The Study of Network Traffic Classification Strategies using Machine Learning

Abstract

The Network Traffic Classification (NTC) is the central topic in the field of computer science. This topic has gone to the front line in recent years since the organizations are expanding utilization of the web and restricted bandwidth. It is the initial step in the network traffic analysis and distinguishes various kinds of applications streaming in a wide area network. Through this technique, Internet Service Providers (ISP) or network administrators can monitor or analyze network performance. There exist numerous conventional strategies to classify network traffic such as Port-Based, Payload Based, etc. Anyway, the ultimate goal of these methods is to increase the performance of the network and the aforementioned methods failed to give optimal results. Machine Learning (ML) on the other hand, an active research area and widely applicable in many kinds of applications to analyze the network traffic. In this paper, we presented a brief overview of NTC and necessity of it in the internet world. Further, we included how specialists have applied machine learning strategies in few classification techniques using the network flow statistical properties.  Furthermore, we also outlined the next phase of our research, includes exploring diverse classification techniques (i.e. supervised, unsupervised, and semi-supervised) that utilize ML algorithms to adapt to real-world network traffic.