Mechanism for Efficient Media Propagation in Event-Driven Cyber-Physical Systems

Many key applications in Cyber-Physical Systems require the transmission of speech, audio, or video. These scenarios involve the use of traditional Real-Time Communication (RTC) protocols and technologies, which cannot always be used in the context of core networks. This is particularly critical in the context of Event-Driven Architectures (EDAs), where RTC protocols require the use of complex topologies that rely on costly infrastructure. One way to avoid this is by encapsulating all media traffic in EDA protocols. However, this approach does not come without challenges. Specifically, the nature of the transport protocols causes the media to be heavily affected by application layer impairments that render their usage highly impractical. To prevent this from happening, this paper introduces a unified scheme that supports the efficient encapsulation of media traffic in EDA scenarios. This is accomplished through a mechanism that relies on a Machine Learning (ML) model that is exercised in an experimental framework.