Estimating Identification, Characterisation and Classification of Digital Health Traffic Using the Shannon Entropy

This paper proposes using the Shannon entropy to
identify, characterise and classify digital health data traffic on a 5G+ digital health testbed. Firstly, the digital health data traffic types are categorised into five main classes. Count-Min sketches are proposed to count traffic instances and map parameters to specific traffic types on a programmable switch. Thirdly, a framework containing three algorithms is proposed to calculate
the information entropy of the data stream, identify traffic types in the stream, and allocate the traffic types to their respective slices via the next-hop IP address. The paper’s main contributions are the proposed classification of digital health data traffic and the usage of the Shannon entropy theorem in classifying the data traffic, which the proposal of three algorithms has demonstrated.