LiteEx: A Lightweight Feature Extraction Tool for Captured Network Traces

Abstract

Network security researcher community widely uses network traffic analysis for various activities such as threat detection, threat prevention and security monitoring. Moreover, machine learning is widely used nowadays in various applications of network security. Therefore, it is important to extract features from network traffic traces which are then fed as input to machine learning algorithms to train the models and further these models are used for aimed specific tasks. However, there is a scarcity of freely available tools that can extract a wide variety of features from network packets and network flows without computationally overburdening the systems. To address this issue, we propose LiteEx which is a lightweight, flexible and GUI based feature extraction tool that can currently extract 68 packet-level features as well as 11 flow-level features from network traffic traces with low RAM and CPU utilization. We tested the performance of LiteEx on publicly available datasets having network traffic traces of PC, Smartphones, IoT devices etc. We tested the performance of LiteEx on two different machines in which one is high configuration and the other is a low configuration machine. After experiments, we found that LiteEx successfully extracted all kinds of features on 13.67 million packets with maximum CPU as 94.124% and maximum RAM utilization as just 8.775%. © 2023 IEEE.