Encrypted 5G Smallcell Backhaul Traffic Classification Using Deep Learning

5G small-cell base stations (“Smallcell”) have been widely deployed to enhance network capacity in densely populated city centers. Generally speaking, IPSec is used to secure the backhaul links between the Smallcell network and the core network. However, encryption sometimes serves as a tunnel to hide malware. Encrypted traffic classification of 5G Smallcell backhaul links is rarely discussed in the literature. To our knowledge, we are the first to classify encrypted 5G Smallcell backhaul links using 1D-CNN. We are able to classify 5G Voice, SMS and Internet data with 99.99% accuracy rate and above, and the model is validated using real network data. The work can be used for classifying real encrypted network traffic in general.