A Survey of Network Traffic Classification

Abstract

This survey is dedicated to the task of network traffic classification, particularly to the use of machine learning algorithms in this task. The survey begins with the description of the task, its variations and possible uses in real-world problems. It then proceeds to the description of the methods used historically to solve this task, their limitations and evolution of traffic making machine learning the main way to solve the problem. Then the most popular machine learning algorithms used in this task are described, with the examples of research papers, providing the insight into their advantages and disadvantages in relation to this field. The task of feature selection is discussed, followed by the more global problem of acquiring the suitable dataset to use in the research; some examples of such popular datasets and their descriptions are provided. The paper concludes with the outline of the current problems in this research area to be solved.

CBS: A Deep Learning Approach for Encrypted Traffic Classification With Mixed Spatio-Temporal and Statistical Features

“…Both port-based and DPI approaches have certain limitations. The port-based approach could be more efficient at classifying traffic due to port obfuscation, random port number assignment, using dynamic ports, and the network address translation (NAT) technique [7], [8]. Furthermore, encryption techniques are widely employed in internet traffic to preserve privacy.…”

“…Consequently, extracting useful information from the packet payload is less effective, which results in less accurate classification [9]. The primary issue with the DPI method is its high computing overhead, which makes it unsuitable for real-time or encrypted traffic classification [7], [10]. Therefore, statistical flow analysis and machine learning approaches have been investigated to overcome the limitations of the two previous methods.…”

“…Feature engineering is a critical aspect of traffic classification, involving feature extraction and selection techniques [5], [7]. Selecting the most effective traffic features for the classifier is essential to achieving high classification accuracy, which depends on several factors.…”