Considering this effective noise, we propose a communication model which can be generalized for various covert channels and takes insertions, deletions, and asynchronous nature of covert channels into account. By leveraging the link between this communication model and information theory literature, we obtain the information leakage capacity that reveals the leakage limit for the worst-case scenarios. Finally, we provide experimental results to demonstrate that the proposed model is an effective and a generalized methodology to score the resilience of a given system to covert channel attacks.