Characterization of Internet traffic has become over the past few years one of the major challenging issues in telecommunication networks. As the Internet continuously grows in size and complexity, the need of an in-depth understanding of the underlying network traffic becomes evident. Internet Service Providers (ISP) must understand the composition and the dynamics of Internet traffic, in order to perform accurate capacity planning, deploy efficient management policies and pricing strategies, assess protocol performance, detect traffic abnormality, and the like. Therefore measurement, modeling, and analysis of Internet traffic have been always facing new challenges as new applications are continuously deployed while access and backbone network link capacities increase. Research studies in this topic have been focusing on scalable traffic identification and characterization, based either on inference methods or deep packet inspection (DPI) techniques. Profiling internet traffic also faces challenges such as scalability and privacy. Researchers always emphasize the importance of developing scalable inference methods for traffic classification, balancing accuracy and data volume in backbone measurements, as well as dealing with legal and ethical issues regarding access to packet payload.

In this talk, I’ll give an overview of some of my research contributions in this field. In general, my research work focuses on understanding the traffic profiles of Internet applications and their impact on the management of the underlying networked infrastructure. Specifically, my research studies aim at investigating tools, techniques, and mechanisms to support high performance traffic measurement and analysis in high-speed networks. Statistical analysis of Internet traffic also plays an important role for traffic measurement, modelling, and analysis. First, I’ll present the main research challenges in the topic of Internet Traffic Profiling for identification and classification purposes. Then, I’ll show recent research developments for traffic identification and classification from different fields, including Machine Learning.