UTMobileNetTraffic2021: A Labeled Public Network Traffic Dataset
Yuqiang Heng, Vikram Chandrasekhar, Jeffrey G. Andrews
Abstract
A high-quality network traffic dataset is essential to the development of accurate network traffic classification algorithms. In this work, we present a new labeled public network traffic dataset with realistic mobile traffic from a wide range of popular applications. An automated platform is constructed to generate and collect data traffic from specified applications in a controlled environment. The dataset contains over 21 million packets from more than 29 hours of mobile traffic with application and activity-level labels. We provide an application classification example using machine learning (ML) models trained on the proposed dataset.
Topics & Concepts
Traffic classificationComputer scienceTraffic generation modelNetwork packetData miningCellular networkDeep packet inspectionArtificial intelligenceMachine learningComputer networkInternet Traffic Analysis and Secure E-votingNetwork Security and Intrusion DetectionAdvanced Malware Detection Techniques