Applications and use Cases of Multilevel Granularity for Network Traffic Classification

Abstract

Network traffic classification is a fundamental process in network management and security. It allows network administrators to classify traffic based on various levels of classification granularity such as the source type or application. Existing literature focuses on analyzing the entire network traffic classification process with emphasis on the classification techniques. However, besides classification techniques, the literature lacks coverage on classification granularity, which deserves proper attention due to its increasing application in modern networks. Understanding the various levels of classification granularity and their use cases allow for more optimized traffic classification. As such, this paper aims to explore the different levels of classification granularity and their use cases. We studied papers published between 2013 and 2019 in order to investigate the different levels of granularity and use cases in the literature. As a result, this paper groups the classification granularity into a systematic multilevel taxonomy to assist in attaining a deeper understanding of their applications. Finally, to motivate future research, we elaborated on the current challenges and future directions for network traffic classification.