Telltale Twin - Predicting Latency of Unseen Subflow of Multipath Transmission Control Protocol Using Machine Learning

In Multipath Transmission Control Protocol (MPTCP), the subflows are not totally independent. Due to the correlation between subflows, information related to the unseen subflow can be extracted from the seen (observed) subflow. This opened a new avenue for network monitoring which was not available with traditional Transmission Control Protocol (TCP) connections. In this study, the network condition (latency) of the unseen subflow was predicted using information observed from the seen subflow. Data flow information from the observed subflow was encoded using six different ways and used across five different machine learning techniques to predict the latency class of the unseen sub-flow which expand the perimeter of the network monitoring.